AU705984B2 - Corticotropin-releasing factor-binding protein inhibitors and their use - Google Patents

Description

WO 96/02569 PCT/US95/08867 1 Description Corticotropin-releasing factor binding protein inhibitors and their use.

Certain aspects of this invention were made with Government support under Grants DK-26741 and HD-13527 awarded by the National Institutes of Health.

The Government has certain rights in the invention. The University of Reading and the Medical Research Council of Great Britain have certain rights to this invention.

Cross-Reference to Related Application This application is a continuation-in-part of United States Patent Application Serial Number 08/276,240, filed July 15, 1994.

Technical Field The present invention relates generally to methods of increasing endogenous levels of neuropeptides, and specifically to increasing corticotropinreleasing factor levels in the brain.

Background of the Invention Recent clinical data have implicated CRF in neuropsychiatric disorders and in neurodegenerative diseases, such as Alzheimer's disease. Alzheimer's disease is a neurodegenerative brain disorder which leads to progressive memory loss and dementia. By current estimates, over two million individuals in the United States suffer from this disease. In particular, several lines of evidence have implicated CRF in Alzheimer's disease First, there are dramatic (greater than 50%) decreases in CRF (Bissette et al., JAMA 254:3067, 1985; DeSouza et al., Brain Research 397:401, 1986; Whitehouse et al., Neurology 37:905, 1987; DeSouza, Hospital Practice 23:59, 1988; Nemeroff et al., Regul. Peptides 25:123, 1989) and reciprocal increases in CRF receptors (DeSouza et al., 1986; DeSouza, 1988) in cerebrocortical areas that are affected in AD, while neither CRF nor CRF receptors are quantitatively changed in non-affected areas of the cortex (DeSouza et al., 1986). Second, chemical affinity cross-linking studies indicate that the increased CRF receptor population in cerebral cortex in AD have normal biochemical properties (Grigoriadis et al., Neuropharmacology 28:761, 1989). Additionally, observations of decreased concentrations of CRF in the cerebrospinal fluid (Mouradian et al., Neural Peptides 8:393, 1986; May et al., Neurology 37:535, 1987) are significantly correlated with the WO 96/02569 PCT/US95/08867 2 global neuropsychological impairment ratings, suggesting that greater cognitive impairment is associated with lower CSF concentrations in cerebrospinal fluid (Pomara et al., Biological Psychiatry 26:500, 1989).

Available therapies for the treatment of dementia are severely limited.

Tacrine T M a recently approved drug, leads to only marginal memory improvement in Alzheimer's patients, and has the undesirable side effect of elevating liver enzymes.

Alterations in brain CRF content have also been found in Parkinson's disease and progressive supranuclear palsy, neurological disorders that share certain clinical and pathological features with AD. In cases of Parkinson's disease, CRF content is decreased and shows a staining pattern similar to cases of AD (Whitehouse et al., 1987; DeSouza, 1988). In progressive supranuclear palsy, CRF is decreased to approximately 50% of control values in frontal, temporal, and occipital lobes (Whitehouse et al., 1987; DeSouza, 1988).

Some depressive disorders are also associated with decreased levels of CRF. Patients in the depressive state of seasonal depression and in the period of fatigue in chronic fatigue syndrome demonstrate lower levels of CRF in the cerebrospinal fluid (Vanderpool et al., J. Clin. Endocrinol. Metab. 73:1224, 1991).

Although some depressions have a high improvement rate and many are eventually self-limiting, there are major differences in the rate at which patients recover. A major goal of therapy is to decrease the intensity of symptoms and hasten the rate of recovery for this type of depression, as well as preventing relapse and recurrence. Anti-depressants are typically administered, but severe side effects may result suicidality with fluoxetine, convulsions with bupropion). (See Klerman et al. in Clinical Evaluation of Psychotropic Drugs: Principles and Guidelines, R.F.

Prien and D.S. Robinson Raven Press, Ltd. 1994, p. 281.) Hypoactivation of the stress system as manifested by low CRF levels may play a role in other disorders as well. For examples, some forms of obesity are characterized by a hypoactive hypothalamic-pituitary-adrenal axis (Kopelman et al., Clin. Endocrinol (Oxford) 28:15, 1988; Bernini et al., Horm. Res. 31:133, 1989), some patients with post-traumatic stress syndrome have low cortisol excretion (Mason et al., J. Neu. Men. Dis. 174:145, 1986), and patients undergoing withdrawal from smoking have decreased excretion of adrenaline and noradrenaline, as well as decreased amounts of cortisol in blood (West et al., Psychopharmacology 84:141, 1984; Puddy et al., Clin.

Exp. Pharmacol. Physiol. 11:423, 1984). These manifestations all point to a central role for CRF in these disorders because CRF is the major regulator of the hypothalamic-pituitary-adrenal axis.

WO 96/02569 PCT/US95/08867 3 Treatments for these disorders have poor efficacy. For example, the most effective approach to treatment of obesity is a behavior-change program.

However, few participants reach goal weight and the relapse rate is high (see Halmi et al. in Clinical Evaluation of Psychotropic Drugs: Principles and Guidelines, R.F.

Prien and D.S. Robinson Raven Press, Ltd. 1994, p. 547).

In view of the deficiencies in treatments for such disorders and diseases, more effective treatments are needed. The present invention exploits the correlation of reduced levels of CRF with various neuro-physiologically based disorders and diseases to effectively treat such diseases by increasing levels of free CRF, and further provides other related advantages.

Summary of the Invention The present invention provides methods for increasing the level of free CRF in the brain by administering to a patient an effective amount of a ligand inhibitor of a CRF/CRF-binding protein (CRF/CRF-BP) complex. Administration of the ligand inhibitor causes release of CRF from the CRF-binding protein. The ligand inhibitor may be a peptide derived from CRF, homologous to CRF, or unrelated to CRF as long as it is capable of causing the "release" of CRF. Ligand inhibitors may also be nonpeptide compounds which are isolated from natural or synthetic chemical libraries.

Within one embodiment of the invention, the ligand inhibitor is a peptide selected from the group consisting of h/rCRF (amino acids 6-33), h/rCRF (amino acids 9-33), and h/rCRF. Within a related aspect, therapeutic compositions are provided comprising the ligand inhibitor in combination with a physiologically acceptable carrier or diluent.

Within other aspects of the invention, methods for improving learning and memory, decreasing food intake, activating CRF neurocircuitry, treating diseases associated with low levels of CRF in the brain, treating symptoms associated with Alzheimer's disease, treating obesity, treating atypical depression, treating post-partum depression, treating age-related memory loss, and treating substance abuse withdrawal are provided. Within such methods, a therapeutically effective amount of a ligand inhibitor of a CRF/CRF-binding protein is administered to a patient as treatment for these conditions. Criteria for choosing candidates for therapy are presented, as well as methods for assessing efficacy of treatment.

It has been found that certain agents having a high affinity to human CRF-BP can be administered which will effectively compete with human CRF in the formation of complexes with CRF-BP and will, in this manner, increase the effective in vivo concentration in a mammal of endogenous hCRF, and/or the effective concentration of a CRF agonist or CRF antagonist optionally administered along with WO 96/02569 PCTIUS95/08867 4 such agent, for the purpose of achieving a particular therapeutic purpose. In other words these agents serve to block the effect of CRF-BP and thus to increase the concentration of endogenous CRF in those regions of the body where CRF-BP is present. More specifically, peptides between about 19 and 28 residues in length have been discovered which have a high affinity to hCRF-BP but, which themselves exhibit relatively little propensity to bind to the CRF receptor. As a result, such peptides can be administered to prevent the clearance of endogenous CRF from particular regions and thereby stimulate the biological effect of CRF in vivo, and in certain instances, it may be advantageous to administer such peptides along with CRF or a CRF agonist.

The very nature of these agents is such that potentially undesirable side effects are minimized or totally obviated. They might also be administered along with CRF antagonists to prevent the clearance of some CRF antagonists from a target region particularly if the CRF antagonist had a fairly high binding affinity to hCRF-BP; however, the effect is counteracted to some extent by the release of endogenous CRF that would otherwise be bound to CRF-BP. These agents are useful for therapeutic treatment to promote parturition in pregnancy, to stimulate the respiratory system, to combat obesity, and to counteract the effects of Alzheimer's disease, and of chronic fatigue syndrome; however, for some of these indications, the agents must be administered in a manner so that they are delivered to the brain.

In addition, methods are provided for screening for particularly effective CRF antagonists by carrying out dual screening assays of such potential antagonists to determine their potential for blocking the ability of CRF to bind to native CRF receptors and to also determine the binding affinity between such a candidate CRF antagonist and hCRF-BP.

Methods are also provided for screening for neuropeptide-binding proteins. In a related aspect, methods are provided for screening for ligand inhibitors of neuropeptide/neuropeptide-binding protein complex, and in particular the CRF/CRF- BP complex. Within one embodiment, this method comprises contacting CRF with CRF-BP in the presence of a ligand inhibitor in an aqueous solution, further mixing in a nonionic detergent, separating the nonionic detergent and the aqueous solution, and detecting the amount of CRF in the aqueous solution, thereby determining whether the ligand inhibitor disrupts the CRF/CRF-binding protein complex. In certain embodiments, the mixing is performed at a temperature above the cloud point of the nonionic detergent and octylphenoxypolyethoxyethanol is a preferred nonionic detergent.

These and other aspects will become evident upon reference to the following detailed description and attached drawings.

WO 96/02569 PCT/US95/08867 Brief Description of the Drawings Figure 1 presents the amino acid sequences of CRF from human (hCRF) (SEQ ID NO:1), and sheep (oCRF) (SEQ ID NO:4).

Figure 2 is a graph showing the levels of CRF bound to CRF-BP and free CRF. CRF levels were determined in brain tissue from Alzheimer's patients and normal age-matched controls. Levels were established with or without the addition of a CRF-BP ligand inhibitor, a-helical ovine CRF(9-41).

Figure 3 is a graph showing the levels of CRF (panel A) and CRF-BP (panel B) in four areas of brain tissue taken from normal controls or Alzheimer's patients.

Figure 4 presents the test results of the Morris water maze and the elevated plus-maze following intracerebroventricular (ICV) injection of vehicle or CRF(6-33) or CRF Rats, in groups of 7-10 animals, received ICV injections 15 min prior to testing. In the Morris water maze test, times to reach the platform were recorded. In the elevated plus-maze test, percent time spent in the open arm was recorded. An asterisk indicates a statistical significant difference.

Figure 5 is a chart showing test results of rats in the Y-maze test.

Groups of 7-10 rats received either 0, 1, 5, or 25 Ljg CRF(6-33) in an ICV injection 15 min prior to testing. Percent correct responses were determined. An asterisk indicates a statistical significant difference.

Figure 6 presents test results of body weight changes and food intake of rats during continuous infusion of dependence-inducing levels of nicotine (dependence phase) and during a two-week abstinence (withdrawal phase). Body weight is represented by the hatched bars and is measured in grams; food intake is represented by a line and is measured in grams.

Figure 7 is a chart showing the effects of withdrawal from nicotine dependence on food intake following administration of a ligand inhibitor, h/r CRF(6-33).

Detailed Description of the Invention Prior to setting forth the invention, it may be helpful to an understanding thereof to set forth definitions of certain terms that will be used hereinafter.

"CRF" refers to a peptide that regulates the release of adrenocorticotropin (ACTH), P-endorphin, and other pro-opiomelanocortin (POMC)derived peptides from the pituitary. In humans, rats, and other species, the amino acid sequence of CRF has been determined and is presented in Figure 1 (SEQ. ID NO:1).

WO 96/02569 PCT/US95/08867 6 The amino acid sequences of rat and human CRFs are identical and the protein is referred to as "h/rCRF." "Free CRF" refers to CRF which is not complexed or bound to CRF-binding protein or CRF receptors.

"CRF-binding protein" (CRF-BP) refers to a protein or proteins present either as a soluble factor in human plasma or associated with cell membranes and that has the ability to inhibit the function of CRF as measured by one of two methods: CRF-induced ACTH release from cultured pituitary cells or from a perfused rat anterior pituitary system, or CRF-induced cAMP formation from cells possessing CRF receptors or from cells which have been transfected with cloned CRF receptors.

Examples of cDNA clones encoding CRF-BP have been isolated from human liver and rat brain (Potter et al., Nature 349:423, 1991).

"CRF/CRF-BP" refers to the complex of CRF and CRF-BP. Binding of CRF and CRF-BP may be through hydrophobic, ionic, or covalent interactions.

"Human CRF binding protein" (hCRF-BP) refers to a 37 kDa serum protein that, by specifically binding hCRF, inactivates hCRF as an ACTH secretogogue in vitro and in vivo. Human CRF-BP has a high affinity for hCRF and a low affinity for oCRF, suggesting hCRF-BP may expedite the elimination of peripheral plasma hCRF. hCRF loses its ability to stimulate ACTH in vitro and in vivo when bound to hCRF-BP. The first 8 amino acids of the CRFs are believed to be involved in receptor activation while the C-terminus is primarily responsible for receptor affinity. hCRF-BP appears to prevent hCRF from stimulating corticotrophs by binding the central domain and thus preventing the ligand from interacting with the receptor and causing ACTH release.

Ligand inhibitor of CRF/CRF-binding protein As noted above, the present invention presents a method for increasing the level of free CRF in the brain by administering an effective amount of a ligand inhibitor of a CRF/CRF-BP complex, such that CRF is released from CRF-BP.

The "ligand inhibitor of the CRF/CRF-BP complex" displaces CRF either in a reversible or irreversible fashion. Displacement may occur by causing a bound CRF molecule to become free CRF. In addition, the binding of the ligand inhibitor may inhibit binding of free CRF to CRF-BP because of a high affinity to human CRF-BP, thus competing with endogenous CRF for binding to hCRF-BP.

Reversible or irreversible displacement of CRF may be mediated by the ligand inhibitor binding directly to the CRF binding site or alternatively by the ligand inhibitor binding to a site that is not the CRF binding site and causing allosteric displacement of the bound CRF. Ligand inhibitors may be peptides derived from CRF or CRF-related WO 96/02569 PCT/US95/08867 7 sequences, or random peptides which are designed to cause displacement of bound CRF. Additionally, ligand inhibitors may be non-peptide molecules derived from natural libraries of small molecules, synthesized analogues of natural molecules, specifically designed small molecules based on physical characteristics of the CRF/CRF-BP binding complex, or other ligand inhibitors. Ligand inhibitors may also be metabolites of administered compounds. The ligand inhibitors must be accessible to the brain, either administered through the CNS or systemically. Preferably, the characteristics of the ligand inhibitor are such that it is a low affinity antagonist at the CRF receptor (Ki 1 pM) or has a 100-fold selectivity to the CRF binding protein (Ki 10 nM). The CRF-BP ligand inhibitor may also exhibit some moderate agonist activity at the CRF receptor (Ki 2 50 nM).

Peptide sequences which bind to CRF-BP and displace endogenous CRF may be derived from CRF peptide sequences, CRF-related peptide sequences, or unrelated peptide sequences. It has been found that relatively short peptides between about 19 and 28 residues in length are effective to competitively bind to hCRF-BP while at the same time exhibiting very low binding affinity to CRF receptors. As a result, this particular group of peptides bind to the hCRF-BP while not substantially binding to and/or interacting with the CRF receptor. When given alone, these peptides have the effect of increasing the amount of endogenous free CRF, which would remain available to interact with the CRF receptors. Thus, these peptides can be used to assure or increase the effect of endogenous CRF in a particular region or body organ. These peptides may similarly be used to increase the efficacy of CRF receptor agonists or antagonists if given together with these substances in a cocktail; however, administration with a CRF antagonist would be somewhat counteracted by the release of endogenous CRF. Examples of CRF-related proteins include urotensin and sauvagine. Preferred peptides are those derived from h/rCRF. In this regard, many different peptides may be used within the context of the subject invention to displace endogenously bound CRF. Such peptides include a-helical oCRF (ovine CRF) (9-41) (numbers in parentheses refer to amino acids), h/rCRF h/rCRF (9-33), urotensin I, sauvagine, and h/rCRF. Preferred peptides are h/rCRF h/rCRF and h/rCRF (1-41)OH, as these peptides bind CRF-BP but do not bind CRF receptors with appreciable affinity. The free acid (OH) at the C-terminal end is particularly preferred, rather than the amidation which is found in native CRF.

Deamidation of the C-terminus amino acid of h/rCRF does not affect the binding affinity to CRF-BP, but vastly reduces the affinity of h/rCRF for the CRF receptor.

The N-terminus of these agents can be protected against degradation by acylation, for example with acetyl and such modified peptides are considered equivalents. The WO 96/02569 PCT/US95/08867 8 minimal sequence of CRF which has been found to bind to CRF-BP without significant interaction with the CRF receptor are amino acids 9-33. In particular, amino acid residues 22-25 appear to play a critical role in binding to CRF-BP.

Other peptides may be designed based on homology to the series of peptides described above. As noted above, when designing peptides, it is preferred that the C-terminus amino acid contains a free acid group. Figure 1 presents a comparison of the amino acid sequence of known CRFs and CRF-related molecules. As mentioned above, residues 9-33 contain the minimal sequence needed to bind to CRF-BP, and residues 22, 23, and 25 are believed to play a critical role in binding. A preferred guideline in designing peptides is that the amino acid residue corresponding to residue 22 is alanine, to residue 23 is a basic amino acid (arginine or lysine), and to residue 25 is glutamic acid.

Examples of suitable peptides include fragments of human CRF, i.e., hCRF(1-41), which has the sequence: Ser-Glu-Glu-Pro-Pro-Ile-Ser-Leu-Asp-Leu-Thr-Phe-His-LeuLeu-Arg- Glu-Val-Leu-Glu-Met-Ala-Arg-Ala-Glu-Gln-Leu-Ala-Gln-Gln-Ala-His-Ser-Asn-Arg- Lys-Leu-Met-Glu-Ile-Ile (SEQ ID NO:1).

One preferred fragment which is employed is hCRF(6-33) having the sequence Ile-Ser-Leu-Asp-Leu-ThrPhe-His-Leu-Leu-Arg-Glu-Val-Leu-Glu-Met-Ala- Arg-Ala-GluGln-Leu-Ala-Gln-Gln-Ala-His-Ser (SEQ ID NO:1). This fragment can be shortened by up to 4 residues at the N-terminus and/or up to 5 residues at the Cterminus by elimination of residues in sequence, hCRF hCRF (6-30) and hCRF Other examples of peptides that may alternatively be used include analogues ofhCRF (6-33) such as: [Nle 2 1]-hCRF [Nle 2 1]-hCRF [Ile 24 hCRF [Asn 26 -hCRF [Nlel 8 21 ]-hCRF [Ile 27 ]-hCRF (6-33), [Va 28 ]-hCRF [Asn 29 30 ]-hCRF [Lys16]-hCRF(6-33), [Asp' 7 ]-hCRF (6- 33), [Leul2]-hCRF (6 3 [Arg' 3 ]-hCRF [Glu 9 ]-hCRF (6-33) [Va 3 1]-hCRF [Thr 33 ]-hCRF [Arg 32 ]-hCRF [IleO]-hCRF (6-33) and [Ilel 4 CRF (6-33).

Other peptides which may be employed for this purpose are defined by the following sequence (SEQ ID NO:2): Xaa 4 -Xaa 5 ,-Xaa 6 -Xaa 7 -Xaa 8 -Xaa 9 -Xaalo- Xaal -Xaal 2 -Xaal 3 -Xaal4-Xaa 5 -Xaal6-Xaal7-Xaa 8 -Xaal 9 -Xaa 20 -Xaa 2 1 -Ala-Xaa 23 Xaa 24 -Glu-Xaa 26 -Xaa 27 -Xaa 28 -Xaa 29 -Xaa3o-Xaa 31 -Xaa 32 -Xaa 33 or a biologically active fragment thereof which is formed by the deletion of from 1 to 8 residues in sequence from the N-terminus, or from 1 to 5 residues in sequence from the C-terminus, or both, wherein Xaa 23 is Arg or Lys, Xaa 24 is Ala, Ile, Asn, Met, Nle or Leu, and each remaining Xaa represents the residue present in the respective position in human CRF WO 96/02569 WO 9602569PCTJUS95/08867 9 (1-4 1) or another naturally occurring amino acid, preferably a conservative substitute for the naturally occurring residue.

Another group of preferred peptides are defined by the following sequence (SEQ ID NO:2): Xaa 4 -Xaa 5 -Xaa 6 -Xaa 7 -Xaa8-Xaag..Xaaio-.XaaI I -Xaal 2 Xaa 1 3 -Xaa] 4 -Xaa 1 5 -Xaa 1 6 -Xaal 17 -Xaa, 8 -Xaa, 9 -Xaa 20 -Xaa 2 -Ala-Xaa 23 -Xaa 24 .Glu.

Xaa 26 -Xaa 27 -Xaa 28 -Xaa 29 -Xaa 30 Xaa 3 -Xaa 32 -Xaa 33 and biologically active fragments thereof which are formed by the deletion of from 1 to 8 residues in sequence from the N-terminus, or from 1 to 5 residues in sequence from the C-terminus, or both, wherein Xaa 6 is le, Met, Leu or Nie; Xaa.

8 is Leu or Ile; Xaa 1 4 is Leu, Met or Nie; Xaa, 7 is Glu or Asn; Xaa, 8 is Val, Met, Leu or Nie; Xaa 1 9 is Leu or Ile; Xaa.

20 is Glu or His; Xaa 2 1 is Met, Leu, Nie or Arg; Xaa 23 is Arg or Lys; Xaa.

24 is Ala, Ile, Asn, Met, Nie or Leu; Xaa, 26 is Gln, Asn or Gly; Xaa.

27 is Leu, Glu or Gln; Xaa, 28 is Ala or Arg; Xaa 29 is Gln or Giu; Xaa 32 is His, Glu or Leu; Xaa, 33 is Ser, Leu or Ile; and each remaining Xaa represents the residue present in the respective position in human CRF (1-41) or another natural amino acid which is preferably a conservative substitution therefor.

From 2 to 5 residues are preferably deleted from the N-terminus. Most preferably, Xaa 7 is Ser, Xaa 9 is Asp, Xaal 0 is Leu, Xaal I is Thr, Xaal 2 is Phe, Xaal 3 is His, Xaal 5 is Leu, Xaa1 6 is Arg, Xaa, 30 is Gin and Xaa 3 is Ala.

Another group of preferred peptides are defined by the following sequence (SEQ ID NO:3): Pro-Pro-Ile-Ser-Xaa 8 -AspLeuTrPheHisLeu-Leu-Ag.

Xaa 1 7 -Xaa 1 8 -Xaa 1 9 -Glu-Xaa 21 l-r-a24GUXa6Xa2-a2-aagGnAa Xaa 32 -Xaa 33 or a biologically active fragment thereof which is formed by the deietion of from 1 to 8 residues in sequence from the N-terminus, or from I to 5 residues in sequence from the C-terminus, or both, wherein Xaa.

8 is Leu or Ile; Xaa 17 is Giu or Asn; Xaa 18 is Val, Met, Leu or Nie; Xaa 19 is Leu or le; Xaa, 2 is Met, Leu. or Nie; Xaa 24 is Ala, le or Asn; Xaa.

26 is Gin or Asn; Xaa 27 is Leu, Glu or Gin; Xaa, 28 is Ala or Arg; Xaa 29 is Gln or Glu; Xaa 32 is His or Glu; and Xaa 33 is Ser or Leu.

Of the latter group, particularly preferred peptides, other than the aforementioned fragments of hCRF, include the following analogues: [11e 8 19 ,24, Asn 7 26, Met 18 Glu 27 29,,Arg 28 Gly 32 Leu 33 ]-hCRF(6-33) [Asn1 7 26, Nle] 8 le' 9 24, GIu 27 29, Arg 28 Gly 32 Lu]-hCRF(9-33) [11e 8 19,24, Asn1 7 26, Met' 8 GIu 27 Arg 28 ]-hCRF(4-28) [le 19 24, Asn' 7 26, Nle' 8 Glu 27 Arg 28 ]-hCRF (7-3 1) [le 8 19, Asn 7 24, 26, Met' 8

GLN

27 Arg 28 Glu 29 Giy 32 Leu 33 ]-hCRJ' (6-33) [Asn' 7 24, 26, Met' 8 Ile' 9 Gln 27 Arg 28 GIu 29 Gly 32 Leu 33 ]-hCRP (9-33) [11e 8 19, Asn 17 24, 26, Met' 8 Gln 27 Arg 28 ]-hCRF (8-28) [11e 8 19, Asn 17 24, 26, Nle 18 Gln 27 Arg 28 Glu 29 Gly 32 ]-hCRF (8-32) WO 96/02569 PCTJUS95/08867 [Ile8, 19, Asn'7, 24, 26, Nle' 8 Gln 27 Arg 28 Glu 29 Gly 32 ]-HCRF (8-32) [Met 6 14,18,24, Ile 8 ,19,33, Asn' 7 His 20 Arg 21 ,28, Lys23, Gly 26 Glu 27 29, Leu 32 ]-hCRF (6-33) [Ile19,33, Met4, 18,24 Asn' 7 His 20 Arg 2 1 28, Lys 2 3 Gly 26 Glu 2 7 ,29, Leu 32 ]-hCRF (9-33) [Nle 6 ,14, 18,24, ile8, 19,33, Asn' 7 His20, Arg 21 28, Lys 23 Gly 26 Glu 27 29, Leu 32 ]-hCRF (6-33) [lie, 19, Nle4, 18, 24, Asn' 7 His 20 Arg 21 28, Lys 23 Gly 26 Glu 27 29 ]-hCRF (8-30) Peptide nomenclature may be found in Schroder and Lubke, "The Peptides," Academic Press (1965) wherein, in accordance with conventional representation, the amino group appears to the left and the carboxyl group to the right.

The standard 3-letter abbreviations are used to identify the alpha-amino acid residues, and where the amino acid residue has isomeric forms, it is the L-form of the amino acid that is represented unless otherwise expressly indicated, Ser L-serine, Or Lornithine, Nle L-norleucine, Nva L-norvaline, Har L-homoarginine and CML L-CaMeLeu.

The peptides are synthesized by a suitable method, such as by exclusively solid-phase techniques, by partial solid-phase techniques, by fragment condensation or by classical solution addition. Chemical syntheses of peptides employ protection of the labile side chain groups of the various amino acid moieties with suitable protecting groups to prevent a chemical reaction from occurring at that site until the group is ultimately removed. Most methods protect an alpha-amino group on an amino acid or a fragment while that entity reacts at the carboxyl group, followed by the selective removal of the alpha-amino protecting group to allow subsequent reaction to take place at that location. Examples of such syntheses of representative peptides are provided in U.S. Patent No. 5,235,036, issued August 10, 1993, the disclosure of which is incorporated herein by reference.

Peptides such as hCRF (6-33) and hCRF (9-33) and the other analogues specifically mentioned hereinbefore are synthesized manually using solid phase methodology or automated with a Beckman Model 990 peptide synthesizer, as described in U.S. Patent No. 4,489,163, the disclosure of which is incorporated herein by reference. Briefly, tertbutoxycarbonyl (Boc) is used for a-amino protection, and

TFA-CH

2 Cl 2 is used for deprotection. Standard couplings are mediated by 1,3diisopropylcarbodimide (DIC), while difficult couplings are accomplished using 2-(1Hbenzotriazol-l-yl)-l,1,3, 3 -tetramethyluronium hexafluorophosphate (HBTU). The protected peptide resin is cleaved using anhydrous hydrofluoric acid (HF) in the presence of 3% methyl sulfide, with the HF subsequently being removed in vacuo.

Crude peptides are purified using multiple-step, reversed-phase HPLC.

A polypeptide analogue includes any polypeptide having an amino acid residue sequence substantially identical to a sequence specifically shown herein in WO 96/02569 PCT/US95/08867 11 which one or more residues have been substituted with an amino acid residue in the positions mentioned hereinbefore. Conservative substitutions may be made with a residue having a functionally similar side chain, as long as the polypeptide displays the ability to cause an increase in free CFR, such as by binding strongly to CRF-BP.

General examples of conservative substitutions include the substitution of one nonpolar (hydrophobic) residue, such as isoleucine, valine, alanine, glycine, leucine or methionine, for another; the substitution of one polar (hydrophilic) residue for another, such as arginine for lysine, glutamine for asparagine, threonine for serine; the substitution of one basic residue such as lysine, arginine or histidine for another; and the substitution of one acidic residue, such as aspartic acid or glutamic acid for the other. The phrase "conservative substitution" also includes the use of a chemically derivatized residue in place of a non-derivatized residue provided that such polypeptide displays the desired binding activity. As previously mentioned, such conservative substitutions can be made for one or more of the residues Xaa 6 -Xaa 21 and Xaa 26 -Xaa 33 examples of preferred conservative substitutions are set forth in Table 1.

WO 96/02569 PCTfUS95/08867 12 TABLE 1 Preferred Original Conservative Most Preferred Residue Substitutions Substitution Ala Val; Leu; Ile Val Arg Lys; Gin; Asn Lys Asn Gin; His; Lys; Arg Gin Asp Glu Glu Cys Ser Ser Gin Asn Asn Glu Asp Asp Gly Pro Pro His Asn; Gin; Lys; Arg Arg Ile Leu; Val; Met; Ala; Phe Ile Leu Nle; Ile; Val; Met; Ala; Ile Phe Lys Arg; Gin; Asn Arg Met Leu; Phe; Ile Leu Phe Leu; Val; Ile; Ala Leu Pro Gly Gly Ser Thr Thr Thr Ser Ser Trp Tyr Tyr Tyr Trp;/ Phe; Thr; Ser Phe Val Ile; Leu; Met; Phe; Ala; Leu Nle "Chemical derivative" refers to a subject polypeptide having one or more residues chemically derivatized by reaction of a functional side group. Such derivatized molecules include, for example, those molecules in which free amino groups have been derivatized to form amine hydrochlorides, p-toluene sulfonyl groups, carbobenzoxy groups, t-butyloxycarbonyl groups, chloroacetyl groups or formyl groups. Free carboxyl groups may be derivatized to form salts, methyl and ethyl esters or other types of esters or hydrazides. Free hydroxyl groups may be derivatized to form WO 96/02569 PCTIUS95/08867 13 O-acyl or O-alkyl derivatives. The imidazole nitrogen of histidine may be derivatized to form N-im-benzylhistidine. Chemical derivatives also include those peptides which contain one or more naturally occurring amino acid derivatives of the standard amino acids. For example, 4-hydroxyproline may be substituted for proline, may be substituted for lysine, 3-methylhistidine may be substituted for histidine, homoserine may be substituted for serine, and ornithine may be substituted for lysine.

In addition, random peptides may be synthesized and subsequently screened to identify peptides that meet the functional criteria of displacement of CRF from the CRF/CRF-BP complex, discussed in more detail below. Random peptides may be generated by biological methods, or by combinatorial chemical technologies (see Gallop et al., J. Med. Chem. 37:1234, 1994).

Biological methods of generating random peptides include at least four different methods. First, random nucleotides may be inserted into a host gene for display of the peptide on the surface of microorganisms (Charbit et al., Embo. Journal 5:3029, 1986; Agterberg et al., Gene 88:37, 1990; Fuchs et al., Bio/Tech 9:1369, 1991; Thery et al., Appl. Environ. Microbiol. 55:984, 1989). In such methods, various cell surface proteins of bacteria are used as fusion partners in which oligonucleotides are inserted to produce peptides fused into one of the extracellular loops of the protein.

LamB, OmpA, PhoE, PAL, and pilin proteins have all served as vehicles for peptide display on bacteria. An alternative system, which presents peptides on the surface of bacteriophage, is a preferred method. The vectors employed are derived from filamentous phages, such as M13, fl, and fd. Typically, coat proteins, such as pill or pVIII, serve as peptide expression vehicles. (Smith, Science 228:1315, 1985; Parmley and Smith, Gene 73:305, 1988; Cwirlaetal., Proc. Natl. Acad Sci. USA 87:6378, 1990; Markland et al., Gene 109:13, 1991, U.S. Patent No. 5,223,409). With both phage display and bacterial display methods, there is a physical link between the peptide and the DNA encoding peptide which allows for easy isolation and propagation of the DNA sequence. Third, peptides may be attached to plasmids (Cull et al., Proc.

Natl. Acad. Sci. USA 89:1865, 1992) by fusing the peptide to the DNA binding protein LacI at its C-terminus. The fusion protein then binds to a Lac operator on the plasmid, such that the peptides become specifically and stably associated with the DNA sequence encoding them. Fourth, peptides may be displayed on polysomes after stalling translation. By causing accumulation of RNA and polysomes containing nascent peptides still linked to their encoding RNA (Gallop et al., J. Medicinal Chem.

37:1233, 1994), the genetic material encoding the peptide may be recovered. In all these methods, libraries of variants based on CRF as a lead peptide may be synthesized by controlling the proportional level of "incorrect" bases which are inserted.

WO 96/02569 PCT/US95/08867 14 In addition to biological methods, approaches based on combinatorial chemical technologies may be utilized to generate random peptides. A variety of methods have been developed to synthesize multiple, homogeneous peptides which are available for assay. These methods include the multi-pin synthesis (Geysen et al., Proc.

Natl. Acad. Sci. USA 81:3998, 1984; Valerio, Anal. Biochem. 197:168, 1991; Bray etal., Tetrahedron Lett. 32:6163, 1991) and the "tea bag" method (Houghten et al., Proc. Natl. Acad. Sci. USA 82:5131, 1985; Houghten et al., Int. J.

Pept. Protein Res. 27:673, 1986). Multiple, degenerate peptides may also be synthesized for use. Simultaneous coupling of mixtures of amino acids to a single resin support is a common strategy for synthesizing multiple peptides. More recently, combinatorial libraries of soluble peptides (Houghtenetal., Nature 354:84, 1991; Houghten et al., Biotechniques 13:412, 1992) and peptides tethered to solid supports (Lam et al., Nature 354:82, 1991) have been synthesized by a "split synthesis" method.

Many variations on these syntheses have been developed (see Gallop et al., supra, for review), including synthesis of peptides on beads that contain identifiers for the peptide structure (PCT WO 93/06121). One such identifier is an oligonucleotide sequence (Needels et al., Proc. Natl. Acad. Sci. USA 90:10700, 1993). Other syntheses of random peptides are well known and may be readily performed by one skilled in the art.

Small, non-peptide molecules, natural or synthesized, may also be used as ligand inhibitors. Libraries of small molecules to be screened for ligand inhibitors of the CRF/CRF-BP complex are obtained from soil samples, plant extracts, marine microorganisms, fermentation broth, fungal broth, pharmaceutical chemical libraries, combinatorial libraries (both chemical and biological) and the like. Such libraries may be obtained from a variety of sources, both commercial and proprietary.

Molecules, which have a similar, but not identical, structure to the candidate compound may be synthesized and tested for ligand inhibition as described below. A small peptide sequence can be modeled around the solution NMR structure of CRF. As previously mentioned, key contact residues in CRF for binding to CRF-BP are residues 22, 23, and 25. Thus, a peptide sequence derived from residues 20-27 can be modeled from a reference solution structure of human CRF, which has been determined by 1H NMR and distance geometry with restrained molecular dynamics (Protein Engineering 6:149, 1993). The intact CRF molecule is used as a basis for modeling because it is possible that the alpha helical structure of CRF is lost as bigger deletions are made. The 3-D structure of the small peptide, which spans the important contact residues, can then be used to search a computer bank for non-peptide molecules which mimic the peptide's structure. Molecules with a lower IC-50 value than the starting molecule are selected. Further molecules are synthesized based upon the WO 96/02569 PCT/US95/08867 physical and biochemical characteristics of the initial compounds and their molecules.

A particularly preferred candidate will have an IC-50 value of 10 nM or less.

The determination of whether any of the inhibitors discussed above will have the requisite properties can be made by assaying the ability of the inhibitor to displace CRF from CRF/CRF-BP binding complex and secondly, by evaluating its ability to bind the CRF receptor.

Candidate ligand inhibitors may be screened for their ability to displace CRF from the CRF/CRF-BP complex by biological assay or by in vitro assay. One suitable biological assay is the measurement of ACTH release from cultured pituitary cells. This assay is performed in the following manner. Anterior pituitary glands from rats are washed six times with sterile HEPES buffer and transferred to a solution containing collagenase. After subsequent transfer to a 25 ml Bellco dispersion flask, the pituitaries are stirred for 30 min at 37 0 C, triturated, incubated for a further 30 min, and again triturated. The partially dispersed cells are then collected by centrifugation.

The cell pellet is resuspended in 10 ml of neuraminidase and again collected by centrifugation. The pellet is reconstituted in 25 ml of BBM-P (BBM (Irvine Scientific) plus 100gg/L, cortisol, ltg/L insulin, O.1gg/L EGF 2 0.4pg/L T 3 0.7gg/L PTH, g/L glucagon, and 2% fetal bovine serum), centrifuged again, and the resultant pellet is finally reconstituted in BBM-P. The cells are then plated at a density of 50,000-100,000/well in a 48 well plate and incubated for 2 days. On the day of assay the cells are washed once with BBM-T in preparation for stimulation with the peptide candidates or CRF. Cells are stimulated with a maximally stimulating dose of h/rCRF (1 nM) and ACTH release is measured by RIA or immunoradiometric assay. When a blocking concentration of CRF-BP is added, the amount of ACTH that is released is reduced and expressed as a fraction of the maximal release. The ligand inhibitor is added at various doses. The potency of the peptide in displacing CRF from CRF-BP is measured by the amount of ACTH release expressed as a fraction of the maximal release caused by CRF given alone.

A preferred mode of screening candidate ligand inhibitors is by an in vitro ligand immunoradiometric assay (LIRMA). For LIRMA, CRF-BP may be isolated from brain tissue, serum or cells expressing a recombinant form. Recombinant hCRF-BP may be produced in Chinese Hamster Ovary (CHO) cells bearing the hA3 and RSV-neo plasmids. Stable CHO transfectants are cloned by dilution under G418 (Sigma Chemical, St. Louis, MO) selection and maintained in Dulbecco's Modified Eagle Medium supplemented with 2 mM L-glutamine and 3% fetal bovine serum. In order to scale up production of hCRF-BP, transfected CHO cells are inoculated into a 10,000 MWCO bioreactor (Cell Pharm Micro Mouse, Unisyn WO 96/02569 PCT/US95/08867 16 Technologies, Tustin, CA). Enriched medium is harvested from the bioreactor daily and stored at -20°C until purification. Closed roller bottles containing recombinant cells and tissue culture medium, which are slowly rotated in a 37 0 C environment, may alternatively be used.

hCRF-BP may be purified by a 3-step process, with fractions from each step being evaluated using the assay described below. First, enriched medium is affinity-purified using a Bio Pilot chromatography device (Pharmacia

LKB

Biotechnology, Uppsala, Sweden). Human CRF is coupled to Affi-Prep 10 (Bio Rad Laboratories, Richmond, CA) via primary amino groups using N-hydroxysuccinimide.

After coupling, the affinity gel is packed into an XK16 or equivalent column (Pharmacia LKB Biotechnology, Uppsala, Sweden). Affinity purification consists of percolating enriched medium through the column at 2 ml/min, washing with 10 bed volumes of 100 mM HEPES HCI (pH 7.5) at 5 ml/min and eluting 1 bed volume fractions using 80 mM triethylammonium formate (pH 3.0) containing 20% acetonitrile at 5 ml/min. Elution under mildly basic conditions, pH about 10.5, may alternatively be used.

Secondary purification utilizes gel chromatography. Affinity-pure hCRF-BP is lyophilized and reconstituted in 6M guanidine-HCl buffered with 0.1M ammonium acetate (pH 4.75). An FPLC device is used in conjunction with two Superose 12 HR 10/30 columns (Pharmacia LKB Biotechnology, Uppsala, Sweden) connected in series for this purification step. The affinity-pure hCRF-BP is loaded in 1 ml and subsequently eluted with 6M guanidine HCI/0.1M ammonium acetate (pH 4.75) at 0.4 ml/min, collecting fractions every minute.

Active fractions from the secondary purification are then subjected to reversed-phase HPLC. The HPLC device consists of 2 model 100A pumps (Beckman, Palo Alto, CA), an Axxiom HPLC controller (Cole Scientific, Calabasas, CA), a Spectroflow 773 absorbance detector set to 214 nm (Kratos Analytical, Ramsey,

NJ),

and a Pharmacia. model 482 chart recorder (Pharmacia LKB Biotechnology, Uppsala, Sweden). Buffer A is 0.1% trifluoroacetic acid (TFA)/5% acetonitrile, and buffer B is 0.1% TFA/80% acetonitrile. Sequential 1 ml injections of affinity-pure, sized hCRF- BP are applied to a semipreparative C4 HPLC column (Vydac, Hesperia, CA) under isocratic conditions with a flow rate of 2.5 ml/minute in 25% B buffer. After the passage of the final salt peak, a single gradient elution is performed starting at 25% B buffer and increasing to 95% B buffer over 30 minutes. The predominant absorbance peak is then quantitated by hCRF-BP IRMA and by amino acid analysis.

In LIRMA CRF-BP isolated from brain tissue, serum, or cells expressing a recombinant form, is added to wells of a 9 6 -well plate, to small WO 96/02569 PCT/US95/08867 17 polypropylene microfuge tubes, or to glass borosilicate tubes in a binding buffer (0.02% NP-40 in 50 mM phosphate-buffered saline). 12 5 I-h/rCRF (New England Nuclear) and the candidate ligand inhibitor at 10 pM are added and the reaction is incubated for one hour at room temperature. An appropriately diluted anti-CRF-BP antibody, such as a rabbit anti-hCRF-BP (Potter et al., Proc. Natl. Acad. Sci. USA 89:4192-4196, 1992) is added to each tube, and after further incubation, bound complexes are precipitated by the further addition of a goat anti-rabbit antibody. The precipitate containing 1 2 5 I-CRF is collected by centrifugation and the amount of radioactivity in the pellet is determined by a gamma counter. If the candidate ligand inhibitor displaces CRF from CRF-BP, the pellets will contain less radioactivity in comparison to controls in which no candidate peptide is added. Maximum inhibition 100%) of the binding of 125 I-h/rCRF to the CRF-BP is defined by the amount of radioactivity left in the pellets after incubation with 10 gM of the CRF-BP peptide ligand h/rCRF Thus, the binding potency of the candidate ligand inhibitor will be measured relative to the potency of the standard h/rCRF Preferably, there is at least 50% inhibition when ligand inhibitor is present.

The inhibitory binding affinity constant is important; it is viewed in proper perspective as per its value relative to the K, for human CRF which, from this assay, is found to be 0.17 0.01 nanomolar Thus, a ligand having a K, of less than that of hCRF will bind more strongly to hCRF-BP than will hCRF itself, and a ligand having a higher value will have a relatively lower binding affinity. Therefore, because the desire is to compete reasonably effectively with hCRF for binding with hCRF-BP, the lower K i value the agent or peptide has, the more valuable it will be for this purpose. Preferably, the agent will have a K i value of about 20 nM or less, more preferably a K i value of about 10 nM or less, and most preferably, a K i value of less than about 5 nM. hCRF (6-33) has been assayed and found to have a K i value of 3.5 0.44 nM. Because this agent also has a low binding affinity for the human CRF receptor, an inhibitory binding constant of greater than 1000 nM, and exhibits a CRF agonism of less than about 0.1% of oCRF, it is considered an excellent choice for employment in the method of the present invention. Human CRF (9-33) has a K i of 11 0.36 nM, and it also has a receptor K i of greater than 1000 nM and is an even weaker CRF agonist, rendering it also very useful in the method of the present invention. Other similar agents and peptides, particularly those which are analogues of hCRF having between 19 and 28 residues may be synthesized and tested in this straightforward manner to determine their usefulness in these valuable methods for increasing the effective concentration of endogenous hCRF in vivo. The peptides specifically enumerated herein are felt to be particularly valuable. For example (Ile 8 19,24, Asn'7, 26 WO 96/02569 PCTIUS95/08867 18 Met 18 Glu 27 Arg 28 ]-hCRF(4-28) has a K, of 1.7 1.2 and relatively low binding affinity for the hCRF receptor.

Thus, a high through-put screening using the LIRMA assay as described, or other methods including ACTH release and 2-site ELISA, may be used to identify small molecules that displace CRF from the CRF/CRF-BP complex. In a first round of screening, all potential candidates are assayed at a single dose of 10 IM. Any compound which gives greater than 50% inhibition at 10 uM is then selected for further screening. The activity of all candidates meeting this criteria is confirmed by a second round of screening using a 6 point-dose response curve. IC-50 values are calculated and those candidates with a value in the range of 10-100 tM are further examined to ensure that the candidate compound is displacing CRF from the CRF/CRF-BP complex and not interfering with antibody binding to the CRF-BP. Specific displacement of CRF is verified in an assay performed as described for LIRMA, except that 0.2 nM 125 I-hCRF-BP is added in place of unlabeled

CRF-BP.

Ligand inhibitors may also be screened by an in vitro assay in which bound and free CRF are separated by detergent phase separation. Briefly, within one embodiment, CRF-BP isolated as described above is incubated with 1 25 I-h/rCRF and the candidate ligand inhibitor at 10pM in a binding buffer (0.02% NP-40 in 50 mM phosphate-buffered saline). Following incubation of 1-2 hours at room temperature, a detergent, such as octylphenoxypolyethoxyethanol, sold as Triton X-114TM, is added and mixed by vortexing. Triton X-114TM and other nonionic detergents are insoluble in water above their cloud point temperature. At this temperature, there occurs a microscopic phase separation. Below this temperature, the detergents form clear micellar solutions and above this temperature, two clear phases, one depleted and one enriched in detergent, are formed. The cloud point temperature of Triton X-114TM is 0 C. As such, Triton X-114TM is preferred. CRF, which has amphiphilic alpha helices, is more soluble in Triton X-114TM and thus partitions to the detergent phase. In contrast, CRF bound to CRF-BP is more soluble in an aqueous solution. Thus, a phase separation of Triton X-114TM and the aqueous solution will segregate bound and free CRF. Phase separation is conveniently accomplished by centrifugation. The aqueous phase (on top) may be removed and the amount of 1 25 I-h/rCRF determined.

A

reduction of radioactivity relative to that obtained in the absence of ligand inhibitor means that the ligand inhibitor displaced CRF from CRF-BP. Maximum inhibition 100%) of the binding of 12 5 I-h/rCRF to the CRF-BP is defined by the amount of radioactivity in the aqueous phase after incubation with 10 pM of the CRF-BP peptide ligand h/rCRF Thus, the binding potency of the candidate ligand inhibitor will WO 96/02569 PCT/US95/0886 7 19 be measured relative to the potency of the standard h/rCRF Preferably, there is at least 50% inhibition when ligand inhibitor is present.

In addition, this assay has broad application in screening for neuropeptide binding proteins in general. Some neuropeptides, such as NPY, have similar physical characteristics to CRF in that they are both very hydrophobic and have alpha helices. As such, NPY should be more soluble in a nonionic detergent, such as Triton X-114TM, than in aqueous solutions. Given that NPY or other neuropeptides of interest will generally partition into the Triton X-114TM detergent phase, the method described above may be generally employed to screen for neuropeptide-binding proteins. Briefly, by way of example, tissue from various organs is homogenized in 1% solubilization buffer. Particulate matter is removed by centrifugation for minutes at 3000 x g. A 50 tl aliquot from the supematant is incubated with 500 pM of the 1 2 5 I-labeled neuropeptide and the assay is performed as above. Serum or plasma may also be used as a potential source of neuropeptide-binding proteins. A range of concentrations (0.1-1000 nM) of unlabeled neuropeptide is coincubated with the radiolabeled neuropeptide to assess whether the putative binding protein specifically binds the radiolabeled neuropeptide. When binding is specific, the radioactivity remaining in the aqueous phase after Triton X- 14TM separation is decreased. Using this method, an IC-50 value can be established for each neuropeptide and tissue extract.

In addition, this method may be employed to screen for ligand inhibitors of the neuropeptide to its neuropeptide-binding protein. Briefly, radiolabeled neuropeptide is incubated with the neuropeptide-binding protein or soluble receptor and the reaction performed as described above. For these assays, either recombinant neuropeptide-binding protein or receptor or crude neuropeptide-binding protein isolated from tissue sample may be used.

A preferred ligand inhibitor either has a low affinity antagonist effect at the CRF receptor or has a 100-fold selectivity to the CRF binding protein. Therefore, compounds with an IC-50 value in the range of 10-100 jM and a specific inhibition of the CRF/CRF-BP complex are further tested for binding to the CRF receptors. The ability of the ligand inhibitor to antagonize the CRF receptor is assessed in a cAMP production assay. The assay compares the potency of the ligand inhibitor to increase levels of free CRF which thereby bind the CRF receptor, and stimulate cAMP production. The test cell lines express the CRF receptor as stable transfectants. The assay is performed according to Battaglia et al. (Synapse 1:572, 1987) with minor modifications. Test cells are incubated for 1 hr with various concentrations of CRF and ligand inhibitors. The cells are washed, and intracellular cAMP is released upon incubation of the cells for 16-18 hrs and is subsequently extracted in 20 mM HC1, WO 96/02569 PCT/US95/08867 ethanol. The lysate is lyophilized and subsequently solubilized in a sodium acetate buffer. The levels of cAMP are measured using a single antibody kit, such as the one from Biomedical Technologies (Stoughton,

MA).

As an alternative to carrying out the foregoing competitive in vitro evaluation assays, the ligand inhibitor can be evaluated in a binding assay with the human CRF receptor. The human CRF receptor and a binding assay for such receptor and human CRF are described in Chen et al., Proc. Natl. Acad. Sci. USA 90:8967-8971, 1993, the disclosure of which is incorporated herein by reference. The agent may be evaluated with radioactively labeled [Nle21, Tyr 32 oCRF to compute an inhibitory binding affinity constant Preferably the agent has a receptor

K

i of at least about 100 nM and more preferably greater than 1000 nM. It may alternatively be satisfactory to use the rat CRF receptor because human CRF and rat CRF have the identical amino acid sequence.

An additional assay using rat anterior pituitary cells to measure

ACTH

secretion can be carried out to determine whether a ligand inhibitor functions as a CRF agonist of hCRF receptors. The procedure which is used is that as generally set forth above except that only ligand inhibitor is added to the cells. Antagonistic action may be determined by performing the assay in the presence of a challenge dose of CRF.

The performance of the ligand inhibitor is compared to the performance of what has become a standard antagonist for this purpose, such as CD-Phe'2, Nle 2 1, 38 ]-rCRF(12- 41) or a fragment of alphahelical CRF(AHC), such as AHC (9-41).

The above-identified in vitro assays to measure CRF agonist and antagonist activity from the standpoint of stimulation of ACTH secretion may be performed using hCRF (6-33) and hCRF As a result of such assays, hCRF (6-33) is shown to have a CRF agonist bioactivity much less than the standard oCRF, which is arbitrarily considered as 1.0. This peptide does not exhibit substantial

CRF

antagonist activity. Because this peptide has less than about 0.1% of the CRF agonist activity of the standard peptide, it is acceptable from this standpoint. The peptide hCRF (9-33) is even a weaker CRF agonist, having substantially less than 0.01% of the activity of oCRF. The desire is that the agent which is employed will not bind strongly to CRF receptors. It is generally believed that an agent should have less than about of the CRF agonist activity of oCRF and that it should not exhibit substantial

CRF

competitive antagonist activity. Preferably, it should have less than 5% of the antagonist activity of the present standard peptide [DPhel2, Nle 2 1, 38 ]-hCRF(12-41).

However, it should be understood that the lower its value in such an assay, the better it should function in this method because its potential blocking effect as a result of binding to CRF receptors will be minimized.

WO 96/02569 PCT/US95/08867 21 In addition to providing methods of therapeutic treatment, the invention also provides methods for screening peptides or other agents to select more effective CRF antagonists for in vivo administration to mammals. To carry out this screening procedure, a candidate peptide is first evaluated in the well-known assay described hereinbefore for determining its biological effectiveness to inhibit a test dosage of CRF from stimulating the secretion of ACTH from a culture of rat interior pituitary cells.

The candidate peptide is then evaluated in the hCRF-BP competitive binding assay described hereinbefore in order to determine its K i which, as explained hereinbefore, is indicative of its affinity for binding to hCRF-BP, which has the tendency to clear the injected peptide from the target cell sites to which it is directed. Based upon the evaluation of the results of these two assays, a particularly effective CRF antagonist can be chosen which has a high value in the CRF antagonism assay and which also has a high (CRF-BP) Ki, indicating that it exhibits a relatively low affinity for binding to hCRF-BP. The present laboratory standard, CRF antagonist [D-Phel 2 Nle 2 1, 38]_ hCRF(12-41) has very good antagonist properties as measured by ACTH secretion from cultured pituitary cells, with a K i value of about 60 10 nM. Its (CRF-BP)Ki is 300 20 nM. Another good CRF antagonist, namely AHC(9-41), which is not as effective as the present standard, has an extremely low K, of 0.10 0.036 nM. [D-Phe' 2 Nle 2138

CML

37 ]-hCRF(12-41) has a (CRF-BP)

K

i of greater than 1000 nM and an IC 50 of 45 11 nM in an ACTH secretion assay performed in pituitary cell cultures, so it ranks even higher than the present laboratory standard. Therefore, based upon these screening assays, the latter peptide or the laboratory standard would be the peptides of choice, compared to AHC(9-41), which would have a far greater propensity in vivo to be complexed and cleared by hCRF-BP. This screening assay thus provides a valuable tool for screening newly synthesized peptides to evaluate their overall relative worth as potential CRF antagonists for in vivo treatment.

Increasing the level of CRF The present invention provides methods for increasing the level of free CRF in the brain through the administration of a ligand inhibitor of a CRF/CRF-BP complex. The increase in level of free CRF may be measured by in vitro assays, such as ELISA, stimulation of ACTH release, or stimulation of cAMP production. In any of these assays, an increase in free CRF due to administration of the ligand inhibitor is measured relative to a reference ligand inhibitor, in this case h/rCRF A minimal acceptable value of increase is 10% of the value for h/rCRF a moderate value is a preferred value is 80%, and a particularly preferred value is 100%.

WO 96/02569 PCT/US95/08867 22 Within the methods described herein, the level of free CRF may be measured by two-site ELISA on homogenates of brain samples, cerebrospinal fluid, or on other bodily tissues and fluids. Total CRF is first quantitated as follows. Wells of an ELISA plate are coated with an anti-CRF antibody, such as protein G purified-sheep anti-CRF. The plates are washed, and unbound sites on the plate are blocked with an irrelevant protein, such as casein, bovine serum albumin, or the like. Prepared tissue samples and standards containing known amounts of CRF are added to wells and allowed to bind at room temperature. Following binding, plates are again washed, and a different anti-CRF antibody, such as RC-70, a rabbit anti-human CRF antibody, is added. RC-70 is not only from a different species, but also detects different epitopes than the sheep anti-CRF used to coat the plates. After washing, an enzyme-conjugated antibody that detects RC-70, or the equivalent, is added. Alternatively, RC-70 antibody may be enzyme-conjugated. Preferred conjugates are horseradish peroxidase and alkaline phosphatase, but one skilled in the art will recognize that many different acceptable alternatives are available, including a radiolabel instead of an enzyme.

Enzyme substrate is added, and color development proceeds. After termination of the reaction, absorbance measurements are used to quantify the amount of total CRF present in the tissue sample. One skilled in the art will recognize that monoclonal antibodies or antibody fragments may be used in place of the polyclonal antibodies in this assay.

In a similar manner, bound CRF may be quantitated by ELISA by coating plates with an anti-CRF-BP antibody followed by detection of the bound CRF with an anti-CRF antibody. Bound CRF can be specifically displaced by the CRF-BP ligand a-helical oCRF(9-41) resulting in a decrease in the signal detected. Alpha helical oCRF(9-41) is used for the displacement as it does not crossreact with anti-CRF antibody. Furthermore, the displaced CRF present in the supernatants may then be assayed by two-site ELISA, as described. Free CRF may then be determined by calculation of the difference between total CRF and bound CRF or by a direct assay.

SIn a direct assay, following capture of the bound complex by the anti-CRF-BP monoclonal antibody, the supernatants are removed and the free CRF measured in a two-site ELISA, as described.

A ligand inhibitor, a-helical ovine CRF, is shown herein to increase free CRF levels in brain tissue of both normal individuals and Alzheimer's disease patients.

A two-site ELISA was used to determine the amount of total and bound CRF. Addition of a-helical ovine CRF resulted in release of all bound CRF (see Figure 2 and Example WO 96/02569 PCT/US95/08867 23 In addition to the described in vitro assays that measure the amount of total, bound, and free CRF in tissues or in cerebrospinal fluid, other procedures may be performed in vivo. These include MRI, PETSCAN, spectscanning or other similar imaging techniques, some of which use a radiolabeled ligand to CRF-BP or to CRF receptors. A preferred method is image analysis using PET position-emitting ligands llC, 18 F) of single photon-emitting ligands 123 I-labeled ligand to CRF-BP or to CRF receptors). Free CRF levels are correlated to the amount of binding of the radiolabeled ligand. An increase in free CRF levels is manifested by a decreased binding of the radiolabeled ligand to the CRF-BP and CRF receptors. Within this imaging technique, an increase in free CRF levels of about 10%-30% or more would be sufficient in the context of the present invention.

Within the context of the present invention, administration of effective amounts of ligand inhibitor of the CRF/CRF-BP complex may be used to treat diseases or syndromes in which there are decreased levels of CRF. CRF levels may be measured directly in cerebrospinal fluid or in the brain by imaging or other methods cAMP production, ACTH release, or two-site ELISA). Such diseases or syndromes include symptoms of dementia or learning and memory loss, obesity, chronic fatigue syndrome, atypical depression, post-partum depression, seasonal depression, hypothyroidism, post-traumatic stress syndrome, nicotine withdrawal, vulnerability to inflammatory disease. Definitions of these syndromes (except for obesity, chronic fatigue syndrome, and vulnerability to inflammatory diseases) are provided in Diagnosis and Statistical Manual of Mental Disorders (4th American Psychiatric Association, Washington, 1994 (hereinafter DSM-IV).

Improving Learning and Memory As noted above, the present invention provides methods for improving learning and memory through the administration to a patient of a therapeutically effective amount of a ligand inhibitor of a CRF/CRF-BP complex. Such patients may be identified through a clinical diagnosis based on symptoms of dementia or learning and memory loss. Individuals with an amnestic disorder are impaired in their ability to learn new information or are unable to recall previously learned information or past events. The memory deficit is most apparent on tasks to require spontaneous recall and may also be evident when the examiner provides stimuli for the person to recall at a later time. The memory disturbance must be sufficiently severe to cause marked impairment in social or occupational functioning and must represent a significant decline from a previous level of functioning.

WO 96/02569 PCT/US95/08867 24 Dementia is characterized by multiple clinically significant deficits in cognition that represent a significant change from a previous level of functioning.

Memory impairment involving inability to learn new material or forgetting of previously learned material is required to make the diagnosis of a dementia. Memory can be formally tested by asking the person to register, retain, recall and recognize information. The diagnosis of dementia also requires at least one of the following cognitive disturbances: aphasia, apraxia, agnosia or a disturbance in executive functioning. These deficits in language, motor performance, object recognition and abstract thinking, respectively, must be sufficiently severe in conjunction with the memory deficit to cause impairment in occupational or social functioning and must represent a decline from a previously higher level of functioning.

In addition, a number of biochemical tests that correlate levels of CRF with impaired learning and memory may be utilized. For instance, the level of free CRF in the cerebrospinal fluid may be measured by ELISA or RIA. Additionally, or in place of the assays, brain imaging as described with a labeled ligand specific to the CRF-BP or CRF receptor may be used to quantitate free receptor or CRF-BP, thus allowing one to know that free CRF is decreased. Finally, imaging of the brain with a ligand specific to unbound CRF may be used to directly assay the amount of free CRF in the brain.

The patient's minimental status is recorded by the Minimental Test for Learning and Memory, a standard test used by clinicians to determine if a patient has impaired learning and memory (Folstein et al., J Psychiatric Res. 12:185, 1975). This test involves a number of simple tasks and written questions. For instance, "pairedassociate" learning ability is impaired in amnesiac patients of several types including those suffering from head trauma, Korsakoffs disease or stroke (Squire, 1987). Ten pairs of unrelated words army-table) are read to the subject. Subjects are then asked to recall the second word when given the first word of each pair. The measure of memory impairment is a reduced number of paired-associate words recalled relative to a matched control group. This serves as an index of short-term, working memory of the kind that deteriorates rapidly in the early stages of dementing or amnesiac disorders.

Improvement in learning and memory constitutes either a statistically significant difference between the performance of ligand-inhibitor treated patients as compared to members of a placebo group; or a statistically significant change in performance in the direction of normality on measures pertinent to the disease model.

This strategy has been successfully employed in identifying therapeutically useful cholinomimetics for memory improvement. Animal models or clinical instances of disease exhibit symptoms which are by definition distinguishable from normal controls.

WO 96/02569 PCT/US95/08867 Thus, the measure of effective pharmacotherapy will be a significant, but not necessarily complete, reversal of symptoms. Improvement can be facilitated in both animal and human models of memory pathology by clinically effective "cognitive enhancing" drugs which serve to improve performance of a memory task. For example, cognitive enhancers which function as cholinomimetic replacement therapies in patients suffering from dementia and memory loss of the Alzheimer's type significantly improve short-term working memory in such paradigms as the paired-associate task (Davidson and Stem, 1991). Another potential application for therapeutic interventions against memory impairment is suggested by age-related deficits in performance which are effectively modeled by the longitudinal study of recent memory in aging mice (Forster and Lal, 1992).

In animals, several established models of learning and memory are available to examine the beneficial cognitive enhancing effects and potential anxietyrelated side effects of activation of CRF-sensitive neurons. The cognitive enhancing effects are measured by the Morris maze (Stewart and Morris, in Behavioral Neuroscience, R. Saghal, Ed. (IRL Press, 1993) p. 107) and the Y-maze (Brits et al., Brain Res. Bull. 6, 71 (1981)) tests; anxiety-related effects are evaluated in the elevated plus-maze. (Pellow et al., J. Neurosci. Meth. 14:149, 1985.) The Morris water maze is one of the best validated models of learning and memory, and it is sensitive to the cognitive enhancing effects of a variety of pharmacological agents (McNamara and Skelton, Brain Res. Rev. 18:33, 1993). The task performed in the maze is particularly sensitive to manipulations of the hippocampus in the brain, an area of the brain important for spatial learning in animals and memory consolidation in humans. Moreover, improvement in Morris water maze performance is predictive of clinical efficacy of a compound as a cognitive enhancer.

For example, treatment with cholinesterase inhibitors or selective muscarinic cholinergic agonists reverse learning deficits in the Morris maze animal model of learning and memory, as well as in clinical populations with dementia (McNamara and Skelton, 1993; Davidson and Stem, 1991; McEntee and Crook, 1992; Dawson et al., 1992). In addition, this animal paradigm accurately models the increasing degree of impairment with advancing age (Levy et al., 1994) and the increased vulnerability of the memory trace to pre-test delay or interference (Stewart and Morris, 1993) which is characteristic of amnesiac patients.

The test is a simple spatial learning task in which the animal is placed in tepid water, which is opaque due to the addition of powdered milk. The animals learn the location of the platform relative to visual cues located within the maze and the testing room; this learning is referred to as place learning.

WO 96/02569 PCT/US95/08867 26 As discussed in more detail below (see Example 15 minutes prior to training on each of days 1-3, groups of animals receive ICV injections of control solution or 0.1, 1, 5, or 25 pg of the ligand inhibitor peptide h/rCRF (6-33) or h/rCRF, which is additionally an agonist at the CRF receptor. When a non-peptide inhibitor is used, amounts injected are adjusted accordingly. Control animals typically reach the platform within five to ten seconds after three days of training. The measure of the memory modulator effects of a ligand inhibitor is a shift of this time period.

Administration of a ligand inhibitor results in a dose-dependent increase in availability of synaptic CRF and a behavioral dose-dependent increase in acquisition and memory retention. Daily pre-test administration of h/rCRF and h/rCRF (6-33) significantly enhanced learning in the Morris water maze test (Figure 4, upper panel). Somewhat higher doses of CRF (6-33) were necessary to produce increases in learning and memory.

The Y-maze test based on visual discrimination is another assay of learning and memory in animals. In this maze, two arms of the maze end in a translucent plastic panel behind which there is a 40-watt electric bulb. The start box is separated from the third arm by a manually-activated guillotine door. In the first trial, all animals are allowed to explore the maze for five minutes, and food pellets are available in each arm. On the second day, each animal is placed in the start box with the door closed. When the door is opened, the animal is allowed to move down the arms and eat the pellets which are located in both arms. On the third day, animals receive six trials in groups of three where one arm is closed at the choice point, no discriminative stimulus is present, and two food pellets are available in the open goal box. On days 4-10, a light at the end of the arm with the food pellets is illuminated and ten trials are run, again in groups of three. The time it takes for the animal to reach the food pellets is recorded.

The effectiveness of a ligand inhibitor to improve learning and memory in the Y-maze is tested as follows. Fifteen minutes prior to each of the blocks of training trials on days 4-10, groups of animals receive ICV injections of control solutions or doses of 1, 5, or 25 gg of peptide ligand inhibitor. Again, if a non-peptide ligand inhibitor is used, dosages are adjusted accordingly. Control animals are expected to make 50% correct choices. The measure of efficacy of treatment on memory- is an increase in correct responses. Daily pre-test administration of CRF (6-33) ligand inhibitor was shown to significantly increase correct responses (Figure The elevated plus maze test measures anxiogenic responses in an approach-avoidance situation involving an exposed, lighted space versus a dark, enclosed space. Both spaces are elevated and are set up as two runways intersecting in WO 96/02569 PCTfUS95/08867 27 the form of a plus sign. This type of approach-avoidance situation is a classical test of "emotionality" and is very sensitive to treatments that produce disinhibition and stress.

Animals are placed in the center of the maze and are allowed free access to all four arms in a five minute testing period. The time spent in each arm is recorded.

Daily pre-test administration of doses of h/rCRF by ICV injection that produced increases in learning and memory also produced anxiety as evidenced by the results of the elevated plus maze test (Figure 4, lower panel). A dose-dependent suppression of exploration was observed. Thus, the usefulness of a CRF-receptor agonist K i 1 nM] for treatment of memory and learning deficits that are due to decreased levels of CRF is of dubious value because of the associated side effects. In marked contrast, the ligand inhibitor, CRF which has low affinity for the CRF receptor, did not alter performance in the elevated plus maze test of anxiety (Figure 4, lower panel). Moreover, there were no overt behavioral alterations such as those seen with administration of h/rCRF. These data demonstrate that cognitive enhancement and anxiety effects may be separately controlled. These data also demonstrate the therapeutic value of administering a ligand inhibitor of the CRF/CRF-BP complex.

In humans, methods for improving learning and memory may be measured by such tests as the Wechsler Memory Scale or a pair-associate memory task.

The Wechsler Memory Scale is a widely-used pencil-and-paper test of cognitive function and memory capacity. In the normal population, the standardized test yields a mean of 100 and a standard deviation of 15, so that a mild amnesia can be detected with a 10-15 point reduction in the score, a more severe amnesia with a 20-30 point reduction, and so forth (Squire, 1987). During the clinical interview, a battery of tests, including, but not limited to, the Minimental test, the Wechsler memory scale, or paired-associate learning are applied to diagnose symptomatic memory loss. These tests provide general sensitivity to both general cognitive impairment and specific loss of learning/memory capacity (Squire, 1987). Apart from the specific diagnosis of dementia or amnestic disorders, these clinical instruments also identify age-related cognitive decline which reflects an objective diminution in mental function consequent to the aging process that is within normal limits given the person's age (DSM IV, 1994). As noted above, "improvement" in learning and memory is present within the context of the present invention if there is a statistically significant difference in the direction of normality in the paired-associate test, for example, between the performance of ligand-inhibitor treated patients as compared to members of the placebo group or between subsequent tests given to the same patient.

WO 96/02569 PCT/US95/08867 28 Decreasing Food Intake As noted above, the present invention provides methods for decreasing food intake through the administration to a patient of a therapeutically effective amount of a ligand inhibitor of a CRF/CRF-BP complex. Such patients may be identified by being obese. An obese individual weighs more than a target weight considered normal for that person's age, gender and height and can be identified objectively by a body mass index (BMI calculated as weight in kilograms/height in meters 2 at or higher than the 85th percentile of the same reference population (National Center for Health Statistics, "Obese and Overweight Adults in the United States." Series 11, No. BO, U.S.

Government Printing Office, Washington, 1983). In addition, evidence that CRF is involved for a particular individual may be obtained by demonstrating decreased CRF levels in the cerebrospinal fluid or by brain imaging as described above. Because the hypothalamus is a common brain area mediating the effects of CRF on food intake and endocrine parameters, alterations in pituitary hormone concentration may also reflect altered levels in hypothalamic

CRF.

A decrease in food intake may be measured both in the delayed initiation of a meal and the reduction in the overall duration or quantity of food consumption.

Smith, "Satiety and the Problem of Motivation," in D.W. Pfaff The Physiological Mechanisms of Motivation, Springer-Verlag, New York, pp. 133-143, 1982. In addition, the selection of particular nutrients in a food choice situation serves as a supplemental measure of specific hunger (Rozin, Adv. Study Behav. 6:21, 1976).

There are two established animal models of appetite regulation. One is a simple measurement of food intake, and the second is a measurement of diet selfselection in a cafeteria environment. In the first method, food intake is limited for 24 hours followed by two hours of access to a preweighed portion of laboratory chow in the animal's home cage. Food intake is measured at 60 and 120 minutes by weighing the remaining pellets. These tests may also be performed on animals that are obese due to genetic mutations and which effectively reproduce symptoms of overeating and deranged nutrient selection (Argil6s, Prog. Lipid Res. 28:53, 1989; Wilding et al., Endocrinol. 132:1939, 1993).

In the cafeteria environment, diets are specially formulated with differing proportions of macronutrients, such as carbohydrate, protein, and fat, so as to measure preference for specific nutrients based on sensory attractiveness or postingestive benefit. Diet selection is altered, in part, by a wide variety of neurochemical systems. These tests are useful for detection of subtle changes in food intake regulation which impact phenomena, such as craving or bingeing, and are relevant for the diagnosis of eating disorders, such as anorexia nervosa and obesity. Following WO 96/02569 PCT/US95/08867 29 establishment of a baseline for animals, 15 minutes prior to testing each animal receives an ICV injection of control solution or a dose of 1, 5, or 25 lg of a peptide ligand inhibitor, or appropriate doses for a non-peptide ligand inhibitor. Food intake is measured as described for the feeding test or the diet self-selection in the cafeteria environment, and test results are compared to baseline.

In humans, obesity is related not only to overeating, but may also be related to consumption of nutritionally imbalanced diets such as a disproportionately large intake of sweet or fatty foods. (Drewnowski et al., Am. J. Clin. Nutr. 46:442, 1987.) Thus, clinical manifestations of appetite regulation are readily detected using controlled experimental diets or cafeteria self-selection protocols which record intake patterns in terms of quantity, meal duration, and choice (Kissileff, Neurosci. Biobehav.

Rev. 8:129, 1984). In these tests, following a baseline determination for each individual, measurement of food intake or self-selection in the cafeteria environment are measured. Improvement in the context of the treatment of obesity constitutes a weight loss or reduction in food intake exhibited by treated patients as compared to members of a placebo group. Moreover, this strategy has been successful in identifying serotonergic agonists for obesity.

Diseases Associated with Low Levels of CRF As noted above, the present invention provides methods for treating diseases associated with low levels of CRF through the administration to a patient of a therapeutically effective amount of a ligand inhibitor of a CRF/CRF-BP complex.

Such patients may be identified through diagnosis of eating disorders, neuroendocrine disorders, and cognitive disorders, such as Alzheimer's disease. In addition, other conditions associated with decreased CRF levels, such as atypical depression, seasonal depression, chronic fatigue syndrome, obesity, vulnerability to inflammation disease, post-traumatic stress disorder, and psychostimulant withdrawal often present a profile of hypothyroidism and decreased stress system activity which is identified characteristically by a decrease in urinary free cortisol and plasma ACTH. Thus, these diseases and conditions would likely be resolved in part by restoration or potentiation of brain CRF levels (Chrousos and Gold, JAMA 267:1244, 1992).

The hallmark of this diverse set of human disease states is dysregulation of the pituitary-adrenal axis with a presumed derangement of brain CRF. Hence, the fact that experimental alternation of CRF/pituitary-adrenal systems in laboratory animals reproduces essential features of the above syndromes, namely behavioral despair (Pepin et al., 1992), exercise fatigue (Rivest and Richard, 1990), obesity (Rothwell, 1989) and hyperarousal associated with psychostimulant withdrawal (Koob WO 96/02569 PCT/US95/08867 et al., 1993; Swerdlow et al., 1991) suggests the broad utility of pharmacotherapies designed to normalize endogenous levels of CRF.

The essential feature of seasonal depression (major depressive disorder with seasonal pattern) is the onset and remission of major depressive episodes at characteristic times of the year. In most cases, the episodes begin in fall or winter and remit in spring. Major depressive episodes that occur in a seasonal pattern are often characterized by prominent anergy, hypersomnia, overeating, weight gain, and a craving for carbohydrates and must persist for a period of at least two weeks during which there is either depressed mood or the loss of interest or pleasure in nearly all activities.

The essential feature of post-traumatic stress disorder is the development of characteristic symptoms following exposure to an extreme traumatic stressor involving direct personal experience of an event that involves actual or threatened death or serious injury to one's own or another's physical integrity. The person's response to the event must involve intense fear, helplessness, or horror. The traumatic event is reexperienced as intrusive recollections or nightmares which trigger intense psychological distress or physiological reactivity. The full symptom picture must be present for more than one month and cause clinically significant distress or impairment in social or occupational functioning.

The essential feature of nicotine withdrawal (nicotine-induced disorder) is the presence of a characteristic withdrawal syndrome that develops after the abrupt cessation of, or reduction in, the use of nicotine-containing products following a prolonged period (at least several weeks) of daily use. Diagnosis of nicotine withdrawal requires identification of four or more of the following: dysphoric or depressed mood, insomnia, irritability or anger, anxiety, difficulty concentrating, restlessness or impatience, decreased heart rate and increased appetite or weight gain.

These symptoms must cause clinically significant distress or impairment in social, occupational functioning.

Improvement constitutes either a statistically significant change in the symptomatic condition of a treated individual as compared to a baseline or pretreatment condition on measures pertinent to the disease model; or a statistically significant difference in the symptomatic condition of ligand-inhibitor treated patients and members of a placebo group. Clinical instances of disease exhibit symptoms which are, by definition, distinguishable from normal controls. For depression, several rating scales of depression are used. (See Klerman et al., Clinical Evaluation of Psychotropic Drugs: Principles and Guidelines, Prien and Robinson Raven Press, Ltd., New York, 1994). One test, the Hamilton Rating Scale for Depression, is widely used to WO 96/02569 PCT/US95/08867 31 evaluate depression and is also used to assess symptom changes in response to treatment. Other tests and ratings can be found in the DSM-IV manual. For nicotine withdrawal, as well as the other disorders, tests for evaluation of the severity of the disorder can be found in the DSM-IV manual.

Alzheimer's Disease As noted above, the present invention provides methods for treating Alzheimer's disease through the administration to a patient of a therapeutically effective amount of a ligand inhibitor of a CRF/CRF-BP complex. Such patients may be identified through clinical diagnosis based on symptoms of dementia or learning and memory loss which are not attributable to other causes. In addition, patients are also identified through diagnosis of brain atrophy as determined by magnetic resonance imaging.

Decreased levels of CRF are shown to be implicated in Alzheimer's disease. Brains obtained post-mortem from ten individuals with AD and ten neurologically normal controls were chosen for study. Standard areas of frontal pole, parietal pole, temporal pole, and occipital pole were dissected from fresh brain, frozen in dry ice, and stored at -70C until they were processed for CRF radioimmunoassay and CRF-BP assay. Formalin-fixed samples of the cerebral cortex and hippocampus were embedded in paraffin and subsequently sectioned and stained with hematoxylin/eosin and silver impregnation. Examination of stained sections from brains of AD patients showed abundant neuritic plaques and neurofibrillary tangles typical of AD, whereas control cases showed none.

Levels of CRF and CRF-BP in the cerebral cortices of Alzheimer's patients and controls were determined. CRF-BP has previously been identified and characterized in rat brain, sheep brain, and human plasma. In the cerebral cortex of brains studied here, the majority of CRF-BP was membrane associated.

Pharmacological characteristics of CRF-BP solubilized from human brain membranes from either controls or AD patients showed no differences in binding characteristics to CRF and ligand inhibitors when compared to a recombinant form of the soluble plasma CRF-BP. In brains of Alzheimer's patients, the levels of CRF in the frontal, parietal, and temporal cerebral cortex were dramatically reduced compared to normal brains, but the levels in the occipital cortex were only slightly decreased (Figure 3A). In contrast, CRF-BP levels were similar in brains of AD patients and normal controls (Figure 3B).

These data provide direct evidence for the presence of CRF-BP in normal and AD brain tissue and suggest that deficits in CRF levels seen in AD may be due to decreased synthesis or increased degradation of CRF rather than neuronal loss. If there is WO 96/02569 PCT/US95/08867 32 neuronal loss in AD, then CRF-BP may be preferentially localized to non-CRF neurons.

The nature of the interaction of CRF with CRF-BP in human brain tissue was determined by measuring the proportions of CRF complexed to CRF-BP and in the free pool. Using assays specific for total CRF and bound CRF, approximately 40% and of the total CRF were found complexed with CRF-BP in normal and Alzheimer's cerebrocortical extracts, respectively. Furthermore, CRF was bound to CRF-BP in a reversible manner because treatment of the tissue with human CRF (6-33) or a-helical oCRF (9-41) displaced CRF from CRF/CRF-BP complex (Figure These data directly demonstrate that CRF-BP ligand inhibitors increase the free concentration of CRF. Moreover, treatment with a CRF-BP ligand inhibitor raised free CRF levels in AD brains to the level found in normal brains.

Several established animal models of Alzheimer's disease which focus on cholinergic deficits are available. The primary role of cholinergic deficits in AD is well established. In AD, there are significant positive correlations between reduced choline acetyltransferase activity and reduced CRF levels in the frontal, occipital, and temporal lobes (DeSouza et al., 1986). Similarly, there are negative correlations between decreased choline acetyltransferase activity and an increased number of CRF receptors in these three cortices In two other neurodegenerative diseases, there are highly significant correlations between CRF and choline acetyltransferase activity in Parkinson's disease, but only a slight correlation in progressive supranuclear palsy (Whitehouse et al., 1987).

In rats, anatomic and behavioral studies evidence interactions between CRF and cholinergic systems. First, in some brain stem nuclei, CRF and acetylcholinesterase are co-localized, and some cholinergic neurons also contain CRF.

Second, CRF inhibits carbachol-induced behaviors (carbachol is a muscarinic cholinergic receptor antagonist), suggesting that CRF has effects on cholinergic systems (Crawley et al., Peptides 6:891, 1985). Treatment with another muscarinic cholinergic receptor antagonist, atropine, results in an increase in CRF receptors (DeSouza and Battaglia, Brain Res. 397:401, 1986). Taken together, these data show that CRF and cholinergic systems interact similarly in humans and animals.

An animal model of Alzheimer's disease which focuses on cholinergic deficits is produced by the administration of scopolamine, a non-selective postsynaptic muscarinic receptor antagonist that blocks the stimulation of postsynaptic receptors by acetylcholine. In these animals, memory deficits are readily apparent as measured by passive avoidance or delayed-matching-to-position tests, which distinguish motor or perceptual deficits from amnesia or cognitive enhancing effects of experimental WO 96/02569 PCT/US95/08867 33 treatments. Thus, the Morris maze and Y-maze tests following scopolamine-induced amnesia are utilized to test memory impairment and subsequent enhancement following administration of ligand inhibitor. In the Morris maze, the design of the experiment is essentially as described above, but is modified to include treatment 30 minutes prior to training on each of days 1 to 3 with an ip injection of scopolamine hydrobromide (0.3 mg/kg). This amnestic dose of scopolamine impairs acquisition and retention of spatial and avoidance learning paradigms in the rat. The anti-amnestic effects of 1, or 25 .g of a peptide ligand inhibitor, or appropriate doses of a non-peptide ligand inhibitor, are measured relative to the concurrent control groups who receive or do not receive scopolamine. The effect of the ligand inhibitors on reversal of scopolamineinduced amnesia using the Y-maze is performed similarly to the Y-maze test described above. Modification of this test includes treatment 30 minutes prior to training on days 5 to 10 with an ip injection of scopolamine hydrobromide (0.3 mg/kg). The antiamnestic effects of 1, 5, or 25 utg of a peptide ligand inhibitor administered ICV or equivalent doses of a non-peptide ligand inhibitor are administered centrally or systemically, are measured relative to concurrent control and scopolamine treatedcontrol groups.

Several tests measuring cognitive behavior in AD have been designed.

(See Gershon et al., Clinical Evaluation of Psychotropic Drugs: Principles and Guidelines, Prien and Robinson Raven Press, Ltd., New York, 1994, p. 467.) One of these tests, BCRS, measures concentration, recent memory, past memory, orientation, and functioning and self-care. The BCRS is designed to measure only cognitive functions. This test, as well as the Weschler Memory Scale and the Alzheimer's Disease-Associated Scale, may be used to determine improvement following therapeutic treatment with ligand inhibition. As noted above, "improvement" in Alzheimer's disease is present within the context of the present invention if there is a statistically significant difference in the direction of normality in the Weschler Memory Scale test, for example, between the performance of ligand-inhibitor treated patients as compared to members of the placebo group or between subsequent tests given to the same patient. In addition, scopolamine-induced amnesia in humans can be used as a model system to test the efficacy of the ligand inhibitors.

Administration of Ligand Inhibitor As used herein, the terms "pharmaceutically acceptable", "physiologically tolerable" and grammatical variations thereof, as they refer to compositions, carriers, diluents and reagents, are used interchangeably. Preferably, the WO 96/02569 PCT/US95/08867 34 materials are capable of administration to a mammal without the production of undesirable physiological effects, such as nausea, dizziness, gastric upset and the like.

A ligand inhibitor of a CRF/CRF-BP complex is administered to a patient in a therapeutically effective amount. A therapeutically effective amount is an amount calculated to achieve the desired effect, either increasing the level of free CRF in the brain, improving learning and memory, decreasing food intake, activating CRF neurocircuitry in the brain, treating diseases associated with low levels of CRF in the brain, treating the symptoms associated with Alzheimer's disease, treating obesity, treating atypical depression, treating substance abuse withdrawal, treating post-partum depression, or age-related memory loss. It will be apparent to one skilled in the art that the route of administration may vary with the particular treatment and also with whether a peptide or non-peptide ligand inhibitor is administered. Routes of administration may be either non-invasive or invasive. Non-invasive routes of administration include oral, buccal/sublingual, rectal, nasal, topical (including transdermal and ophthalmic), vaginal, intravesical, and pulmonary. Invasive routes of administration include ICV, intraarterial, intravenous, intradermal, intramuscular, subcutaneous, intraperitoneal, intrathecal and intraocular.

Intracerebroventricular (ICV) injections are performed on animals as follows. Animals are anesthetized with halothane and secured in a KOPF stereotaxic instrument. A guide cannula aimed above the lateral ventricle is implanted and anchored to the skull with two stainless steel screws and dental cement. For injections, a 30 gauge stainless steel cannula attached to 60 cm of PE 10 tubing is inserted through the guide to 1 mm beyond its tip. Two microliters of ligand inhibitor are injected by gravity flow over a one minute period simply by raising the tubing above the head of the animal until flow begins. Procedures for the other routes of administration are well known in the art.

The required dosage may vary with the particular treatment and route of administration. In general, dosages for peptide ligand inhibitors are given to achieve an end concentration approximately 50 to 125 pg per 1.5g of brain tissue or 15 to 38 nmoles per 1.5g of tissue. Dependent, however, on the size of the protein or polypeptide, a relatively larger or smaller amount is employed. These treatments are conducted two or three times a week. Treatments may need to be continuous for retention of therapeutic benefit. Patients are monitored by assessing CRF levels in cerebral spinal fluid or in the brain by imaging as described above. In addition, patients are monitored by assessing performance under various tests as described for each of the treatments.

WO 96/02569 PCTIS95/08867 Therapeutic administration is performed under the guidance of a physician, and pharmaceutical compositions contain the ligand inhibitor in a pharmaceutically acceptable carrier. These carriers are well known in the art and typically contain non-toxic salts and buffers. Such carriers may comprise buffers like physiologically-buffered saline, phosphate-buffered saline, carbohydrates such as glucose, mannose, sucrose, mannitol or dextrans, amino acids such as glycine, antioxidants, chelating agents such as EDTA or glutathione, adjuvants and preservatives. Acceptable nontoxic salts include acid addition salts or metal complexes, with zinc, iron, calcium, barium, magnesium, aluminum or the like (which are considered as addition salts for purposes of this application). Illustrative of such acid addition salts are hydrochloride, hydrobromide, sulphate, phosphate, tannate, oxalate, fumarate, gluconate, alginate, maleate, acetate, citrate, benzoate, succinate, malate, ascorbate, tartrate and the like. If the active ingredient is to be administered in tablet form, the tablet may contain a binder, such as tragacanth, corn starch or gelatin; a disintegrating agent, such as alginic acid; and a lubricant, such as magnesium stearate.

If administration in liquid form is desired, sweetening and/or flavoring may be used, and intravenous administration in isotonic saline, phosphate buffer solutions or the like may be effected.

The peptides being administered under the guidance of a physician will usually be in the form of pharmaceutical composition that contains the ligand inhibitor and a conventional, pharmaceutically-acceptable carrier. Usually, the dosage will be from about 1 to about 1000 micrograms of the peptide per kilogram of the body weight of the host animal per day; frequently it will be between about 100 .tg and about 1 mg but may vary up to about 10 mg. Treatment of subjects with these peptides can be carried out to alleviate symptoms and correct detrimental manifestations which are characteristic of the relatively low ambient CRF levels which occurs in Alzheimer's disease patients and chronic fatigue syndrome patients. In the former case, treatment improves short and medium term memory. However, for many of these indications including suppression of appetite, it is necessary for the peptide to be delivered to the brain and preferably it is coupled with an agent capable of penetrating the blood-brain barrier. Administration by iv, im or sc injection effects an increase in cortisol level and can effect a lessening of fatigue. Treatment of subjects with these ligand inhibitors can also be carried out to boost the effective biological concentration of free CRF in order to stimulate the human respiratory system by administration to reach the brain. For treatment of conditions in medical emergencies, such as cardiovascular arrest and shock due to substances or pathological agents, an increase in cortisol level can be achieved by administration by iv, im or sc injection. To promote parturition in pregnancy, the WO 96/02569 PCT/US95/08867 36 ligand inhibitor, optionally with additional hCRF (or an analogue thereof), is administered to cause the concentration of free hCRF in plasma to rise to at least about 250 pmole/liter or preferably to at least about 0.1 ng/ml. It may also be similarly administered to patients afflicted with AIDS who frequently have low levels of cortisol so that such a CRF-BP blocker could elevate ACTH and cortisol.

When the agent is administered along with CRF or a CRF agonist, such CRF peptide may be administered as a daily dosage of from about 1 to about 200 micrograms of the CRF peptide per kilogram of body weight of the host animal.

Examples of suitable CRF agonists include those described in U.S. Patents Nos.

4,415,558, 4,489,163, 4,594,329, 5,112,809, 5,235,036 and 5,278,146, the disclosures of which are incorporated herein by reference. When the agent is administered along with a CRF antagonist, the CRF antagonist may be administered in an amount between about 0.01 to about 10 mg of the peptide per kilogram of body weight of the host animal. Suitable CRF antagonists are disclosed in U.S. Patents Nos. 4,605,642, 5,109,111 and 5,245,009, the disclosures of which are incorporated herein by reference.

Preferred CRF antagonists include [D-Phel 2

N

21 38 ]-hCRF(12-41) and [D-Phe' 2 Nle 2 1 l 38, CML 37 ]-hCRF(1241). Preferred CRF agonists include [His 20 Nle 21 Leu 38 ]-hCRF, [D-Phe' 2 Nle 2 1, 38, Leu 36 ]-hCRF, [D-Pro 4 D-Phel2, Asp 25 Nle 2 1, 38 ]-hCRF and [D- Pro 4 D-Phel 2 Nle 2 1, 38, CML 37 ]-hCRF.

The following examples are offered by way of illustration and not limitation.

Example 1 Ligand immunoradiometric assay (LIRMA) to assay for ability of ligand inhibitor displacement of CRF from CRF-BP The assay is performed in 600 l1 polypropylene microfuge tubes or a 96-well plate. First, 50 pl of a 250 ng/ml of purified recombinant CRF-BP is added to 150 gl of PBS binding buffer (50 mM sodium phosphate, 0.15 M NaC1, and 0.02% NP-40). Next, 12 5 I-h/r CRF at a final concentration of 200 pM and 50 pl of a 10-100 g M concentration of the ligand inhibitor are added and incubated for 1 hr at room temperature. To the reaction, 50 pl of anti-CRF-BP antibody, 5144, diluted 1:1000 with assay buffer is added to each tube and allowed to bind for a further 1 hr at room temperature. The total volume in all tubes is adjusted to 300 gl with assay buffer.

Bound complexes are precipitated by the addition of 200 gl of preprecipitated goat antirabbit (GAR) second antibody at 1:20 in 1% normal rabbit serum, 4% PEG, 50 mM sodium phosphate, 0.1% sodium azide, followed by incubation for 1 hr at room

F-

WO 96/02569 PCT/US95/08867 37 temperature. The antibody-bound 12 5 I-CRF precipitate is then collected by centrifugation (3000 x g) at 4 0 C for 20 min in a Beckman GS-15R centrifuge. Using a Beckman Biomer 1000 robotic workstation, the tubes are aspirated and washed once with 600 pl of PBS plus 0.02% NP-40. Tubes containing the pellets are then transferred to 12 x 74 mm counting tubes and counted in a gamma counter.

Inhibitory binding affinity constant (Ki) values are determined using parameters calculated by the LIGAND computer program, Munson et al., Anal.

Biochem. 107:220, 1980, and a Vax/VMS computer system. Errors are shown which are the standard error of the mean from 3 replicate binding assays.

Example 2 Detergent phase separation to assay for the ability of ligand inhibitor displacement of CRF from CRF-BP Recombinant human CRF-BP at 200 ng/ml is incubated in binding buffer (phosphate buffered saline, pH 7.4/0.02% NP-40) with radiolabeled 2 I-h/r CRF pM) for 2 hours at room temperature. Following incubation, bound and free CRF are separated by the addition of a 1:10 dilution of Triton X-114TM in assay buffer octylphenoxypolyethoxyethanol (SIGMA). Triton X-114TM is insoluble in water at room temperature and in aqueous solution can be separated into a detergent phase.

After addition of the Triton X-l 14TM, the tube is vortexed and immediately centrifuged at room temperature at 12,000 x g for 5 minutes. The detergent phase is found at the bottom of the tube while the aqueous phase remains at the top. CRF, which as amphiphilic alpha helices, segregates to the detergent phase. However, when CRF is bound to CRF-BP, the CRF/CRF-BP complex remains in the aqueous phase. Thus, a l1 aliquot of the aqueous phase is transferred to a 12 x 74 mm plastic tube and counted. The amount of radioactivity left in the supernatant is determined.

Example 3 ACTH release assay for free CRF Four rats are killed by decapitation and the anterior pituitary glands are removed. The anterior pituitary glands are washed 6 times with sterile HEPES buffer and transferred to 20 ml of collagenase solution (4 mg/ml). The pituitaries in collagenase are then transferred to a 25 ml Bellco dispersion flask and stirred for min at 37 0 C. After 30 minutes, the pituitary cell suspension is then triturated by drawing the pituitaries through a 10 ml pipette and incubated for 30 min more before WO 96/02569 PCT/US95/08867 38 more trituration. The cell suspension is incubated for a further 45 min and the partially dispersed cells are transferred to a sterile 50 ml tube and centrifuged at 4000 rpm for 4 min. The cell pellet is reconstituted in 10 ml of neuraminidase (8 ug/ml) and vortexed. The suspension is placed in a water bath for 9 min, vortexed again for 4 min and centrifuged again. The supernatant is poured off and the cell pellet is reconstituted in 25 ml of BBM-P (250 ml BBM-T plus 5 ml of 2% fetal calf serum) by vortexing.

The cells are collected by centrifugation, and finally the suspension is reconstituted in 22 ml of BBM-P. The cells are then plated at a density of 50-100,000/well in a 48 well plate and incubated in a humidified

CO

2 chamber for 2 days. On the day of assay, the cells are washed once with BBM-T in preparation for stimulation with the various relevant analogues.

The cells are stimulated with a maximally stimulating dose of h/rCRF (1 nM) in the presence and absence of a blocking concentration of CRF-BP (5 nM).

This concentration of CRF-BP reduces the amount of ACTH released from the pituitary cells by binding to h/rCRF. The reduction is expressed as a fraction of the amount of ACTH released by 1 nM CRF in the absence of CRF-BP. The CRF-BP (5 nM), which is bound to h/rCRF (1 nM), is incubated with a range of concentrations of ligand inhibitors typical concentrations for the CRF-BP ligand h/rCRF (6-33) range from 0.1-1000 nM). The ligand inhibitor binds to CRF-BP and displaces CRF from the complex resulting in a dose-dependent reversal of the inhibition of h/rCRF induced-ACTH secretion by CRF-BP. The potency of the CRF-BP ligand is expressed as a fraction of ACTH release obtained by stimulation with 1 nM CRF alone.

Example 4 cAMP production assay to measure free CRF The assay for detection of CRF-stimulated adenylate cyclase activity is carried out as previously described (Battaglia et al., 1987) with minor modifications.

The standard assay mixture contains 2 mM L-glutamine, 20 mM HEPES, 1 mM IBMX (isobutylmethyl xanthine) in DMEM buffer. In stimulation studies, cells that have been transfected with a clone encoding CRF receptor are plated in 24 well plates and incubated for 1 hr at 37 0 C with various concentrations of CRF-related and unrelated peptides. Following the incubation, the media is aspirated, the wells rinsed once gently with fresh media and aspirated. The amount of intracellular cAMP is determined after lysing the cells in 300 1 l of a solution of 95% ethanol and 20 mM HCI at 20 0 C for 16-18 hrs. The lysate is transferred into 1.5 ml Eppendorf tubes, the wells are washed with an additional 200 gl of EtOH/HCI, and the wash is pooled with the lysate. The WO 96/02569 PCT/US95/08867 39 lysates are lyophilized and resuspended in 500 pl of sodium acetate buffer, pH 6.2.

cAMP is measured with a single antibody kit from Biomedical Technologies Inc.

(Stoughton, MA). For the functional assessment of CRF receptor antagonists, a single concentration of CRF or related peptides causing 80% stimulation of cAMP production is incubated along with various concentrations of competing compounds (10-12 to 10-6 The incubation and measurement conditions for cAMP are performed as described.

Example Two-site ELISA to measure CRF levels A. Preparation of brain tissue samples.

Autopsy samples were weighed and homogenized in 5 ml of sucrose. One ml of each sample was centrifuged at 10,000 x g for 10 min at room temperature and the resultant membrane pellets were reconstituted in SPEA (50 mM sodium phosphate pH 7.4, 0.1M NaCI, 25 mM EDTA, 0.1% sodium azide, containing 0.25% bovine serum albumin (BSA) and 1% NP-40). Eight hundred microliters of cerebral cortex homogenate (in 10% sucrose) was further extracted by the addition of 200 pl of TTBS (Tris-buffered saline with 0.5% Tween -20, 1% NP-40, 1% BSA) followed by vortexing for 1 min. The sample was centrifuged at 10,000 x g for 10 min at room temperature. The resultant supernatant was kept for analysis of "total CRF," "bound CRF" CRF bound to CRF-BP) and "free CRF" using a two-site CRF

ELISA.

B. Measurement of total CRF.

Briefly, ELISA plates were coated for 2 hr at 37 0 C with protein G-purified sheep anti-CRF antibody (20 pg/ml) diluted in 50 mM sodium bicarbonate buffer, pH 9.5. Plates were washed once with TTBS and blocked with 1% casein in TBS for 1 hr at room temperature. One hundred microliters of the samples or standard were added to each well and allowed to bind at room temperature. Plates were washed five times with TTBS. RC-70 rabbit anti-human CRF antibody (diluted 1:1000 in TTBS/1% BSA) was added. Following incubation for 1 hr at room temperature, plates were washed five times with TTBS buffer and then exposed to horseradish peroxidase conjugated-goat anti-rabbit (GAR) second antibody for 1 hr at room temperature. Plates were finally washed five times with TTBS and developed by the addition of 100 pl of TMB microwell peroxidase substrate solution (Kikegaard and Perry Laboratories, Inc.). Absorbance at 450 nM was determined.

WO 96/02569 PCTIUS95/08867 C. Measurement of bound CRF.

Bound CRF was determined by capture of the CRF/CRF-BP complex in wells which had been pre-coated with an anti-human CRF-BP monoclonal antibody (5 gg/ml of antibody in 50 mM sodium bicarbonate buffer, pH 9.5) followed by detection of the bound CRF with RC-70 anti-human CRF antibody essentially as described for the total CRF ELISA.

D. Measurement of free CRF.

Free CRF is measured in the supernatant following capture of the bound complex by the anti-CRF-BP monoclonal antibody. Following binding of sample material, the supernatant is removed to a new ELISA plate coated with protein G-purified sheep anti-CRF antibody. The assay is then performed as described for determining total CRF levels.

Example 6 Screening for ligand inhibitors Candidate ligand inhibitors may be screened for their ability to displace CRF from CRF/CRF-BP complex. A suitable assay, such as ACTH release from cultured pituitary cells (see Example 2) or two-site LIMRA (see Example is used to measure free CRF and CRF-BP levels, respectively.

In the LIMRA assay, generally the procedure from Example 1 is followed. The ligand inhibitor at a 10 pM concentration is added to the reaction along with the 125 h/r CRF. If the candidate ligand inhibitor displaces CRF from CRF-BP, the pellets will contain less radioactivity in comparison to controls in which no candidate peptide is added. Candidate peptides are re-screened using a 6 point-dose curve. values are calculated.

The ligand inhibitors human CRF, a-helical ovine CRF human CRF and ovine CRF were screened in this manner against CRF/CRF-BP complex. The affinities of these peptides for cloned human pituitary CRF receptor was determined in membrane preparations of stable transfectants of the receptor in Ltkmouse fibroblast cells using a previously characterized radioligand binding assay (DeSouza, J. Neurosci. 7:88, 1987).

WO 96 02569 PCT/US95/08867 41 Results of these assays is shown in the following Table. Moreover, these ligand inhibitors were also tested on CRF-BP in cerebral cortices of individuals with Alzheimer's disease and controls.

Table I Ligand Inhibitors of CRF/CRF-BP

IC

50 Values (nM) Peptide CRF-BP in Cerebral Cortex Recombinant Human Controls Alzheimer's CRF-BP

CRF-R

Disease human CRF 0.30 0.18 0.19 a-helical ovine CRF (9-41) 0.16 0.04 0.20 9.9 human CRF (6-33) 1.6 1.1 1.6 >1000 ovine CRF 667 595 471 2.4 These results show that human CRF, a-helical ovine CRF (9-41) and human CRF (6-33) were nearly equally effective at displacing CRF from CRF/CRF-BP complex. In contrast, ovine CRF was not as effective in displacing CRF. The CRF-BP in cerebral cortices from diseased or normal individuals was equivalent to the recombinant form. Of the three best ligand inhibitors, human CRF and a-helical ovine CRF (9-41) bound human CRF-R approximately equivalently. In marked contrast, human CRF (6-33) binds 2-3 logs less efficiently.

Example 7 Treatment with ligand inhibitor Five cerebrocortical samples from Alzheimer's patients and five samples from age-matched, normal controls were prepared as in Example 4 and pooled. Levels of total CRF, bound CRF, and free CRF were measured as described in Example 4. As can be seen in Figure 2, 40% of the total CRF was complexed to CRF-BP in brain extracts from normal individuals and 60% was complexed in brain extracts from Alzheimer's patients.

WO 96/02569 PCT/US95/08867 42 The effect of a high affinity CRF-BP ligand to displace bound CRF was assessed after monoclonal capture of the CRF/CRF-BP complex in the presence of nM a-helical ovine CRF Displaced free CRF was measured in the supernatants remaining after CRF-BP monoclonal antibody capture by CRF ELISA.

As seen in Figure 2, treatment of brain tissues with the ligand inhibitor caused release of all bound CRF in both Alzheimer's and control tissues. Moreover, treatment of the Alzheimer's disease cerebral cortex with the ligand inhibitor replenished the free CRF levels to the level seen in age-matched controls.

Tissue was also incubated with 50 nM of the ligand inhibitor a-helical ovine CRF The supernatant was then assayed for displaced CRF by two-site ELISA assay as described in Example 4.

Example 8 Morris water maze test The Morris Water Maze Test is a simple spatial learning task that requires a minimal amount of stress and experience. No motivational constraints such as shock or food deprivation are necessary. The animal is placed in tepid water, which is opaque due to the addition of milk powder. The latency time to find a hidden platform is monitored. The animals learn the location of the platform relative to visual cues located within the maze and the testing room; this learning is referred to as place learning. This test is particularly sensitive to manipulations of the hippocampus, a critical brain area involved in spatial learning in animals and memory consolidation in humans.

The apparatus used in this test is a pool (46.4 cm in diameter, 45.7 cm high) filled to a depth of 23 cm with opaque water (22 0 C-25 0 The top of a weighted target platform, 10 cm in diameter, is located 1-2 cm beneath the water surface. Four equal quadrants of the pool are distinguished by designs located on the inner surface.

The animal is placed into a designated quadrant of the tank and the time to approach and ascend the hidden platform is measured; the location of subject placement and platform remain constant throughout the experiment. After climbing on top of the platform, the animal is allowed to rest for 20 sec. Subjects that do not find the platform within 60 sec are placed onto the platform and allowed to rest for 20 sec.

Rats were treated by ICV injection 15 min prior to testing with either the ligand inhibitor h/rCRF (6-33) or the CRF receptor agonist h/rCRF. Doses of h/rCRF (6-33) were 0, 1, 15, 25, 50, or 125 tg; doses of h/rCRF were 0, 0, 1, 1 or 2.5 g.

WO 96/02569 PCT/US95/08867 43 Seven to 10 rats per group were treated. Statistical analysis confirmed significant improvement in performance following treatment with either h/rCRF (6-33) (p<0.05) or CRF (p<0.05) (Figure There was a significant improvement in performance over time as well (p<0.0001).

Example 9 Y-maze visual discrimination test The Y-Maze visual discrimination test is a learning test using positive reinforcement to study learning with minimal stress to the animals. Subjects are meal deprived and fed only after the training session; animals have the option of not responding, but do so in most cases because the positive reinforcing properties of the food pellets, which rats prefer to regular chow.

The Y-maze contains three arms of equal length (61 cm long, 14 cm wide, 30 cm high). One arm is used as a start box and is separated from the other two goal arms by guillotine doors which are manually operated. The vertical surface at the ends of the two distal arms is equipped with an eight watt electric bulb. On the first day of training, rats, which have been food-deprived to 80% body weight, were allowed to explore the maze for 5 min with two food pellets (45 mg Noyes) available at the end of each goal arm. On the second day, each rat was allowed one trip down each of the goal arms which were baited with pellets. On the third day, rats received six spaced trials in squads of three animals where one goal arm was closed at the choice point and two mg pellets were available in the open goal box, but no discriminative visual stimulus was provided (light off). The open arm alternated from left to right over the six trials, as well as from subject to subject. On days 4-10, both goal arms were open and the light at the end of one goal arm was illuminated. Ten trials were run daily, again in squads of three so that the intertrial interval was about 90 sec. Subjects were fed a 15 g portion of laboratory chow in the home cage daily at the conclusion of training.

On days 4-10, ICV injections of ligand inhibitor were given immediately prior to testing. Groups of 7-9 rats received either 0, 1, 5, or 25 jgg of the ligand inhibitor h/rCRF Percent correct responses were recorded. Rats receiving 5 and ftg CRF (6-33) had statistical significant better performance than rats receiving 0 or 1 gg of ligand inhibitor.

WO 96/02569 PCTUS95/08867 44 Example Elevated plus-maze test The Elevated plus-maze test predicts how animals respond to an approach-avoidance situation involving an exposed, lighted space versus a dark, enclosed area. In the maze, both spaces are elevated off the ground and constitute two runways intersecting in the form of a plus sign. This type of approach-avoidance situation is a classical test of "emotionality" and is very sensitive to treatments that produce disinhibition (such as sedative or hypnotic drugs) and stress. No motivational constraints are necessary and the animal is free to remain in the dark or venture out onto the open arms.

The elevated plus-maze apparatus has four arms (50 cm long, 10 cm wide) situated at right angles to each other and elevated from the floor (50 cm). Two of the arms are enclosed with walls (40 cm high) and two arms have no walls (open arms).

Subjects were placed individually into the center of the maze and allowed free access to all four arms for a 5 min testing period. The time spent in each arm was recorded automatically by photocell beams and a computer interface.

Groups of 7-10 rats received ICV injections of the ligand inhibitors h/rCRF (6-33) or h/rCRF Rats received 0, 0.1, 1, or 25 gg of h/rCRF (1-41).

Doses of 1 and 25 gg of h/rCRF (1-41) produced statistically significant more time on the open arms, indicating increased anxiety (Figure In marked contrast, memoryenhancing doses of CRF as well as doses two- to five-fold higher (50-125 jig) did not alter performance or produce overt behavioral alterations comparable to h/rCRF h/rCRF (1-41) is known to be a CRF-receptor agonist. Therefore, these data demonstrate a clear-cut functional dissociation of the efficacious cognitive enhancing and anxiogenic side effects for a ligand inhibitor.

Example 11 Animal Models of Obesity Obesity involves an excess of body fat arising from a level of energy intake which exceeds energy expenditure. The complex etiology of obesity in clinical populations may involve overeating, abnormal lipid metabolism, insulin excess and diminished physical activity. These phenomena and the resulting increase in body mass can be modeled in animals using genetically obese rats and mice, animals with lesions of hypothalamic regions of the brain, and animals having various long-term pharmacological treatments.

WO 96/02569 PCTJUS95/08867 Centrally administered CRF exerts a beneficial anorexic action in halting excessive weight gain in genetically obese Zucker rats. Appetite suppressive effects of endogenous CRF have recently been explored using a ligand inhibitor of a CRF/CRFbinding protein complex, h/r CRF This ligand inhibitor acts as an indirect CRF agonist, which increases synaptic levels of free, unbound CRF in brain. Central administration of h/r CRF (6-33) immediately prior to a two hour meal of laboratory chow in animals deprived of food for 24 hours produces a dose dependent suppression in appetite. Relative to CRF, h/r CRF (6-33) is significantly less potent and less effective in reducing appetite and, in contrast to CRF, does not alter the intake of a two hour mean in non-deprived subjects. Moreover, in separate experiments the anorexic dose of h/r CRF (6-33) did not induce fear-like behaviors which would be expected following central administration of CRF itself. These results disclose that pharmacological treatment with h/r CRF (6-33) exerts selective and moderate appetite suppression by stimulating a physiologically relevant increase in levels of endogenous

CRF.

Weight gain and overeating are undesirable features of nicotine withdrawal which may be resolved by pharmacological targeting of disregulated biological and neurochemical substrates for energy balance and appetite control.

Augmented hunger and weight gain often persist for at least six months and result in an average weight gain of four to six pounds over the first year after stopping smoking.

This situational obesity which arises in over three quarters of smokers who quit smoking is not effectively remedied by standard behavioral weight loss strategies. As an example of the clinical phenomenon, one investigator reported that the body weight of abstinent women at 26 weeks post-cessation was nine pounds over baseline although caloric intake registered in the normal range. Such long term withdrawal symptoms may play a major role in replace to smoking.

Appetite and weight disorders following smoking cessation are a reproducible component of the nicotine abstinence syndrome modeled in animals.

Figure 6 illustrates the changes in body weight and food intake of laboratory rats produced by continuous infusion of nicotine over two weeks at dependence-inducing levels and by a two week abstinence following discontinuation of nicotine administration. Chronic nicotine administration diminishes food intake and the rate of body weight gain relative to a vehicle-treated group while subsequent nicotine withdrawal induces overeating and normalization of body weight relative to controls. It seems likely that the effects of smoking cessation on energy balance are elicited by the abrupt removal of nicotine since nicotine replacement therapy is antidotal to appetite and weight disturbances in withdrawal. This clinical observation is confirmed by P:\OPER\JMS\31283-95.CLM 1/4/99 -46animal models in which the behavioral disruption and overall abstinence signs measured at 24 hours following termination of nicotine infusion are alleviated by acute systemic administration of nicotine.

In an experiment designed to reveal beneficial anorexic properties of CRF-BP antagonists in the context of nicotine withdrawal-induced overeating, both non-dependent and nicotine abstinent subjects were administered h/r CRF (6-33) immediately prior to a two hour meal occurring four days following cessation of nicotine administration. Figure 7 reveals that while the ligand inhibitor Wr CRE (6-33) did not alter appetite in nicotine-naive controls, nicotine abstinent subjects exhibited an increased level of food intake which was significantly reduced by administration of h/r CR Taken together, these results disclose that a dose of ligand inhibitor which does not alter appetite in either non-food deprived or spontaneously hungry animals, effectively attenuates the excessive appetite stimulated by either food deprivation or nicotine withdrawal.

15 Throughout this specification and the claims which follow, unless the context requires otherwise, the word "comprise", and or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated integer or step or group S i of integers or steps but not the exclusion of any other integer or step or group of integers or S steps.

S° From the foregoing, it will be appreciated that, although specific embodiments of the invention have been described herein for purposes of illustration, various modifications may be made without deviating from the spirit and scope of the invention. Accordingly, the invention is not limited except as by the appended claims.

WO 96/02569 PCT/US95/08867 47 SEQUENCE LISTING GENERAL INFORMATION: APPLICANT: Behan, Dominic P.

Heinrichs, Stephen C.

Sutton, Steven W.

Lowry, Phillip J.

Rivier, Jean E.F.

DeSouza, Errol B.

Vale Jr,, Wylie W.

(ii) TITLE OF INVENTION: METHODS FOR INCREASING ENDOGENOUS

LEVELS

OF CORTICOTROPIN-RELEASING

FACTOR

(iii) NUMBER OF SEQUENCES: 4 (iv) CORRESPONDENCE ADDRESS: ADDRESSEE: SEED and BERRY STREET: 6300 Columbia Center, 701 Fifth Avenue CITY: Seattle STATE: Washington COUNTRY: USA ZIP: 98104-7092 COMPUTER READABLE FORM: MEDIUM TYPE: Floppy disk COMPUTER: IBM PC compatible OPERATING SYSTEM: PC-DOS/MS-DOS SOFTWARE: PatentIn Release Version #1.30 (vi) CURRENT APPLICATION DATA: APPLICATION NUMBER: US FILING DATE: 07-JUN-1995

CLASSIFICATION:

WO 96/02569 PCTIS95/8867 48 (viii) ATTORNEY/AGENT INFORMATION: NAME: Nottenburg, Carol REGISTRATION NUMBER: 39.317 REFERENCE/DOCKET NUMBER: 690068.403 (ix) TELECOMMUNICATION INFORMATION: TELEPHONE: (206) 622-4900 TELEFAX: (206) 682-6031 TELEX: 3723836 SEEDANDBERRY INFORMATION FOR SEQ ID NO:1: SEQUENCE CHARACTERISTICS: LENGTH: 41 amino acids TYPE: amino acid STRANDEDNESS: single TOPOLOGY: unknown (ii) MOLECULE TYPE: peptide SEQUENCE DESCRIPTION: SEQ ID NO:1: Ser Glu Glu Pro Pro Ile Ser Leu Asp Leu Thr Phe His Leu Leu Arg Glu Gin Leu Ala Gin Gin Ala His 25 Glu Val Leu Glu Met Ser Asn Arg Lys Leu Ala Arg Ala Met Glu Ile lle INFORMATION FOR SEQ ID NO:2: WO 96/02569 PCTIS95/08867 SEQUENCE CHARACTERISTICS: LENGTH: 30 amino acids TYPE: amino acid STRANDEDNESS: single TOPOLOGY: unknown (ii) MOLECULE TYPE: peptide (xi) SEQUENCE DESCRIPTION: SEQ ID NO:2: Xaa Xaa Xaa Xaa Xaa Xaa Xaa Xaa Xaa Xaa Xaa Xaa Xaa Xaa 1 5 10 Xaa Xaa Ala Xaa Xaa Glu Xaa Xaa Xaa Xaa Xaa Xaa Xaa Xaa 25 Xaa Xaa INFORMATION FOR SEQ ID NO:3: SEQUENCE CHARACTERISTICS: LENGTH: 30 amino acids TYPE: amino acid STRANDEDNESS: single TOPOLOGY: unknown (ii) MOLECULE TYPE: peptide (xi) SEQUENCE DESCRIPTION: SEQ ID NO:3: Pro Pro Ile Ser Xaa Asp Leu Thr Phe His Leu Leu Arg Xaa Xaa Xaa 1 5 10 1 WO 96/02569 PCT/US95/08867 Glu Xaa Ala Arg Xaa Glu Xaa Xaa Xaa Xaa Gin Ala Xaa Xaa 25 INFORMATION FOR SEQ ID NO:4: SEQUENCE CHARACTERISTICS: LENGTH: 41 amino acids TYPE: amino acid STRANDEDNESS: single TOPOLOGY: unknown (ii) MOLECULE TYPE: peptide (xi) SEQUENCE DESCRIPTION: SEQ ID NO:4: Ser Gin Glu Pro Pro Ile Ser Leu Asp Leu Thr Phe His Leu Leu Arg Glu Val Leu Glu Met Thr Lys Ala Asp Gin Leu Ala Gin Gin Ala His Ser Asn Arg Lys Leu Leu Asp Ile Ala

Claims (6)

1. A method of screening for ligand inhibitors comprising: incubating a candidate ligand inhibitor with an isolated CRF/CRF-binding protein complex; and measuring free CRF by an assay.

2. The method according go Claim 1, wherein free CRF is measured by an in vitro ligand immunoradiometric assay.

3. The method according to Claim 1, wherein free CRF is measured by a biological assay.

4. A method according to any one of Claims 1 to 3 for screening for ligand inhibitors of a CRF/CRF-binding protein complex, comprising: a a. contacting CRF with CRF-binding protein in the presence of a candidate S. "ligand inhibitor in an aqueous solution; Sb. adding a nonionic detergent to the solution of step c. separating the nonionic detergent from the aqueous solution; and d. detecting either the amount of bound CRF is the aqueous solution or the amount of free CRF in the nonionic detergent, and thereby determining whether the candidate ligand inhibitor disrupts the CRF/CRF-binding protein complex. a A method according to Claim 4 wherein step is performed at a temperature above the cloud point of the nonionic detergent.

P:\OPER\JMS\31283-95.CLM 1/4/99 52

6. A method according to Claim 4 or 5 wherein the nonionic detergent is octylphenoxypolyethoxyethanol. Dated this I1st day of April, 1999. University of Reading AND Neurocrine Biosciences, Inc. AND The Salk Institute By their Patent Attorneys Davies Collison Cave 0 a