This application is divided out of original Indian application no. 885/KOLNP/2006 filed
on 10/04/2006 being the National phase entry of PCT/US04/29888
1. CROSS-REFERENCE TO RELATED APPLICATIONS
Priority is claimed under 35 U.S.C. 119(e) to U.S. provisional patent applications
Serial No. 60/501,864, filed on September 10,2003; Serial No. 60/559,898, filed April 5,
2004; and Serial No. 60/603,187, filed August 20,2004. The contents of these provisional
applications are hereby incorporated by reference, in their entireties.
2. FIELD OF THE INVENTION
The present invention relates to compositions and methods for modulating the growth
and repair of the central nervous system (CNS), including processes such as neuronal
survival, axonal growth and synaptic plasticity. More specifically, the invention relates to
compounds (including cyclic peptides and peptidomimetic compounds) that are either
agonists or antagonists of a family of receptors, known as Trk receptors, that are expressed
on the surface of neuronal cells and which regulate such processes of CNS growth and repair.
Further, the invention relates to methods for promoting CNS growth and repair using a p75
binding agent.
3. BACKGROUND OF THE INVENTION
Injury to the central nervous system (CNS) can have devastating consequences due to
the poor regenerative capacity of neurons in that environment. This contrasts markedly with
the comparatively good regenerative capacity of neurons in the peripheral nervous system.
See, for example, Horner & Gage, Nature 2000,407:963-970.. Numerous diseases, such as
Alzheimer's disease, Parkinson's disease, stroke, head and spinal cord trauma to name a few,
are all associated with damage to the CNS mat is often severe, even debilitating, long lasting
or even permanent No cure is presently available for these conditions, and even palliative
treatments are lacking.

3.1. Neurotrophins
It is now understood that the growth and regeneration of neurons is regulated at least
in part by certain polypeptide growth factors, known as neuroptrophins or "NTs," which bind
to and activate cell surface receptors having an intrinsic tyrosine kinase activity. Upon
neurotropbin binding, these receptors are believed become autophosphorylated on one or
more amino acid residues and subsequently, associate with intracellular molecules important
for signal transduction. For areview, see Ulrich & Schlessinger, Cell 1990,61:203-212.
The first identified neurotrophin is known in the art as nerve growth factor (NGF) and
has a prominent effect on developing sensory and sympathetic neurons of the peripheral
nervous system. See, Levi-Montalcini & Angeletti, Physiol. Rev. 1968,48:534-569;
Thoenen et al., Rev. Physiol Biochem. Pharmacol 1987,109:145-178; Thoenen & Barde,
Physiol Rev. 1980,60:1284-1325; Whittemore & Seiger, Brain Res. 1987,434:439-464;
Angeletti & Bradshaw, Proc. Natl. Acad. Sci. U.SJ.. 1971, 68:2417-2420; Angeletti et al,
Biochemistry 1973,12:100-115. NGF orthologs have also been isolated and characterized in
a number of other species, including mice, birds, reptiles and fishes (Scott et al, Nature
1983,302:538-540; Schwartz et al, J. Neurochem. 1989,52:1203-1209; and Hallbook et
al, Neuron 1991,6:845-858.
A number of other NTs are also known in the art. These include brain-derived
neurotrophic factor (BDNF), which is also known as neurotrophin-2 (NT-2). See, Leibrock
et al, Nature 1989, 341:149-152. Still other NTs include a factor originally called neuronal
factor (NF) and now commonly referred to as neurotrophin-3 or "NT-3" (Ernfors et al, Proc.
Natl. Acad. Sci. U.S.A. 1990, 87:5454-5458; Hohnetal, Nature 1990, 344:339;
Maisonpierre et al, Science 1990,247:1446; Rosenthal et al, Neuron 1990,4:767; Jones &
Reichardt, Proc. Natl Acad. Set U.S.A. 1990,87:8060-8064; and Kaisho et al, FEBSLett.
1990,266:187). Neurotrophins-4 and -5 (NT-4 and NT-5) are also known. See, Hallbook et
al, Neuron 1991,6:845-858; Berkmeier et al, Neuron 1991,7:857-866; Ip et al, Proc.
Natl Acad. Sci. U.SJ.. 1992,89:3060-3064. See also, U.S. Patent No. 5,364,769 issued
November 15,1994 to Rosenthal. Because it was subsequently seen to be a mammalian
ortholog of the Xenqpus NT-4 described by Hallbrook et al, supra, the mammalian NT-5
molecule described by Berkmeier et al, supra, is also commonly referred to as NT-4/5. An
alignment of NT's BDNF, NT4, NT3, and NGF is provided in Figure 1.

3.2. Trk Receptors
Neurotrophins mediate their effect through a family of receptor tyrosine kinases that
are expressed on the surface of neuronal cells and referred to collectively as Trk-receptors.
At least three different Trk-receptors are known and have been described in the art: TrkA,
TrkB and TrkC. For a review, see U.S. Patent Nos. 5,844,092; 5,877,016; 6,025,166;
6,027,927; and 6,153,189 all by Presta et al Although the structure and sequences of the
different Trk-receptors are similar, alternate splicing increases the complexity of this family
giving rise to several different isoforms of each receptor. An alignment of the different Trk-
receptor amino acid sequences is provided here at Figure 2A-2C setting forth the consensus
sequences and boundaries for the various domains of each receptor. See also, Figures 16A-
16C in U.S. Patent No. 5,877,016.
Each of the different Trk-receptors exhibits particular binding affinity for the
different neurotrophins, although there is Some overlap. Hence, TrkA is believed to bind not
only NGF, but also NT-3 and NT-4/5 (but not BDNF). TrkB is believed to bind BDNF, NT-
3, NT-4 and NT-4/5, but not NGF. By contrast, TrkC is believed to bind only NT-3 and not
any of the other neurotrophins.
A number of studies have validated the Trk-receptors as therapeutic targets for brain
repair. See, for example, Liu et al, J. Neurosci. 1999,19:4370-4387; Menei et al, Eur. J.
Neurosci. 1998,10:607-621; and Kobayashi et/al., J. Neurosci. 1997,17:9583-9595. The
Trk-receptors and their ligands have also been studied using X-ray crystallography to obtain
three-dimensional structures of the ligand-receptor binding .complexes. Wiesmann et al,
Nature 1999,401:184-188; Banfield et al, Structure (Camb) 2001,9:1191-1199. These and
other studies suggest that neurotrophin binding to the Trk-receptors induces dimerization of
receptor monomers, resulting in an increase of the receptors' intrinsic tyrosine kinase
activity. This increased activity triggers, in turn, signaling cascades that are believed to be
beneficial to neurons by promoting neuronal survival, axonal growth, and synaptic plasticity.
Snider, Cell 1994,77:627-638; Kaplan & Miller, Curr. Opin. Neurobiol 2000,10:381-391.
There has therefore been considerable recognition that therapeutic compounds which
target and activate Trk-receptors (i.e., Trk-receptor "agonists") would be beneficial and
desirable. See, for example, Lindsay et al, Exp. Neurol 1993,124:103-118; Olson,

Neurochein. Int. 1994,25:1-3. Moreover, increased levels of certam neurotrophins (e.g.,
BDI- F) are also associated with medical conditions such as epilepsy (Binder et al, Trends
Neurosci. 2001,24:47-53). Hence, even compounds that inhibit Trk-receptor activity (i.e.,
Trk-receptor "antagonists") would be beneficial. Despite this long felt need, such
compounds have been elusive at best. As large-molecules, the therapeutic delivery of
effective levels of neurotrophins themselves presents considerable, possibly insuimountable,
challenges. Moreover, natural neurotrophins may interact with other receptors, such as the
p75 receptor in neurons, which is associated with neuronal apoptosis and growth cone
collapse. Lee et al, Curr. Opin. Neurobiol 2001,11:281-286.
However, previous efforts to design peptidornimetic agonists and/or antagonists of
Trk-receptors have also been unsuccessful. For example, cyclic peptides derived from loop 1
of the neurotrophin NGF have been reported to moderately mimic the survival activity of
NGF. However, these peptides appear to function in a p75, rather Trk-receptor, dependent
manner. Long et al, J. Neurosci. Res. 1997,48:1-17. Some NGF loop 4 cyclic peptides are
said to show NGF-like survival activity that is blocked by a Trk antagonist. However, the
maximal survival response induced by those peptides is reported to be only 10-15% of the
maximal response promoted by the NGF neurotrophin itself. See, Xie et al, J. Biol. Chem.
2000,275:29868-29874; and Maliartchouk et al, J. Biol Chem. 2000,275:9946-9956.
Bicyclic and tricyclic dimeric versions of BDNF loop 2 peptides have been shown to have
BDNF-like activity. Again, however, the maximal survival response they induce is reported
to be only 30% of the maximal response promoted by the natural neurotrophin. O'Leary et
al, J. Biol Chem. 2003,278:25738-25744 (Electronic publication May 2,2003).
There continues to exist, therefore, a long felt need for compositions that can
modulate (i.e., increase or inhibit) neuronal growth and recovery. There also exists a need
for processes and methods (including therapeutic methods) that effectively modulate
neuronal growth and recovery.
3.21. The p75 receptor Neurotrophin Receptor
The p75 receptor is known to play roles in signaling complexes for neuronal
apoptosis and growth inhibition. Barker, Neuron 2004,42:529-533. The p75 receptor is a
member of the tumor necrosis factor (TNR) superfamily and is characterized by cysteine-rich

domains (CRDs) in its extracellular portion. These CRDs are required for neurotrophin
binding, and p75 receptor serves as a low affinity receptor for neurotrophins such as NGF,
BDNF, NT-3, and NT-4. Huang and Reichardt, Annu. Rev. Biochem. 2003,72:609-642.
NGF, BDNF, NT-3 and NT-4 can effectively compete with each other for binding to p75
receptor. In inhibitory environments, these neurotrophins can be used to compete out each
other's binding to p75 receptor in order to reveal responses that depend solely on Trk
signaling. Barker and Shooter, Neuron 1994,13:203-215.
3.22. The p75 receptor and the NGF TDIKGKE Motif
It is known that the TDIKGKE motif that constitutes the first {3 hairpin loop of NGF
plays a crucial role in the binding of NGF to the p75 receptor. He and Garcia, Science 2004,
304:870-875; Ibanez et al, Cell 1992,69:329-341. Furthermore, constrained TDIKGKE
motifs interact with the p75 receptor and are expected to compete for neurotrophin binding to
this receptor. Longo et al, J. Neurosci. Re's. 1997,48:1-17.-
The cyclic peptides and peptidomimetic compounds derived from loop 1 of NGF
have been reported to moderately .mimic NGF's neuron growth-promoting activity (see U.S.
Patent No. 6,017,878 to Saragovi et al.), and these peptides appear to function in a p75
receptor-dependent manner (Longo et al., J. Neurosci. Res. 1997,48:1-17). Some NGF
loop 4 cyclic peptides are said to show NGF-like neuron growth promotion that is blocked by
a Trk antagonist. However, the maximal response induced by those peptides is reported to be
only 10-15% of the maximal response promoted by the NGF neurotrophin itself. See Xie et
al., J. Biol. Chem. 2000,275:29868-29874; and Maliartchouk et al, J. Biol. Chem. 2000,
275:9946-9956. Bicyclic and tricyclic dimeric versions of BDNF loop 2 peptides have been
shown to have BDNF-like activity. Again, however, the maximal response they induce is
reported to be only 30% of the maximal response promoted by the natural neurotrophin.
O'Leary et al, J. Biol Chem. 2003,278:25738-25744 (Electronic publication May 2,2003).
3.3. Inhibitory Signals
The central nervous system's limited ability to repair injuries is thought to be at least
partly due to the presence of inhibitory products that prevent axonal regeneration - including
inhibitors associated with damaged myelin (Berry, Bibl. Anat. 1982,23:1-11). Indeed,

biochemical studies on central myelin have identified two protein fractions that contain
inhibitory activity for cell spreading (Caroni & Schwab, J. Cell Biol. 1988,106:1281-1288)
and monoclonal antibodies that bind to those fractions enhance the growth of cultured
sensory and sympathetic neurons in what are otherwise non-permissive substrates for neurite
growth (Caroni & Schwab, Neuron. 1988,1:85-96). Studies with these same antibodies in
lesioned animals have also shown that functional recovery can be obtained by blocking the
function of inhibitory molecules associated with myelin (Bregman et al., Nature 1995,
378:498-501; Schnell & Schwab, Nature 1990,343:269-272). A much more robust
regeneration response has been obtained in mice immunized with whole myelin (Huang et
al, 1999) further demonstrating that CNS recovery and repair can be enhanced in vivo, by
blocking inhibitory factors.
At least three myelin derived molecules are known that are potent inhibitors of axonal
growth: the myelin-associated glycoprotein, which is also referred to as "MAG" (described
by McKerracher et al., Neuron 1994,13:805-811; and by Mukhopadhyay et al., Neuron
1994,13:757-767); Nogo-A (see, Chen et al, Nature 2000,403:434-439; GrandPre et al,
Nature 2000,403:439-444; and Prinjha et al, Nature 2000,403:383-384) and the
oligodendrocyte myelin glycoprotein (Wang et al, Nature 2002,417:941-944). The Nogo
receptor (also referred to as "NgR"), the ganglioside GTlb and the p75 neurotrophin receptor
(also referred to as "p75NTR" or "p75NTR") have been implicated in mediating responses to
all three of these inhibitory molecules. Specifically, binding to NgR is said to be required for
inhibitory activity by all three inhibitors MAG, Nogo-A and oligodendrocyte glycoprotein
(Domeniconi et al, 2002; Liu et al, 2002; Wang et al, 2002b). However, MAG can also
bind directly to the GTlb receptor (Vyas & Schnaar, Biochimie 2001, 83:677-682).
Moreover, antibody induced clustering of GTlb receptor can mimic the inhibitory response
produced by MAG (see, Vinson et al, J. Biol Chem. 2001,276:20280-20285; and Vyas et
al,Proc. Natl Acad. Sci. U.S.A. 2002,99:8412-8417).
The p75 receptor is the signaling component of a multimeric receptor complex than
can bind all three myelin receptors. See Domeniconi et al, Neuron 2002,35:283-290; Liu et
al, Science2002,297:1190-1993; Wang et al,Nature2002,417:941-944. Interactions
between the GTlb and p75NTR receptors have been reported (Yamashita et al, 2002), as have
interactions between the NgR and p75NTR receptors (see, Wang et al, 2002a; and Wong et

al, 2002). Such interactions with p75NTR are thought to be important in the transmission of
inhibitory signals (e.g., from MAG, Nogo-A and/or oligodendrocyte glycoprotein) across the
cell membrane. For example, interactions of MAG or a Nogo-A peptide with cells that
express NgR increases association of p75NTR with Rho-GDI, and induces the release of RhoA
from that complex (Yamashita & Tohyma, Nat. Neurosci. 2003,6:461-471). This step is a
pre-requisite for activation of RhoA and inhibition of growth (Id), and the inhibition of RhoA
and/or Rho kinase (a downstream effector of RhoA) effectively circumvents inhibitory
activity, e.g., of myelin in cultured neurons (see, for example, Dergham et al, J. Neurosci.
2002,22:6570-6577; Fournier et al, J. Neurosci. 2003,23:1416-1423; and Lehmann et al,
J. Neurosci. 1999,19:7537-7547).
As noted above, the various neurotrophins (e.g., NGF, BDNF, NT-3 and NT-4/5) do
have dramatic effects on neuronal survival and axonal growth during development. It has
been recently suggested that neurotrophins and inhibitory molecules (for example, MAG,
Nogo-A and oligodendrocyte glycoprotein) may have an opposing effect on the coupling of
p75NTR receptor to Rho-GDI (see, Yamashita & Tohyania, Nat. Neurosci. 2003, 6:461-467).
Nevertheless, it has not as of yet been possible to promote robust, long range axonal
regeneration using neurofrophins. This is believed to be at least partly due to the inability of
neurofrophins to effectively counteract inhibitory signals such as those described above. For
example, the treatment of cultured neurons with neurotrophins such as NGF, BDNF or
GDNF (glial derived neurotrophic factor) does not normally counteract the inhibitory activity
of myelin unless the neurons are first "primed" by exposure to the neurotrophin for several
hours before exposure to the inhibitory signal (Cai et al, Neuron 1999,22:89-101). Such
priming, however, is of limited effect, time consuming, cumbersome to apply, and
impractical for clinical and other in vivo applications. Moreover (and as noted above), the
therapeutic delivery of neurofrophins themselves, which are large molecules, presents
considerable and possibly insurmountable technical challenges. Furthermore, neurotrophins
may be compromised in their ability to promote regeneration because they bind to the
inhibitory complex through their interaction with the p75 receptor. Neurofrophins, which are
bound to p75 receptor, cannot activate Trk receptors to overcome inhibitory signaling and to
promote neuronal growth.

Hence, there additionally exists a need for compounds that can effectively modulate
the effects of inhibitory signals on neuronal growth and recovery including compounds that
effectively modulate effects of inhibitory signals such as those produced by MAG, Nogo-A,
oligodendrocyte glycoprotein, NgR, GTlb, p75 NTR and/or downstream effectors of these
signaling molecules. In particular, there exists a need for compounds that can effectively
counteract such inhibitory signals, and/or stimulate neuronal growth and recovery. There
also exists a need for processes and methods (including therapeutic methods) that modulate
effects of such inhibitory signals and, in particular, for processes and methods that counteract
such inhibitory signals and/or stimulate neuronal growth and recovery.
The citation and/or discussion of a reference in this section and throughout the
specification is provided merely to clarify the description of the present invention and is not
an admission that any such reference is "prior art" to the invention described herein.
4. SUMMARY OF THE INVENTION
The present invention provides at least a partial solution to the above-mentioned
problems in the art by providing compounds and formulations thereof which modulate (e.g.,
enhance or inhibit) activity mediated by a Trk receptor such as TrkA, TrkB or TrkC. For
example, in one embodiment the invention provides compounds that are Trk antagonists and,
as such, inhibit Trk mediated activity. In other embodiments, the invention provides
compounds that are Trk agonists and, as such, enhance or increase Trk mediated activity.
As noted above, Trk receptors and their ligands (i.e., neurotophins such as NGF,
BDNF, NT-3, NT-4, NT-5 and NT-4/5) are associated with the growth and repair of the
central nervous system (CNS). As such, Trk modulator compounds of the present invention
can be used to modulate such processes, including processes of neuronal growth and
survival, axonal growth, neurite outgrowth, and synaptic plasticity. In one aspect, therefore,
the present invention provides methods (including therapeutic methods) that use Trk
modulator compounds of the invention to modulate (e.g., enhance or inhibit) such processes.
In one particular embodiment, the invention provides cyclic peptide compounds that
modulates Trk receptor mediated activity. These cyclic peptides preferably comprise, within

a cyclic peptide ring, the amino acid sequence: Arg-Gly-Glu. La a more particular
embodiment, the cyclic peptide comprises the formula:

In Formula I, the elements Y1 and Y2 are independently selected amino acids with a
covalent bond formed between Y1 and Y2. The element X1 and X2 are optional and, if
present, are independently selected amino acids or sequences of amino acids joined by
peptide bonds. Preferably X1 and/or X2 are each between zero and about 10 amino acids in
length, and are more preferably about 1,2,3,4 or 5 amino acids in length. Moreover, X1 and
X2 are also preferably selected so that the size of the cyclic peptide ring ranges from about 5
to about 15 amino acids in length, and is more preferably between about 5-10 amino acids in
length.
The invention further provides, in particular embodiments, cyclic peptides having the
formula:
where Y1 and Y2 are as described above, for Formula I Particularly preferred cyclic
peptides of the invention are ones comprising the amino acid sequence: CSRRGEC (SEQ ID
NO:l), N-Ac-CSRRGEC-NH? (SEQ ID NO:2), CARRGEC (SEQ ID NO:3), N-Ac-

CARRGEC-NH2 (SEQ ID N0:4), CFHRGEC (SEQ ID N0:5), N-Ac-CFHRGEC-NH2 (SEQ
ID N0:6), CSBRGEC (SEQ ID N0:7), N-Ac-CFHRGE-NH2, (SEQ ID N0:8), CRGEC (SEQ
ID NO:9), N-Ac-CRGEC-NH2 (SEQ ID NO: 10). N-Ac-KRGED-NH2 (SEQ ID NO:ll), H-
C(O)CRGEC-NH2 (SEQ ID NO:12), CH.-SO2-NH-CRGEC-NH2 (SEQ ID NO:13), N-Ac-
CRGEC-Y-NH2 (SEQ ID NO:14), H-C(O)-CRGEC-Y-NH2 (SEQ ID NO:15) and CH3-SO2-
NH-CRGEC-Y-NH2 (SEQ ID NO:16), (where the underlined portion of each amino acid
sequence indicates that portion of the peptide that is cyclized).
Preferred cyclic peptides of the above formulas and sequences are Trk antagonists.
However, the invention also provides, in other embodiments, cyclic peptides that are Trk
agonists. Such cyclic peptides preferably have the formula:
(Formula H)
In Formula II, above, the elements Y1 and Y2 are independently selected amino acids
with a covalent bond formed between Y1 and Y2. The elements Z1, Z2 and Z0 are optional
and, if present, are independently selected amino acids or sequences of amino acids joined by
peptide bonds. Preferably, Z1, Z2 and/or Zo are each no more than about ten amino acids in
length, and are more preferably only about 1,2,3,4,5 or 10 amino acids in length.
Moreover, the lengths of Z\, Z2 and/or Zo are preferably selected so that the size of the cyclic
peptide ring ranges from about 10-50 amino acids in length, and more preferably from about
10-25 or from about 15-20 amino acids in length. In particularly preferred embodiments, the
elements Z1, Z2 and Z0 are selected such that the tandem Arg-Gly-Glu sequences in
Formula I adopt a conformation where they are adjacent and anti-parallel to each other.
In preferred embodiments, the invention provides cyclic peptides according to
Formula II that have the formula

where the elements Y1 and Y2 are as set forth, supra, for Formula II. Particularly
preferred peptides according to Formula II, which are also a part of the present invention, are
cyclic peptides comprising the amino acid sequence: CSRRGELAASRRGELC (SEQ ID

N0:17) and N-Ac-CSKRGELAASRRGELC-NH2 (SEQ ID N0:18), (where the underlined
portion of each amino acid sequence indicates that portion of the peptide that is cyclized).
In accordance with the invention, cyclic peptides are provided comprising, within a
cyclic ring of the cyclic peptide, the D amino acid sequence:
dGm-Gly-dArg
wherein the cyclic peptide modulates Trk receptor mediated activity. Preferred cyclic
peptides that modulate Trk receptor mediated activity comprising D amino acid sequences
are c[dLdEGdRdRdSdLdEGdRdRdS] (SEQ ID NO:40), (where the bracketed portion of the
amino acid sequence indicates that portion of the peptide that is cyclized by a peptide
bond)and Ac-dCdLdEGdRdRdSdAdAdLdEGdRdRdSdC-MI2 (SEQ ID NO:41).
In a further embodiment, the invention provides the cyclic peptide having the amino
acid sequence c[SRRGELSRRGEL] (SEQ ID NO:39).
In addition to the cyclic peptides, the invention also provides methods for identifying
other compounds (i.e., "candidate compounds") mat modulate Trk receptor mediated activity
or are likely to modulate such activity. These methods involve comparing a three-
dimensional structure of the candidate compound with the mree-dimensional structure of a
cyclic peptide of the invention. Similarity between the structure of the candidate compound
and the structure of the cyclic peptide is indicative of the candidate compound's ability to
modulate Trk receptor mediated activity. Hence, a candidate compound having a
substantially similar structure to the three-dimensional structure of the cyclic peptide is likely
to be a compound which modulates Trk receptor mediated activity.
The above methods are ideally suited for identifying peptidomimetic compounds that
modulate Trk receptor mediated activity. Accordingly, the invention provides
peptidomimetic compounds that are Trk modulators, and such compounds are considered
another aspect of the invention. In particular, the peptidomimetic compounds of the
invention are compounds having a three-dimensional structure that is substantially similar to
the three-dimensional structure of a cyclic peptide of the invention (i.e. a cyclic peptide that
modulates Trk mediated activity and comprises, within a cyclic ring thereof, the amino acid
sequence Arg-Gly-Glu).
The invention additionally provides methods, including therapeutic methods, that use
cyclic peptides and peptidomimetic compounds to modulate Trk mediated activity. In one

such embodiment, the invention provides methods for inhibiting Trk mediated activity. Such
methods involve contacting a cell (in vitro or in vivo) with an amount of a cyclic peptide or
peptidomimetic compound of the invention that inhibits Trk mediated activity. The amount
of the cyclic peptide or peptidomimetic compound contacted to the cell should be an amount
that effectively inhibits the Trk receptor mediated activity.
In another embodiment, the invention provides methods for enhancing Trk mediated
activity. Such methods involve contacting a cell (in vitro or in vivo) with an amount of a
cyclic peptide or peptidomimetic compound of the invention that enhances Trk mediated
activity. The amount of the cyclic peptide or peptidomimetic compound contacted to the cell
should be an amount that effectively enhances the Trk receptor mediated activity.
Examples of Trk mediated activities that can be modulated (e.g., enhanced or
inhibited) by such methods include: neuronal growth and survival, axonal growth, neurite
outgrowth and synaptic plasticity and well as other processes of central nervous system
(CNS) growth and/or repair. Accordingly, the invention additionally provides methods for
enhancing growth or repair of the central nervous system in an individual. These methods
involve adroinistering to the individual an amount of a cyclic peptide or a peptidomimetic
compound of the invention that enhances Trk mediated activity. The amount of the cyclic
peptide or peptidomimetic compound administered should be an amount that effectively
enhances CNS growth or repair.
The invention additionally provides methods that use Trk agonists and antagonists to
modulate responses that inhibit CNS growth and repair, including responses that normally
inhibit processes such as neuronal growth, neuronal survival, axonal growth, neurite
outgrowth and synaptic plasticity. In particular, Trk agonists and antagonists of the invention
can be used to modulate inhibitory factors and/or inhibitory signals generated by such
factors. Examples include factors associated with myelin, including the myelin associated
glycoprotein (MAG), Nogo-A and the oligodendrocyte myelin glycoprotein, In general, the
invention provides methods using- Trk agonists and/or antagonists to modulate a CNS
inhibitor response mediated by a signal cascade with one or more components that are
themselves modulated by a factor or factors involved in signaling by a Trk receptor. These
include, for example, components such as Rho that are modulated by protein kinase A (PKA)
and/or by phosphoinositide-3 kinase (PI3K). In preferred embodiments, therefore, the

invention provides methods for reducing such "CNS inhibitor" responses by contacting a cell
with a Trk agonists {e.g., a cyclic peptide or peptidomimetic) of the invention in an amount
that is effective for reducing the CNS-inhibitor response. The invention also provides
methods for reducing a CNS inhibitor response in an individual, by administering to the
individual an amount of a Trk agonists (e.g., a cyclic peptide or peptidomimetic) of the
invention in an amount that effectively reduces the CNS inhibitor response.
In still other embodiments, the invention provides pharmaceutical compositions that
can be used in therapeutic methods, such as those described above. Such pharmaceutical
compositions comprise an amount of a cyclic peptide or peptiomimetic compound of the
invention, along with one or more carriers, diluents or excipients that are pharmaceutically
and/or physiologically acceptable.
Further, the present invention is based on the discovery that agents which interfere
with the binding of neurotrophins to the p75 receptor promote CNS neuron growth in an
inhibitory environment.
Trk receptors and their ligands (i.e. neurotrophins such as NGF, BDNF, NT-3, NT-4
and NT-5) are associated with the growth and repair of CNS neurons. As such, when
neurotrophins bind to and activate Trk receptors, Trk activity triggers signaling cascades
which promote neuronal growth. However, the p75 receptor binds neurotrophins with low
affinity and, when the p75 receptor is engaged in an inhibitory complex, this interaction
compromises the ability of neurotrophins to promote CNS neuron growth. The invention
provides methods for promoting CNS neuron growth using a p75 receptor binding agent,
which interferes with the binding of a neurotrophin to 'the p75 receptor.
According to the present invention, a method is provided for promoting CNS neuron
growth in an inhibitory environment, which comprises administering to an individual a
therapeutically effective amount of a p75 receptor binding agent In one embodiment, the p75
receptor binding agent includes a neurotrophin binding motif or a peptidomimetic thereof. In
a particular embodiment, the p75 receptor binding agent comprises a cyclic peptide or
peptidomimetic comprising, within a cyclic ring thereof, the amino acid sequence Thr-Asp-
IIe-Lys-Gly-Lys-Glu (TDDCGKE) (SEQ ID NO:42). A preferred p75 receptor binding agent
is N-Ac-CTDIKGKEC-NH2 (SEQ ID NO:43). The individual is preferably a mammal and
more preferably a human.

The present invention provides methods for promoting CNS neuron growth in an
inhibitory environment, which comprise administering to an individual a therapeutically
effective amount of a p75 receptor binding agent in combination with a neurotrophin. In one
embodiment, the neurotrophin is selected from the group consisting of NGF, BDNF, NT-3,
NT-4 and NT-5. In a further embodiment, the p75 receptor binding agent is administered in
an amount about 10 to about 100 fold greater than the neurotrophin. In an aspect of the
invention, the p75 receptor binding agent is a neurotrophin that interferes with another,
different neurotrophin for binding to the p75 receptor, but does not interfere with binding of
the another, different neurotrophin to a Trk receptor expressed on an injured neuron. In a
particular aspect, the p75 receptor binding agent is NGF and the neurotrophin is BDNF
wherein NGF is administered in an amount about 10 to about 100 fold greater than BDNF.
The individual is preferably a mammal, and more preferably, a human.
5. BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 shows the amino acid sequence alignment of NTs BDNF (SEQ ID NO:19),
NT4 (SEQ ID NO:20), NT3 (SEQ ID NO:21), and NGF (SEQ ID NO:22). Mature chains
are denoted by bold lettering and the RGE motif is underlined.
Figures 2A-2C show an alignment of the full length amino acid sequences of human
TrkA (SEQ ID NO:23), TrkB (SEQ ID NO:24) and TrkC (SEQ ID NO:25) receptors.
Consensus sequences for the receptors are boxed, and the boundaries of the receptors'
various domains are marked by vertical lines. See also, U.S. Patent No. 5,844,092 by Presta
et al
Figures 3A-3B illustrate representative backbone modifications that may be present
within a peptidomimetic. See also, Figures 4A and 4B in WO 01/53331.
Figure 4 illustrates representative unusual amino acids and dipeptide surrogates that
maybe incorporated into apeptilomimetic. See also, Figure 5 in WO 01/53331.

Figure 5 illustrates representative secondary structure mimics that maybe
incorporated into a peptidomimetic. See also, Figure 6 in WO 01/53331.
Figures 6A-6C illustrates the analysis of aNT/Trk crystal structure to identify linear
regions of the ligand that interact with the Trk receptor. Figure 6A shows a ribbon image of
the crystal structure (reported by Banfield et al, Structure (Camb) 2001,9:1191-1199) of an
NT-4 dimer (denoted as chain a1 and a2 in a complex with two membrane proximal Ig
domains from the TrkB receptor (denoted as chain b1 and b2). Figure 6B shows the results
of examining this crystal structure (solid line) and theNGF/TrkA crystal structure (reported
by Wiesman etal, Nature 1999,401:184-188) (dotted line) for linear peptide sequences
(LIPs) that make contact between the a1 and b2 chains. Figure 6C shows the results of
examining the NGF (dotted line) and NT-4 (solid line) crystal structures for LPs that make
contact between the a1 and b1 chains.
Figures 7A-7D show data from experiments where cerebellar neurons were cultured
over monolayers of 3T3 cells in control media or in media supplemented with NT-4 or
BDNF in the presence of various peptides for 18 hours before being fixed and stained for
GAP-43. The mean length of the longest neurite was determined from between about 100-
120 neurons under each culture condition. Figure 7A shows data from experiments testing
the effects of various concentrations of NT-4 and BDNF on neurite outgrowth. Figure 7B
shows data from experiments testing the effects of increasing concentrations of the cyclic
peptide N-Ac-CRGEC-NH2 in control media and in media containing 5 ng/ml BDNF or
5 ng/ml NT-4, as indicated. Figure 7C shows data from experiments testing the effects of
the NT-4 derived cyclic peptide N-Ac-CSRRGEC-NH2 (SEQ ID NO:26), the NT-3 derived
cyclic peptide N-Ac-CSHRGEC-NH? and the NGF derived cyclic peptide N-Ac-CFHRGEC-
NH2 in cerebellar neurons cultured with 5 ng/ml BDNF. Figure 7D shows data from
experiments identical to that shown in Figure 7C, but where the cerebellar neurons are
cultured with 5 ng/ml NT-4.
Figure 8 illustrates results from experiments were cerebellar neurons were cultured
over monolayers of 3T3 cells in control media or in media supplemented with NT-4, BDNF,

FGF2 (all at 5 ng/ml), with the CGI receptor agonist WIN55,2122-2 (0.2 uM) or over
monolayers of 3T3 cells the express fransfectedN-cadherin (NGAD) on their cell surface.
The experiments were performed, and data plotted, in the presence and absence of: (a) the
cyclic peptide N-Ac-CRGBC-NB2 at 440 μM; (b) the cyclic peptide N-Ac-CSRRGEC-NH2
at 125 μM; and (c) the linear peptide N-Ac-SRRGELA-NH2 (SEQ ID NO: 27) at 125 μM.
Figures 9A-C show modeled structures of the BAG peptide. Figure 9A shows the
native structure of the SRRGELA motif from one monomer of the NT-4 dimer in the NT-
4/TrkB crystal structure. The native structure of the ASRRGEL (SEQ ED NO:28) motif from
the partner NT-4 monomer in that crystal structure is shown in Figure 9C. A modeled
structure of the BAG peptide N-Ac-CSRRGEIAASRRGELC-NH2 incorporating these
"tandem-repeat" motifs is shown in Figure 9B.
Figures 10A-10B show results from neurite outgrowth experiments where cerebellar
neurons were cultured in media supplemented with a range of concentrations of the BAG
peptide N-Ac-CSRRGELAASRRGELC-NH2. Figure 10A shows the mean value of
absolute neurite lengths determined from 100-120 neurons sampled in a single experiment.
Figure 10B shows a histogram comparing effects of the BAG peptide (6 uM) with the
response to established growth promoting agents, including NT-4, BDNF, and FGF2 (all at
5 ng/ml) as indicated.
Figure 11 shows results from neurite outgrowth experiments where cerebellar
neurons were cultured in control media or in media supplemented with: (a) the BAG peptide
N-Ac-CSRRGELAASRRGELC-NH2 at 6 μM; (b) the TrkB antagonist peptide N-Ac-
CSSRGEC-NH2 (SEQ ID NO:29) at 125 uM; (c) the Trk specific tyrosine kinase inhibitor
K252a at 100 nM; or the linear version of the TrkB antagonist peptide N-Ac-SRRGELA-
NEfe at 125 μM, as indicated.
Figure 12 shows a bar graph depicting results from neurite outgrowth experiments
testing the effects of various agents. In particular, cerebellar neurons were cultured over
monolayers of N-cadherin expressing 3T3 cells in media supplemented with a soluble MAG-

Fc fusion construct at final concentrations of 0,5 or 25 p.g/ml (as indicated below each bar in
the graph). Experiments were done in control media (z.e., in media supplemented with
MAG-Fc only) and in media additionally supplemented with the Eho kinase inhibitor
Y27632 (10 uM final concentration) BAG polypeptide (6 uM final concentration) or BDNF
(5 ng/ml final concentration). Cultures were maintained for 22 hours before being fixed and
stained for GAP-43. The mean length of the longest neurite was determined from
measurements of between about 100-120 neurons under each culture condition. Each
column of the graph depicts pooled results from a number of independent experiments
(indicated above the column), and the bars indicate standard error of the mean (SEM).
Figure 13 illustrates a dose response curve for BAG polypeptide on the MAG-Fc
response in cultured neurons. In particular, each data point indicates the mean length
measured from about 120-150 neurons when cultured over monolayers of N-cadherin
expressing 3T3 cells in media supplemented with MAG-Fc (25 u,g/ml final concentration)
and BAG polypeptide at the final concentration indicated on the horizontal axis. Each point
indicates data from a single, representative experiment and the .bars on each point indicate the
SEM.
Figure 14 shows a bar graph depicting results from experiments testing the effects of
BAG polypeptide and BDNF on neurite outgrowth in cerebellar neurons that were cultured
over mpnolayers of 3T3 cells that do not express N-cadherin and in media supplemented with
0 or 25 u-g/ml final concentration MAG-Fc (as indicated in the figure). Experiments were
done in control media (i.e., in media supplemented with MAG-Fc only) and in media
t
additionally supplemented with BAG polypeptide (6 uM final concentration) or BDNF
(5 ng/ml final concentration). Cultures were maintained for 22 hours before being fixed and
stained for GAP-43. The mean length of the longest neurite was determined from
measurements of between about 100-120 neurons under each culture condition. Each
column of the graph depicts pooled results from three independent experiments, and the bars
indicate standard error of the mean (SEM).

Figure 15 shows a bar graph depicting results from experiments testing the BAG
polypeptide's effect on neurite outgrowth in cerebellar neurons that were cultured over
monolayers of N-cadherin expressing 3T3 cells in either: (C) control media without
supplements; (1) media supplements with monoclonal antibody for GTlb (20 μg/ml final
concentration); or (2) media supplemented with both the GTlb antibody (20 μg/ml final
concentration) and BAG polypeptide (6 uM final concentration). Cultures were rnaintained
for 22 hours before fixing and staining for GAP-43. The mean length of the longest neurite
was determined from between about 100 and 120 neurons under each culture condition.
Each column in the figure indicates pooled results from between 7 and 10 independent
experiments, and the bars on each column indicate the SEM.
Figure 16 shows a bar graph depicting results from experiments testing the effects of
various agents on neurite outgrowth in cerebellar neurons that were cultured over monolayers
of N-cadherin expressing 3T3 cells either in control media without supplements (column C)
or in media pre-treated with antibody to p75NTR (columns 1-4). These antibody treated
neurons were cultured after treatment in either: (1) control media without supplements; (2)
the Rho kinase inhibitor Y27632 (10 uM final concentration); (3) BAG polypeptide (6 uM
final concentration); or (4) BDNF neurotrophin (5 ng/ml final concentration). Cultures were
maintained for 22 hours before being fixed and stained for GAP-43. The mean length of the
longest neurite was determined from between about 100 and 120 neurons under each culture
condition. Each column in the figure indicates pooled results from the number of
independent experiments indicated above it, and the bars on each column indicate the SEM.
Figure 17 shows a bar graph depicting results from experiments testing the effects of
various kinase inhibitors on neurite outgrowth in cerebellar neurons cultured under different
conditions. In particular, cerebellar neurons were cultured over monolayers of 3T3 cells in
media supplements with either BAG polypeptide at a final concentration of 6 uM (blank bars),
BDNF neurofrophin at a final concentration of 5 ng/ml (striped bars) or FGF2 at a final •
concentration of 5 ng/ml (black bars). To test the effects of various agents, the experiments
were done as indicated in the figure using control media' containing no additional
supplements, or using media additionally supplement with K252a (100 nM final

concentration), a PKA inhibitor (KT5720 at 200 nM final concentration or H-89 at 400 nM
final concentration), or a PI3K inhibitor (Wortmannin or Ly294002, each, at a final
concentration of 10 uM). Cultures were maintained for 18 hours before being fixed and
stained for GAP-43, and the mean length of the longest neurite was determined from between
about 100 and 120 neurons under each culture condition. Data are pooled for results
obtained with each of the PKA inhibitors and each of the PDK inhibitors (which produced
the same results). Each column in the figure indicates pooled results from at least three
independent experiments, and the bars on each column indicate the SEM.
Figures 18A-18B show graphs depicting the results from experiments testing the
effects on neurite growth in cerebellar neurons that were cultured in an "inhibitory
environment" of wells coated with polylysine at 17 jig/ml in distilled water (dH^O); a
mixture of goat anti-human IgG (Fc-specific) and fibronectin (both at 10 μg/ml in DMEM);
and MAG-Fc at 0.25 jag/ml in DMEM/10% FCS. Cultures were maintained for 27 hours
before being fixed and stained for GAP-43. Figure ISA shows a dose-response curve of
mean neurite length of cerebellar neurons grown in the presence of hriBAG2, hBAG2 or riBAG-
Figure 18B shows a bar graph depicting the mean neurite length of cerebellar neurons grown
in the presence of BDNF, BAG, hriBAG2 hBAG2 or riBAG-
Figure 19 shows a bar graph depicting results from neurite outgrowth experiments
testing the effects of various agents in an inhibitory environment. In particular, cerebellar
neurons were cultured over monolayers of N-cadherin expressing 3T3 cells in media
supplemented with a soluble MAG-Fc fusion construct at a final concentration of 25μg/ml.
The culture was further supplemented with BDNF (lng/ml), NGF (10 ng/ml or 100 ng/ml),
BDNF (lng/ml) in combination with NGF (10 ng/ml or 100 ng/ml), a constrained monomer
of the NGF loop 1 binding motif (N-Ac-CTDKGKEC-NH2) (SEQ ID NO:43) at 100μg/ml,
or the NGF loop 1 peptide (at 100μg/ml) in combination with BDNF (at lng/ml). Cultures
were maintained for 23 hours before being fixed and stained for GAP-43.. The mean length
of the longest neurite was determined from measurements of between about 100-120 neurons
under each culture condition. Each column of the graph depicts pooled results from a

number of independent experiments (indicated above the column), and the bars indicate
standard error of the mean (SEM).
6. DETAILED DESCRIPTION
As noted above, the present invention provides compounds, including peptides and
peptidornimetics, that modulate (e.g, increase or decrease) activity mediated by Trk-
receptors such as TrkA, TrkB and TrkC. Such compounds are generally referred to here as
Trk-receptor modulator compounds or "Trk modulators."
Trk modulators of the invention are useful, e.g., for modulating processes such as
neuronal growth and survival, axonal growth, neurite outgrowth, synaptic plasticity and other
processes that are mediated, at least in part, by a Trk-receptor. These uses include
therapeutic methods that may involve modulating the growth and repair of the central
nervous system in vitro (e.g., in a cell culture) or in vivo (such as in a patient or other
individual). Trk modulators of the invention therefore have utility in the treatment of
diseases such as stroke, Alzheimer's disease, Parkinson's disease, head trauma, spinal cord
injury, and epilepsy to name a few.
Applicants have discovered that a key interaction between Trk receptors and their
neurotrophin ligands occurs through a conserved short linear sequence motif of three amino
acid residues - Arg-Gly-Glu (i.e., "RGE" in the single letter amino acid code) found at the
N-terminal of mature neurotrophin amino acid sequences. The RGE motif is present in all
neurotrophins and, when bound to the Trk receptor, exists as half a helix in what is
considered a tight loop.
Applicants have also discovered that properly constrained peptides (for example,
cyclic peptides) of the small linear RGE motif have a high structural overlap with the native
NT structure and are able to function as Trk receptor antagonists. Similarly, peptidomimetic
compounds having high structural overlap with such constrained RGE peptides are also
expected to have high structural overlap with the native NT structure and, as such, can also
function as Trk receptor antagonists.
As noted above, the RGE motif is conserved among all neurotrophins, and
> interactions with this motif are important for the binding of those neurotrophins to their
respective Trk receptor(s). Hence, constrained peptides and peptidornimetics comprising the

RGE motif are useful as antagonists of a wide variety of Trk receptors, including TrkA, TrkB
and TrkC. However, Trk modulators of the present invention can also be targeted to specific
Trk receptors, by selecting flanking amino acid sequences from an NT ligand that preferably
binds to the desired Trk receptor. In preferred Trk antagonist compounds (i.e., Trk
modulator compounds that inhibit Trk receptor mediated activity), such flanking residues
preferably range in length from no more than about 0 to 10 amino acid residues in length,
with sizes between about 2-5 or 2-3 amino acid residues being particularly preferred.
t
Moreover, the size of the cyclic peptide ring (or the corresponding peptidomimetic structure)
preferably ranges from only about 4 to 15 amino acid residues, with sizes from about 5 to 10
amino acid residues being particularly preferred.
Applicants have also determined that, in crystal structures of NT dimers in complex
with their binding domain of a Trk receptor, the RGE motif runs anti-parallel to itself in the
NT dimer. That is to say, the RGE helix in the first NT molecule is aligned with and in an
anti-parallel orientation to the RGE helix in the second NT molecule in that dimer. See, in
particular, Figures 6A-6C. Applicants have moreover discovered that, when a tandem repeat
peptide or peptidomimetic of the RGE motif is properly constrained (as in a cyclic peptide or
peptidomimetic), it adopts the same anti-parallel alignment conformation and has a high
structural overlap with the native NT structure. Such "tandem-repeat" RGE cyclic peptides
and peptide mimetics are, surprisingly, able to function as Trk receptor agonists (i.e., they are
able to increase activity mediated by a Trk receptor). As such, these compounds are also
among the Trk modulator compounds of the invention.
As with the RGE antagonists, described, supra, constrained peptides and
peptidomimetics comprising a tandem repeat of the RGE motif are useful as agonists for a
wide variety of Trk receptors, including TrkA, TrkB and TrkC. However, the compounds
can also be targeted to specific Trk receptors, for example, by selecting flanking amino acid
sequences from a NT ligand that preferably binds to the desired Trk receptor. In preferred
Trk agonists compounds (i.e., the-Trk modulator compounds that increase Trk receptor
mediated activity) such flanking residues preferably range in length from no more than about
0 to 10 amino acid residues in length, with sizes from 2-5 or 2-3 being more preferred.
Tandem repeat cyclic peptides and peptidomimetics of the invention may, optionally,
contain additional amino acid residues situated between the two tandem repeats of the RGE

motif Such additional amino acid residues therefore function as "spacer" moieties to join
the two RGE motifs together in such a way that they adopt the anti-parallel alignment
conformation having a high structural overlap with the RGE motif in the native NT
stnicture(s). The exact identity of the spacer amino acid residue(s) is not important and their
identities may or may not correspond to identities of amino acid residues flanking the RGE
motif in a particular neurofrophin. Preferably, the spacer moiety (if present) in a tandem
repeat cyclic peptide or peptidomimetic is short; e.g., not longer than five amino acid
residues in length, with spacer moieties between about 0-3 amino acid residues in length
being more preferred. Particularly preferred spacer moieties are only about 1 or 2 amino acid
residues in length.
The total size of such "tandem-repeat" cyclic peptides and peptidomimetics is
moreover, typically about twice that of a Trk antagonist cyclic peptide or peptidomimeric of
the invention. Hence, preferred sizes of the peptide ring (or corresponding peptidomimetic
structure) are preferably from about 8 to 30 amino acid residues in length, with sizes from
about 10 to 20 amino acid residues in length being particularly preferred.
Preferred cyclic peptides that comprise the RGE motif and/or tandem repeats thereof
are described in Section 6.1, infra. Section 6.2 then describes routine experimental methods
by which a person skilled in the art can determine, e.g., by X-ray crystallography or NMR
spectroscopy, three-dimensional 'pharmacophore" structures for these and other cyclic
peptides, and methods using such pharmacophore structures to design appropriate
peptidomimetic compounds are provided in Section 6.3, infra, along with exemplary
peptidomimetic modifications. Still further modifications to the Trk modulator compounds
of the invention are described in Section 6.4, including pharmaceutical formulations and
medicinal uses. Section 6.5 describes routine assays by which a person skilled in the art may
verify the Trk modulator activity of a compound, such as a peptidomimetic compound.
Section 6.6 then describes certain preferred exemplary uses of such compounds in methods
for modulating Trk mediated activities such as neuronal survival, axonal growth and synaptic
plasticity. Formulations of Trk modulators (including pharmaceutical formulations) that are
particularly suitable for such uses are provided in Section 6.7. The specification concludes
with a series of examples, at Section 7, demonstrating certain exemplary embodiments of the
invention.

The present invention also provides methods for the administration of a p75 receptor
binding agent to an individual, wherein the p75 receptor binding agent interferes with the
binding of a neurotrophin to a p75 receptor engaged in an inhibitory complex and, thus,
, promotes neurotrophin binding to a Trk receptor. These methods are useful in the treatment
of conditions wherein CNS neurons are damaged or injured, for example, diseases such as
stroke, Alzheimer's disease, Parkinson's disease, traumatic brain injury and spinal cord
injury.
The present invention is based, in part, on the discovery that agents designed to
interfere with neurotrophin binding to a p75 receptor engaged in an inhibitory complex
facilitate neurotrophin-mediated CNS neurite outgrowth. As described in the Examples
section below, neurite outgrowth assays were performed using cultured rat cerebellar neurons
in an inhibitory environment. The inhibitory culture media was supplemented with NGF,
NGF-in combination with BDNF, a constrained monomer of the NGF first binding loop (N-
Ac-CTDIKGKEC-NH^ (SEQ ID NO:43), or the NGF loop 1 peptide in combination with
BDNF. Cultures were maintained for 23 hours, fixed, and stained for GAP-43. The mean
length of the longest neurite was then determined from 100-120 neurons under each culture
condition. It was found that NGF alone and the constrained monomer alone did not promote
neurite outgrowth, but the combinations ofNGF with BDNF in ratios of 10:1 and 100:1 and
the NGF loop 1 peptide at 100 μg/ml with BDNF at 1 ng/ml promoted neurite outgrowth.
These findings show that administration of a p75 receptor binding agent, which is a
neurolxophin that does not bind the expressed Trk receptor in combination with another,
different neurotrophin, which binds a Trk receptor expressed on the injured neurons results in
CNS neuron growth in an inhibitory environment. Here, the neurotrophin, which does not
bind to an expressed Trk receptor (i.e., a p75 receptor binding agent) was administered in an
amount about 10 to about 100 fold greater than the neurotrophin, which binds to an expressed
Trk receptor (i.e., not a p75 receptor binding agent). These results also show that
administration of a constrained monomer of the NGF first binding loop, a non-neurotrophin
p75 receptor binding agent, promotes neurotrophin-mediated CNS neuron growth.
Definitions
The following defined terms are used throughout the present specification, and should
be helpful in understanding the scope and practice of the present invention.

As used herein "an inhibitory environment" means an environment in which growth
of a damaged or injured neuron is inhibited. An inhibitory environment is present in the
milieu surroimdmg damaged or injured neurons. Damaged or injured neurons are present in
conditions, which include, for example, diseases and disorders that are associated with
damage to or impaired function of the CNS. Exemplary conditions include, but are not
limited to, Alzheimer's disease, Parkinson's disease, Huntington's disease, amyotrophia
lateral sclerosis, stroke, traumatic brain injury, and spinal cord injury. Alternatively, an
inhibitory environment is an environment wherein a p75 receptor is engaged in an inhibitory
complex (i.e., a p75 receptor is engaged with a molecule that binds a p75 receptor, e.g., a
myelin-derived molecule such as MAG or Nogo-A), and this binding results in inhibition of
nein-otrophin-mediated neuron growth).
The term "therapeutically effective" means that quantity of a compound or
pharmaceutical composition that is sufficient to result in a desired activity upon
administration to an individual in need thereof. Preferably, a therapeutically effective
amount can ameliorate or prevent a clinically significant deficit in the activity, function and
response of the individual. Alternatively, a therapeutically effective amount is sufficient to
cause an improvement in a clinically significant condition in the individual. For example,
"therapeutically effective" means an amount or dose of a p75 receptor inhibitor sufficient to
promote the growth of a neuron in the CNS in an inhibitory environment.
As used herein, the phrase "pharmaceutically acceptable" refers to molecular entities
and compositions that are "generally regarded as safe", e.g., that .are physiologically tolerable
and do not typically produce an allergic or similar untoward reaction, such as gastric upset,
dizziness and the like, when adniinistered to a human. Preferably, as used herein, the term
"pharmaceutically acceptable" means approved by a regulatory agency of the Federal or a
state government of listed in the U.S. Pharmacopeia or other generally recognized
pharmacopeia for use in animals, and more particularly in humans.
The term "carrier" refers to a diluent, adjuvant, excipient, or vehicle with which the
compound is administered. Such pharmaceutical carriers can be sterile liquids, such as water
and oils, including those of petroleum, animal, vegetable, or synthetic origin, such as peanut
oil, soybean oil, mineral oil, sesame oil and the like. Water, or aqueous saline solutions and
aqueous dextrose and glycerol solutions are preferably'employed as carriers, particularly for

injectable solutions. Suitable pharmaceutical carriers are described in 'Remington's
Pharmaceutical Sciences" by E.W. Martin.
An "individual" or "patient" as used herein is preferably a mammal and more
preferably a human, but can be any animal, including a laboratory animal in the context of a
clinical trial or screening or activity experiment Thus, as can be readily appreciated by one
of ordinary skill in the art, the methods of the present invention are particularly suited to
administration to any animal, particularly a mammal, and including, but not limited to
domestic animals, wDd animals and research animals.
6.1. TRK RECEPTOR MODULATORS: CYCLIC PEPTIDES
The term "cyclic peptide," as used herein, refers to a peptide or salt thereof that
comprises: (1) an intramolecular covalent bond between two non-adjacent residues; and (2)
at least one Trk-receptor recognition sequence RGE (i. e., Arg-Gly-Glu) within a cyclic ring
of the cyclic peptide. It is understood that preferred peptides of the invention which function
as either Trk receptor agonists or antagonists will be constrained and, hence, are preferably
cyclic peptides. However, non-cyclic or "linear" peptides are also useful (e.g., as
intermediate compounds for making cyclic peptides of the invention). Hence, non-cyclic
versions of the cyclic peptides described throughout this application are also considered part
of the present invention.
The intramolecular bond may be a backbone to backbone, side-chain to backbone or
side-chain to side-chain bond (i.e., terminal functional groups of a linear peptide and/or side-
chain functional groups of a terminal or interior residue may be linked to achieve
cyclization). Preferred intramolecular bonds include, but are not limited to; disulfide, amide
and thioether bonds. A variety of means for cyclizing polypeptides are well known in the art,
as are many other modifications that can be made to such peptides. For a general discussion,
see International Patent Publication Nos. WO 01/53331 and WO 98/02452. Such cyclic
bonds and other modifications can also be applied to the cyclic peptides and derivative
compounds of this invention. For convenience, cyclic peptides of the invention are
frequently illustrated in this application showing particular cyclic bonds, which may or may
not be preferred. However, other embodiments of these cyclic peptides comprising

additional and/or alternative cyclic bonds, will be apparent to those persons skilled in the art
and are therefore considered part of this invention.
• Within certain embodiments a cyclic peptide of the invention preferably comprises an
N-acetyl group (i.e., an amino group present on the amino terminal residue of the peptide is
acetylated, preferably prior to cyclization). Alternatively, a cyclic peptide of the invention
may comprise an N-formyl group (i.e., the amino group present on the amino terminal
residue of the peptide is formylated, preferably prior to cyclization). Alternatively, the amino
group present on the amino terminal residue of the peptide may be mesylated; again,
preferably prior to cyclization. The presence of such terminal groups may, for example,
enhance cyclic peptide activity or stability in certain applications. In addition, witihin certain
embodiments a cyclic peptide of the invention may comprise a C-amide group.
In certain embodiments, preferred cyclic peptides of the present invention satisfy the
general formula:
(Formula I)
where Yi and Y2 are amino acid residues whose identities are independently selected and
having a covalent bond between the residues Yi and Y2. The elements Xi and X2 are
optional and, if present, they are independently selected from the amino acid residues and
combinations thereof that are linked by peptide bonds. Hence, either Xi or X2 or both Xi and
X2, if present, maybe a single amino acid residue or, alternatively, may each be a sequence
comprising a plurality of amino acid residues linked by peptide bonds.
In preferred embodiments, a cyclic peptide satisfying Formula I, above, will modulate
one or more Trk-receptor mediated activities. For example, in certain preferred embodiments
a peptide satisfying Formula I will inhibit one or more Trk-receptor mediated activities and,
as such, will be a Trk antagonist In other embodiments, a peptide satisfying Formula I will
increase one or more Trk receptor mediated activities and, as such, will be a Trk agonist.
In addition to the RGE consensus sequence(s), cyclic peptides of the invention
generally comprise at least one additional residue within the cyclic ring so that preferably at
least one of either Xi or X2 in Formula I is present. Generally, the size of Xi and/or X2 will
depend upon the desired activity of the cyclic peptide. For example, where a cyclic peptide

that is a Trk antagonist is desired, shorter peptide sequences are preferred. Accordingly, in
such embodiments, X1 and/or X2 are each preferably between 0 and about 10 amino acids in
length, with sizes of about 1,2,3,4 or 5 amino acid residues being particularly preferred.
Moreover, in such embodiments the lengths of X1 and/or X2 are also preferably selected so
that the size of the cyclic peptide ring ranges from about 5 to about 15 amino acid residues,
and is more preferably between about 5-10 amino acid residues in length. Peptide ring sizes
of about 5-7 amino acid residues in length are particularly preferred. Such additional
residues {i.e., X1 and/or X2 in Formula I, supra) may be present on either the N-terminal or
C-terminal side of the RGE sequence, or they may be present on both sides of the RGE
sequence.
In preferred cyclic peptides of the invention, the additional residues are derived from
sequences that flank the RGE sequence within one or more naturally occurring neurotrophins
(e.g., NGF, BDNF, NT-3, NT-4, NT-5 and NT-4/5) with or without amino acid substitutions
and/or other modifications. In particular, the presence of flanking sequences from a
neurotrophin may help target a cyclic peptide for a particular Trk receptor of interest. Hence,
in embodiments where an antagonist for a particular Trk receptor is desired, a cyclic peptide
of the invention may comprise amino acid residues flanking either the N-terminal, C-terminal
or both sides of the RGE sequence that are derived from flanking sequences in a
neurotrophin that preferably binds to the targeted Trk receptor.
As an example, and not by way of limitation, Table I, infra, list certain preferred
cyclic peptides that comprise additional amino acid residues derived from particular
neurotrophins whose identities are also indicated in the table. The right-hand column in
Table I also indicates a Trk receptor to which the neurotrophin preferably binds (or, rather,
binds with the highest binding affinity). Hence, each cyclic peptide listed in Table I may, in
one embodiment, be used to inhibit the particular Trk receptor indicated along side it, in the
right hand column of Table L Those skilled in the art will appreciate, however, that there is
some overlap in binding specificity of the different neurotrophin ligands for various Trk
receptors. Hence, the cyclic peptides listed in Table I can also be used as antagonists of other
Trk receptors. As a particular example, and not by way of limitation, it is demonstrated in
the Examples, infra, that the cyclic peptide N-Ac-CSRRGEC-NH?, which contains additional
residues from the neurotrophin NT-4, is a more potent TrkB antagonist than are the peptides

The underlined portion of each foregoing amino acid sequence indicates that portion
of the peptide that is cyclized. "N-Ac" denotes an acetylated N4erminal amino group and
"NH2" denotes a C-erminal amide group.
Within certain embodiments, relatively small cyclic peptides of the invention that do
not contain significant sequences flanking the RGE consensus sequence are particularly
preferred. Such peptides may or may not contain an N-acetyle group and they may or may
not contain a C-amide group. Examples of preferred, small cyclic peptides of the invention
include:

In other embodiments of the invention, where a Trk agonist is desired, longer peptide
sequences are generally preferred. In particular, preferred cyclic peptides of the invention
that are Trk agonists comprise at least one "tandem repeat" of the RGE motif. Accordingly,
where such cyclic peptides satisfy Formula I, supra, at least one of X1 and X2 will be present
and comprises a second RGE sequence. More specifically, such cyclic peptides of the
invention preferably satisfy the following general formula:
(Formula IT)
As in Formula I, Y1 and Y2 are amino acid residues whose identities are
independently selected and having a covalent bond between the residues Y1 and Y2. The
elements Z1 and Z2 are optional and, if present, they are independently selected from the
amino acid residues and combinations thereof that are linked by peptide bonds. The element
Zo is also optional, and if present is an amino acid residue or some combination thereof,
linked by peptide bonds. Hence, either Z1 Z2, Zo or any combination thereof, if present, may

each be a single amino acid residue or, alternatively, they may each be a sequence
comprising a plurality of amino acid residues linked by peptide bonds.
In addition to a tandem repeat of the RGE consensus sequence, cyclic peptides of the
invention generally comprise one additional residues within the cyclic ring so that,
preferably, at least one of either Z1, Z2 and/or Zo is present In embodiments where a cyclic
peptide that is a Trk agonist is desired, Z1, Z2 and/or Zo are each preferably no more than
about ten amino acid residues in length, and more preferably are each only 1,2,3,4 or 5
amino acid residues in length. Moreover, the lengths of Z1, Z2 and/or Zo are preferably
selected so that the size of the cyclic peptide ring ranges from about 8-50 amino acid
residues, and more preferably from about 8-25 or from about 15-20 amino acid residues.
As with the cyclic peptides of Formula I, in preferred cyclic peptides of Formula II
the additional residues (i.e.} Z1, Z2 and/or Zo) can be derived from sequences that flank the
RGE sequence within one or more naturally occurring neurotrophin (e.g., NGF, BDNF, NT-
3, NT-4, NT-5 or NT-4/5), with or without amino acid substitutions and/or other
modifications. In particular, the presence of flanking amino acid residues from a particular
neurofrophin may help target a cyclic peptide for a particular Trk receptor of interest. Hence,
in embodiments where an antagonist for a particular Trk receptor is desired, a cyclic peptide
of the invention may comprise amino acid residues flanking the N-terminal and/or C-terminal
of one or both tandem repeat RGE sequences, and these flanking sequences may be derived
from a neurotrophin that preferably binds to the targeted Trk receptor of interest.
As noted above, preferred tandem repeat cyclic peptides of the invention (including
cyclic peptides according to Formula E) have the two RGE sequences aligned anti-parallel to
each other. Accordingly, in preferred cyclic peptides according to Formula n the element Zo
is present and can function as an effective "spacer moiety" to align the two RGE sequences
together in an anti-parallel alignment conformation, hi preferred embodiments, Zo is not
more than 10 amino acid residues in length, and is preferably five or fewer amino acid
residues in length. Preferred sizes for Zo are about 1,2,3,4 or 5 amino acid residues in
length. The exact sequence of amino acid residues in Zo is not critical. As such, the element
Zo may or may not comprise a sequence of amino acid residues corresponding to a sequence
from either the N-terminal or C-terminal of the RGE motif in a natural neurotrophin (e.g.,
NGF, BDNF, NT-3, NT-4, NT-5 and NT-4/5). Where Z0 does comprise sequences from a

neurotrophin, those sequences may or.may not comprise amino acid substitutions and/or
modifications.
Examples of particularly preferred cyclic peptide sequences of the invention, which
are preferably Trk agonists, include:

The cyclic peptide sequences identified by SEQ ID NOs: 30,31, and 32 are TrkA,
TrkB, and TrkC agonists, respectively. SEQ ID NO:30 is derived from NGF. SEQ ID NO:
31 is derived from BDNF. SEQ ED NO: 32 is derived from NT-3.
Cyclic peptides as described herein may comprise residues of L-amino acids, D-
amino acids, or any combination thereof. The amino acids may from natural or non-natural
sources provided that at least one amino group and at least one carboxyl group are present in
the molecule. A- and β-aramo acids are generally preferred. The 20 L-amino acids
commonly found in proteins are particularly preferred in the present invention. These amino
acids are identified herein by their conventional three-letter and one-letter abbreviations,
whereas the corresponding D- amino acids are designated by the prefix "d".
Jn certain embodiments, cyclic peptides of the invention may comprise a sequence of
D-amino acid residues that is the opposite of a sequence of L-amino acid residues provided
herein. For example, the invention provides certain Trk receptor agonist polypeptides,
referred to herein as riBAGi and briBAG2 (SEQ 3D NOS:40-41) that comprise sequences of D-
amino acid sequences which are the reverse sequence of another Trk receptor agonist
polypeptide refereed to as the BAG polypeptide (SEQ ID NO: 17). Hence, in addition to the
polypeptides of L-amino acid residues described supra, the present invention also
contemplates polypeptides having the reverse sequence of L-amino acid residues. Hence, in
one preferred embodiment peptides and peptidomimetics of the present invention comprise
sequences of L-amino acid residues including the Arg-Gly-Glu (z'.e., "RGE") motif
described, supra. Accordingly, the invention also provides, in an alternative embodiment)

peptides and peptidomimetics comprising sequences of D-amino acid residues including the
short linear sequence motif dGlu-Gly-dArg (i.e., "dEGdR").
Cyclic peptides may also contain one or more rare amino acids (such as 4-
hydroxyproline or hydroxylysine), organic acids or amides and/or derivatives of common
amino acids, such as amino acids having the C-tenninal carboxylate esterified (e.g., benzyl,
methyl or ethyl ester) or amidated and/or having modifications of the N -erminal amino
group (e.g., acetylation or alkoxycarbonylation), with or without any of a wide variety of
side-chain modifications and/or substitutions (e.g., methylation, benzylation, t-butylation,
tosylation, alkoxycarbonylation, and the like). Preferred derivatives include amino acids
having an N-acetyl group (such that the amino group that represents the N-terminus of the
linear peptide prior to cyclization is acetylated) and/or a C-terminal amide group (z'.e., the
carboxy terminus of the linear peptide prior to cyclization is amidated). Residues other than
common amino acids that may be present with a cyclic peptide include, but are not limited to,
penicillamine, β,β-tetramethylene cysteine, β,β-pentamethylene cysteine,β-
mercaptopropiomc acid,β,β-pentamemylene-β-mercaptopropionic acid, 2-mercaptobenzene,
2-mercapioamline, 2-mercaptoproline, ornithine, diaminobutyric acid, a-aminoadipic acid,
m-aminomemylbenzoic acid and a,β-diaminopropionic acid.
CycHc peptides as described herein may be synthesized by methods well known in
the art, including recombinant DNA methods and chemical synthesis. Chemical synthesis
may generally be performed using standard solution phase or solid phase peptide synthesis
techniques, in which apeptide linkage occurs through the direct condensation of the a-amino
group of one amino acid with the a-carboxy group of the other amino acid with the
ehtnmation of a water molecule. Peptide bond synthesis by direct condensation, as
formulated above, requires suppression of the reactive character of the amino group of the
first and of the carboxyl group of the second amino acid. The masking substituents must
permit their ready removal, without inducing breakdown of the labile peptide molecule.
In solution phase synthesis, a wide variety of coupling methods and protecting groups
may be used (see Gross & Meienhofer, eds., "The Peptides: Analysis, Synthesis, Biology,"
Vol. 1-4 (Academic Press, 1979);- Bodansky & Bodansky, "The Practice of Peptide
Synthesis," 2d ed. (Springer Verlag, 1994)). In addition, intermediate purification and linear
scale up are possible. Those of ordinary skill in the art will appreciate that solution synthesis

requires consideration of main chain and side chain protecting groups and activation method.
In addition, careful segment selection is necessary to minimize racemization during segment
condensation. In particular, a high percentage of racemization may be observed when
residues such as Phe-Gly are coupled. Such situations are, however, uncommon. Solubility
considerations are also a factor.
Solid phase peptide synthesis uses an insoluble polymer for support during organic
synthesis. The polymer-supported peptide chain permits the use of simple washing and
filtration steps instead of laborious purifications at intermediate steps. Solid-phase peptide
synthesis may generally be performed according to the method of Merrifield et al, J. Am.
Chem. Soc. 1963,85:2149. These methods involve assembling a linear peptide chain on a
resin support using protected amino acids. Solid phase peptide synthesis typically utilizes
either the Boc or Fmoc strategy. The Boc strategy uses a 1% cross-linked polystyrene resin.
The standard protecting group for a-amino functions is the tert-butyloxycarbonyl (Boc)
group. This group can be removed with dilute solutions of strong acids such as 25%
trifiuoroacetic acid (TFA). The next Boc-amino acid is typically coupled to the amino acyl
resin using dicyclohexylcarbodiimide (DCC). Following completion of the assembly, the
peptide-resin is treated with anhydrous HF to cleave the benzyl ester link and liberate the free
peptide. Side-chain functional groups are usually blocked during synthesis by benzyl-
derived blocking groups, which are also cleaved by HF. The free peptide is then extracted
from the resin with a suitable solvent, purified and characterized. Newly synthesized
peptides can be purified, for example, by gel filtration, HPLC, partition chromatography
and/or ion-exchange chromatography, and may be characterized by, for example, mass
spectrometry or amino acid sequence analysis. In the Boc strategy, C-terminal amidated
peptides can be obtained using benzhydrylamine or memylbenzhydrylamine resins, which
yield peptide amides directly upon cleavage with HF.
In the procedures discussed above, the selectivity of the side-chain blocking groups
and of the peptide-resin link depends upon the differences in the rate of acidolytic cleavage.
Orthoganol systems have been introduced in which the side-chain blocking groups and the
peptide-resin link are completely stable to the reagent used to remove the a-protecting group
at each step of the synthesis. The most common of these methods involves the 9-
fluorenylmetbyloxycarbonyl (Fmoc) approach. Within this method, the side-chain protecting

groups and the peptide-resin link are completely stable to the secondary amines used for
cleaving the N-a-Fmoc group. The side-chain protection and the peptide-resin link are
cleaved by mild acidolysis. The repeated contact with base makes the Merrifield resin
unsuitable for Fmoc chemistry, and p-alkoxybenzyl esters linked to the resin are generally
used. Deprotection and cleavage are generally accomplished using TFA.
Those of ordinary skill in the art will recognize that, in solid phase synthesis,
deprotection and coupling reactions must go to completion and the side-chain blocking
groups must be stable throughout the entire synthesis. In addition, solid phase synthesis is
generally most suitable when peptides are to be made on a small scale.
Acetylation of the N-terminal can be accomplished by reacting the final peptide with
acetic anhydride before cleavage from the resin. C-amidation is accomplished using an
appropriate resin such as methylben2hydrylamine resin using the Boc technology.
Following synthesis of a linear peptide, with or without N-acetylation and/or C-
amidation, cyclization may be achieved by any of a variety of techniques well known in the
art. Within one embodiment, a bond may be generated between reactive amino acid side
chains. For example, a disulfide bridge maybe formed from a linear peptide comprising two
thiol-containing residues by oxidizing the peptide using any of a variety of methods. Within
one such method, air oxidation of thiols can generate disulfide linkages over a period of
several days using either basic or neutral aqueous media. The peptide is used in high dilution
to minimize aggregation and intermolecular side reactions. This method suffers from the
disadvantage of being slow but has the advantage of only producing H2O as a side product.
Alternatively, strong oxidizing agents such as I2 and K3Fe(CN)6 can be used to form disulfide
linkages. Those of ordinary skill in the art will recognize that care must be taken not to
oxidize the sensitive side chains of Met, Tyr, Trp or His. Cyclic peptides produced by this
method require purification using standard techniques, but this oxidation is applicable at acid
pHs.
Oxidizing agents also allow concurrent deprotection/oxidation of suitable S-protected
linear precursors to avoid premature, nonspecific oxidation of free cysteine.
DMSO, unlike I2 and K3Fe(CN)6, is a mild oxidizing agent which does not cause
oxidative side reactions of the nucleophilic amino acids mentioned above. DMSO is
miscible with H20 at all concentrations, and oxidations can be performed at acidic to neutral

pHs with harmless byproducts. MethyltricMorosilane-dipiieriylsiilfoxide may alternatively
be used as an oxidizing agent, for concurrent deprotection/oxidation of S-Acm, S-Tacm or S-
t-Bu of cysteine without affecting other nucleophilic amino acids. There are no polymeric
products resulting from intermolecular disulfide bond formation.
Suitable tori-containing residues for use in such oxidation methods include, but are
not limited to, cysteine, p,p-dimethyl cysteine (peiricinamine or Pen), P,p-tetramethylene
cysteine (Tmc), P,P-pentamethylene cysteine (Pmc), p-mercaptopropionic acid (Mpr), P,P-
pentamethylene-p-mercaptopropionic acid (Pmp), 2-mercaptobenzene, 2-mercapto aniline
and 2-mercaptoproline.
It will be readily apparent to those of ordinary skill in the art that, within each of these
t
representative formulas set forth supra, any of the above tHol-contaming residues may be
employed in place of one or both of the Mol-containing residues recited.
Within further embodiments, cyclization may be achieved by amide bond formation.
For example, a peptide bond may be formed between terminal functional groups (i.e., the
amino and carboxy termini of a linear peptide prior to cyclization). Examples of such
peptides include c(SRRGE) (SEQ ID NO:33), c(ARRGE) (SEQ ID NO:34), cflFHRGE)
(SEQ ID NO:35) and c(SHRGE) (SEQ ID NO:36). An example of one particularly preferred
peptide having such a cyclic amide bond is the peptide c(SRRGELSRRGEL) (SEQ ID
NO:39). This peptide, which is described in the Examples, infra, is referred to here as the
1LBAG2 peptide. Within another such embodiment, the linear peptide comprises a D-arrdno
acid. For example, the Examples, infra, describe another preferred peptide that is referred to
as the hriBAG2 peptide. This peptide, which contains a cyclic amide bond as described,
supra, comprises the following sequence of D amino acid residues:
c[dLdEdGdRdRdSdLdEdGdRdRdS] (SEQ ID NO:40). Alternatively, cyclization may be
accomplished by linking one terminus and a residue side chain or using two side chains, as in
KRGED (SEQ K> NO:37) or KSRRGED (SEQ ID NO:38), with or without an N-tenninal
acetyl group and/or a C4erminal amide. Residues capable of forming a lactam bond include
lysine, orm'thine (Orn), a-amino adipic acid, m-aminomethylbenzoic acid, a,P-
diaminopropionic acid, glutamate or aspartate.
Methods for forming amide bonds are well known in the art and are based on well
established principles of chemical reactivity. Within one such method, carbodiimide-

mediated lactam formation can be accomplished by reaction of the carboxylic acid with
DCC, DIC, EDAC or DCCI, resulting in the formation of an O-acylurea that can be reacted
immediately with the free amino group to complete the cyclization. The formation of the
inactive N-acylurea, resulting from an O—»N migration, can be circumvented by converting
the O-acylurea to an active ester by reaction with an N-hydroxy compound such as 1-
hydroxybenzotriazole, 1-hydroxysuccinimide, 1-hydroxynorbomene carboxamide or ethyl 2-
hydroximino-2-cyanoacetate. In addition to minimizing 0—>N migration, these additives
also serve as catalysts during cyclization and assist in lowering racemization. Alternatively,
cyclization can be performed using the azide method, in which a reactive azide intermediate
is generated from an alkyl ester via a hydrazide. Hydrazinolysis of the terminal ester
necessitates the use of a t-butyl group for the protection of side chain carboxyl functions in
the acylating component. This limitation can be overcome by using diphenylphosphoryl acid
(DPPA), which furnishes an azide directly upon reaction with a carboxyl group. The slow
reactivity of azides and the formation of isocyanates by their disproportionation restrict the
usefulness of this method. The mixed anhydride method of lactam formation is widely used
because of the facile removal of reaction by-products. The anhydride is formed upon
reaction of the carboxylate anion with an alkyl chloroformate or pivaloyl chloride. The
attack of the amino component is then guided to the carbonyl carbon of the acylating
component by the electron donating effect of the alkoxy group or by the steric bulk of the
pivaloyl chloride t-butyl group, which obstructs attack on the wrong carbonyl group. Mixed
anhydrides with phosphoric acid derivatives have.also been successfully used. Alternatively,
cyclization can be accomplished using activated esters. The presence of electron
withdrawing substituents on the alkoxy carbon of esters increases their susceptibility to
aminolysis. The high reactivity of esters of p-nitrophenol, N-hydroxy compounds and
polyhalogenated phenols has made these "active esters" useful in the synthesis of amide
bonds. The last few years have witnessed the development, of benzotriazolyloxytris-
(dimethylamino)phosphonium hexafluorophosphonate (BOP) .and its congeners as
advantageous coupling reagents. Their performance is generally superior to that of the well
established carbodiimide amide bond formation reactions.
Within a further embodiment, a thioether linkage may,be formed between the side
chain of a thiol-containing residue and an appropriately derivatized a-amino acid. By way of

example, a lysine side chain can be coupled to bromoacetic acid through the carbodiirnide
coupling method (DCC, ED AC) and then reacted with the side chain of any of the thiol
containing residues mentioned above to form a thioether linkage. .la order to form
dithioethers, any two thiol containing side-chains can be reacted with dibromoethane and
diisopropylamine in DMF. Examples of thiol-containing linkages include:
and
Cyclization may also be achieved using 5i§i-Ditryptophan (i.e., Ac-Trp-Gly-Gly-Trp-
OMe), as shown below:


The structures and formulas recited herein are provided solely for the purpose of
illustration, and are not intended to limit the scope of the cyclic peptides described herein.
6.2. TRK-RECEPTOR PHARMACOPHORES
For designing peptidomimetics, it is beneficial to obtain a three dimensional structure
for the pharmacophore of one or more cyclic peptides described above. The term
"pharmacophore" refers to the collection of functional groups on a compound that are
arranged in three-dimensional space in a manner complementary to the target protein, and
that are responsible for biological activity as a result of compound binding to the target
protein. Useful three-dimensional pharmacophore models are best derived from either
crystallographic or nuclear magnetic resonance structures of the target, but can also be
derived from homology models based on the structures of related targets or three-
dimensional quantitative stracture-activity relationships derived from a previously discovered
series of active compounds.
The present invention provides pharmacophores of certain representative cyclic
peptides (i.e., three-dimensional conformations of the neurotrophin consensus sequence RGE
within such peptides). Such three-dimensional structures provide the information required to
most efficiently direct the design and optimization of peptidomimetics.
In one embodiment, the three-dimensional structures of cyclic peptides are generally
determined using X-ray crystallography. These techniques are well known and are within the
routine skill in the art. For example, see Cantor&Schimmel, Biophysical Chemistry 1980
(Vols. I-III) W. H. Freeman and Company (particularly Chapters 1-13 in Vol. I, and Chapter
13 in Vol. IT). See, also, Macromolecular Crystallography, Parts A-B (Carter&Sweet, Eds.)

In: Methods Enzymol 1997, Vols. 276 -211; JanDrenth, Principles of Protein X-Ray
Crystallography (NewYork: Springer-Verlag, 1994).
The term "crystal" refers, generally, to any ordered (or at least partially ordered)
three-dimensional array of molecules. Preferably, the ordering of molecules within a crystal
is at least sufficient to produce a sharp X-ray diffraction pattern so that the molecules' three-
dimensional structure may be determined.
The molecules in a crystal may be of any type, and it will be understood that a crystal
may contain molecules of only one type or may comprise a plurality of different types of
molecules. In preferred embodiments, crystals of the present invention comprise at least one
biomolecule, such as a cyclic peptide described, supra, in Section 6.1. Crystals of the
invention may even comprise a complex or assembly of two or more proteins or other
biomolecules. For example, a crystal may comprise molecules of a ligand, such as a
neurotrophin, bound to molecules of a receptor, such as a Trk receptor. Typically, crystals
that contain biological molecules such as proteins will contain other molecules as well, such
molecules of solvent (e.g., water molecules) and/or salt. Other molecules such as drugs, drug
candidates or compounds that bind to the protein may also be present in a crystal.
Indeed, crystal structures for the binding domain of Trk receptors complexed with a
neurotrophin are already available in the art. See, for example, Wiesmann et al, Nature
1999,401:184; and Banfield et al, Sturcutre (Comb.) 2001, 9:1191. The coordinates of
these X-ray structures can be readily obtained, for example, from the Protein Data Bank at
(Accession Nos. lwww and lhcf, respectively). Hence, in particularly
preferred embodiments, which are demonstrated in the Examples, infra, pharmacophore
structures of the invention are determined using the X-ray crystal structure(s) of a
neurotrophin bound to an appropriate Trk receptor (or fragment thereof). These three-
dimensional structures can then be used to design peptidomimetics of the invention or,
alternatively, to design additional cyclic peptides that are likely to be Trk modulators.
Alternatively, the three-dimensional structures of cyclic peptides may generally be
determined using nuclear magnetic resonance (NMR) techniques that are well known in the
art. NMR data acquisition is preferably carried out in aqueous systems that closely mimic
physiological conditions to ensure that a relevant structure is obtained. Briefly, NMR
techniques use the magnetic properties of certain atomic nuclei (such as 1H, 13C, 15N and 31P),

which have a magnetic moment or spin, to probe the chemical environment of such nuclei.
The NMR data can be used to determine distances between atoms in the molecule, which can
be used to derive a three-dimensional model or the molecule.
For determining three-dimensional structures of cyclic peptides (and candidate
peptidomimetics, as discussed below) proton NMR is preferably used. More specifically,
when a molecule is placed in a strong magnetic field, the two spin states of the hydrogen
atoms are no longer degenerate. The spin aligned parallel to the field will have a lower
energy and the spin aligned antiparallel to the field will have a higher energy. At
equilibrium, the spin of the hydrogen atoms will be populated according to the Boltzmann
distribution equation. This equilibrium of spin populations can be perturbed to an excited
state by applying radio frequency (RF) pulses. When the nuclei revert to the equilibrium
state, they emit RF radiation that can be measured. The exact frequency of the emitted
radiation from each nucleus depends on the molecular environment of the nucleus and is
different for each atom (except for those atoms that have the same molecular environment).
These different frequencies are obtained relative to a reference signal and are called chemical
shifts. The nature, duration and combination of applied RF pulses can be varied greatly and
different molecular properties can be probed by those of ordinary skill in the art, by selecting
an appropriate combination of pulses.
For three-dimensional structure determinations, one-dimensional NMR spectra are
generally insufficient, as limited information pertaining to conformation may be obtained.
One-dimensional NMR is generally used to verify connectivity within a molecule and yields
incomplete data concerning the orientation of side chains within a peptide. Two-dimensional
NMR spectra are much more useful in this respect and allow for unambiguous determination
of side-chain-to-side-chain interactions and the conformation of the peptide backbone.
Two-dimensional NMR spectra are generally presented as a contour plot in which the
diagonal corresponds to a one-dimensional NMR spectrum and the cross peaks off the
diagonal result from interactions between hydrogen atoms that are directly scalar coupled.
Two-dimensional experiments generally contain a preparation period, an evolution period
where spins are "labeled" as they process in the XY plane according to their chemical shift, a
mixing period, during which correlations are made with other spins and a detection period in
which a free induction decay is recorded.

Two-dimensional NMR methods are distinguished by the nature of the correlation
that is probed during the mixing period. A DQF-COSY (double quantum filtered correlation
spectroscopy) analysis gives peaks between hydrogen atoms that are covalently connected
through one or two other atoms. Nuclear Overhauser effect spectroscopy (NOESY) gives
peaks between pairs of hydrogen atoms that are close together in space, even if connected by
way of a large number of intervening atoms. In total correlation spectroscopy (TOCSY),
correlations are observed between all protons that share coupling partners, whether or not
they are directly coupled to each other. Rotating-frame Overhauser Spectroscopy (ROESY)
experiments may be thought of as the rotating frame analogue of NOESY, and yields peaks
between pairs of hydrogen atoms that are close together in space. One or more such methods
may be used, in conjunction with the necessary water-suppression techniques such as
WATERGATE and water flip-back, to determine the three-dimensional structure of a cyclic
peptide or candidate peptidomimetic under aqueous conditions. Such techniques are well
known and are necessary to suppress the resonance of the solvent (HDO) during acquisition
of NMR data.
By way of example, both TOCSY and NOESY may be applied to representative
cyclic peptides for the purpose of determining the conformation and the assignment. The
water solvent resonance may be suppressed by application of the WATERGATE procedure.
A water flipback pulse may also be applied at the end of the mixing period for both TOCSY
and NOESY experiments to maintain the water signal at equilibrium and to minimize the loss
of amide proton resonances due to their rapid exchange at the near neutral Ph conditions (i.e.,
Ph 6.8) used in the experiment. NMR data may be processed using spectrometer software
using a squared cosine window function along both directions. Baseline corrections may be
applied to the NOESY, ROESY and TOCSY spectra using the standard Bruker polynomial
method.
NOESY data maybe acquired at several mixing times ranging from 80 ms to 250 ms.
The shorter mixing time NOESY may be acquired to ensure that no diffusion effects were
present in the NOESY spectrum acquired at the longer mixing times. The interproton
distances may generally be determined from the 250 ms NOESY. The sequence-specific
assignment of the proton resonances may be determined by standard methods (see Wuthrich,

NMR of Proteins and Nucleic Acids, "Wiley & Sons, New York, 1986), making use of both
the results of the TOCSY and NOESY data.
For conformational calculations, the NOE cross peaks may be initially converted to a
uniform distance upper and lower bounds of 1.8-5.0 angstroms regardless of the NOE
intensities. The NOE distances may be refined iteratively through a comparison of computed
and experimental NOEs at the various mixing times. This refinement may be much in the
spirit of the PEPFLEX-II procedure (Wang et al, Techniques in Protein Chemistry IV, 1993,
Evaluation of NMR Based Structure Determination for Flexible Peptides: Application to
Desmopressin p. 569), although preferably initial NOE-based distances with very loose upper
bounds (e.g., 5 angstroms) are used to permit the generation of a more complete set of
conformations in agreement with experimental data. Dihedral-angle constraints may be
derived from the values of the 3JCH coupling constants. A tolerance value of 40 degrees
may be added to each of the dihedral angle constraints to account for the conformational
flexibility of the peptide. Distance geometry calculations may be carried out utilizing fixed
bond lengths and bond angles provided in the ECEPP/2 database (Ni et al, Biochemistry
1992,31:11551-11557). The co-angles are generally fixed at 180 degrees, but all other
dihedral angles may be varied during structure optimization. .
Structures with the lowest constraint violations may be subjected to energy
minimization using a distance-restrained Monte Carlo method (Ripoll & Ni, Biopolymers
1992,32:359-365; Ni, J. Magn. Reson. B 1995,106:147-155), and modified to include the
ECEPP/3 force field (Ni et al, J. Mol Biol. 1995,252:656-671). All ionizable groups may
be treated as charged during constrained Monte Carlo minimization of the ECEPP/3 energy.
Electrostatic interactions among all charges may be screened by use of a distance-dependent
dielectric to account for the absence of solvent effects in conformational energy calculations.
In addition, hydrogen-bonding interactions can be reduced to 25% of the full scale, while van
der Waals and electrostatic terms are kept to full strengths. These special treatments help to
ensure that the conformational search is guided primarily by the experimental NMR
constraints and that the computed conformations are less biased by the empirical
conformational energy parameters (Warder et al, FEBSLett. 1997,411:19-26).
Low-energy conformations of the peptide from Monte Carlo calculations may be used
in NOE simulations to identify proximate protons with no observable NOEs and sets of

distance upper bounds that warrant recalibration The refined set of NOE distances including
distance lower bounds derived from absent NOEs are used in the next cycles of Monte Carlo
calculations, until the resulting conformations produced simulate NOE spectra close to those
observed experimentally (Ning et al, Biopolymers 1994,34:1125-1137; Ni et ah, J. Mol.
Biol. 1995,252:656-671). Theoretical NOE spectra may be calculated using a tumbling
correlation time of 1.5 ns based on the molecular weight of the peptide and the experimental
temperature (Cantor& Schimmel (1980) Biophysical Chemistry, W. H. Freeman & Co., San
Francisco). All candidate peptide conformations are included with equal weights in an
ensemble-averaged relaxation matrix analysis of interconverting conformations (Ni & Zhu, J.
Magn, Reson. B 1994,102:180-184). NOE simulations may also incorporate parameters to
account for the local motions of the methyl groups and the effects of incomplete relaxation
decay of the proton demagnitizations (Ning et al, Biopolymers 1994,34:1125-1137). The
computed NOE intensities are converted to the two-dimensional FID's (Ni, Magn. Reson. B
1995,106:147-155) using the chemical shift of assignments, estimated Hnewidths and
coupling constants for all resolved proton resonances. Calculated FIDs may be converted to
simulated NOESY spectra using identical processing procedures as used for the experimental
NOE data sets.
. 6.3. TRK RECEPTOR MODULATORS: PEPTIDOMIMETICS
As noted above, peptidomimetics are compounds in which at least a portion of the
RGE sequence within a cyclic peptide is modified, such that the three dimensional structure
of the peptidomimetic remains substantially the same as that of the RGE sequence.
Peptidomimetics may be peptide analogues that are, themselves, cyclic peptides containing
one or more substitutions or other modifications within the RGE sequence. Alternatively, at
least a portion of the RGE sequence may be replaced with a nonpeptide structure, such that
the three-dimensional structure of the cyclic peptide is substantially retained. In other words,
one, two or three amino acid residues within the RGE sequence may be replaced by a non-
peptide structure. In addition, other peptide portions of the cyclic peptide may, but need not,
be replaced with a non-peptide structure. Peptidomimetics (both peptide and non-peptidyl
analogues) may have improved properties {e.g., decreased proteolysis, increased retention or
increased bioavailability). Peptidomimetics generally have improved oral availability, which

makes them especially suited to treatment of conditions such as cancer. It should be noted
that peptidomimetics may or may not have similar two-dimensional chemical structures, but
share common three-dimensional structural features and geometry. Each peptidomimetic
may further have one or more unique additional binding elements. The present invention
. provides methods for identifying peptidomimetics. A variety of modifications of peptide
modifications (including modifications to cyclic peptides as described supra) are known in
the art and can be used to generate peptidomimetic compounds. See, for instance,
International Patent Publication No. WO 01/53331. Such modifications can also be used in
the present invention to generate peptidomimetic compounds, as well as the specific
modifications described below.
All peptidomimetics provided herein have a three-dimensional structure that is
substantially similar to a three-dimensional structure of a cyclic peptide as described above.
In general, two three-dimensional structures are said to be substantially structurally similar to
each other if their pharmacophore atomic coordinates have a root-mean square deviation
(RMSD) less than or equal to 1 angstrom, as calculated using the Molecular Similarity
module within the QUANTA program (QUANTA, available from Molecular Simulations
Inc., San Diego, Calif.). All peptidomimetics provided herein have at least one low-energy
three-dimensional structure that is substantially similar to at least one low-energy three-
dimensional structure of a cyclic peptide as described above.
Low energy conformations may be identified by conformational energy calculations
using, for example, the CHARMM program (Brooks et al, J. Comput. Ckem. 1983,4:187:
217). The energy terms include bonded and non-bonded terms, including bond length
energy, angle energy, dihedral angle energy, Van der Waals energy and electrostatic energy.
It will be apparent that the conformational energy can be also calculated using any of a
variety of other commercially available quantum mechanic or molecular mechanic programs.
A low energy structure has a conformational energy that is within 50 kcal/mol of the global
minimum.
The low energy conformations) of candidate peptidomimetics are compared to the
low energy conformations of the cyclic peptide (as determined, for example, by NMR or X-
ray crystallography) to determine how closely the conformation of the candidate mimics that
of the cyclic peptide. In such comparisons, particular attention should be given to the

locations and orientations of the elements corresponding to the crucial side chains. If at least
one of the candidate low energy conformations is substantially similar to a solution
conformation of a cyclic peptide (i.e., differs, with a root-mean square deviation (RMSD) of 1
angstrom or less), the candidate compound is considered a peptidomimetic. Within such
analyses, low energy conformations of candidate peptidornimetics in solution maybe studied
using, for example, the CHARMM molecular mechanics and molecular dynamics program
(Brooks et al, J. Comput. Chan. 1983,4:187-217), with the TIP3P water model (Jorgensen
etal.,J. ChemPhys. 1983,79:926-935) used to represent water molecules. TheCHARM22
force field may be used to represent the designed peptidornimetics.
By way of example, low energy conformations may be identified using a combination
of two procedures. The first procedure involves a simulated annealing molecular dynamics
simulation approach. In this procedure, the system (which includes the designed
peptidornimetics and water molecules) is heated up to above room temperature, preferably
around 600 K, and simulated for a period of 100 picoseconds (ps) or longer; then gradually
reduced to 500 K and simulated for a period of 100 ps or longer, then gradually reduced to
400 K and simulated for a period of 100 ps or longer; gradually reduced to 300 K and
simulated for a period of 500 ps or longer. The trajectories are recorded for analysis. This
simulated annealing procedure is known for its ability for efficient conformational search.
The second procedure involves the use of the self-guided molecular dynamics
(SGMD) method (Wu & Wang, /. Physical Chemistry 1998,102:7238-7250). The SGMD
method has been demonstrated to have an extremely .enhanced conformational searching
capability. Using the SGMD method, simulation may be performed at 300 K for 1000 ps or
longer and the trajectories recorded for analysis.
Conformational analysis may be carried out using the QUANTA molecular modeling
package. First, cluster analysis may be performed using the trajectories generated from
molecular dynamic simulations. From each cluster, the lowest energy conformation may be
selected as the representative conformation for this cluster and may be compared to other
conformational clusters. Upon cluster analysis, major conformational clusters may be
identified and compared to the solution conformations of the cyclic peptide(s). The
conformational comparison may be carried out using the Molecular Similarity module within
the QUANTA program.

Similarity in structure may also be evaluated by visual comparison of the three-
dimensional structures displayed in a graphical format, or by any of a variety of
computational comparisons. For example, an atom equivalency maybe defined in the
peptidomimetic and cyclic peptide three-dimensional structures, and a fitting operation used
to establish the level of similarity. As used herein, an "atom equivalency" is a set of
conserved atoms in the two structures. A "fitting operation" may be any process by which a
candidate compound structure is translated and rotated to obtain an optimum fit with the
cyclic peptide structure. A fitting operation maybe a rigid fitting operation (e.g., the cyclic
peptide three-dimensional structure can be kept rigid and the three-dimensional structure of
the peph^omimetic can be translated and rotated to obtain an optimum fit with the cyclic
peptide). Alternatively, the fitting operation may use a least squares fitting algorithm that
computes the optimum translation and rotation to be applied to the moving compound
structure, such that the root mean square difference of the fit over the specified pairs of
equivalent atoms is a minimum. Preferably, atom equivalencies may be established by the
user and the fitting operation is performed using any of a variety of available software
applications (e.g., QUANTA, available from Molecular Simulations Inc., San Diego, Calif).
Three-dimensional structures of candidate compounds for use in estabhshing substantial
similarity may be determined experimentally (e.g., using NMR techniques as described
herein or x-ray crystallography), or may be computer-generated using, for example, methods
provided herein.
Certain peptidomimetics may be designed, based on the cyclic peptide structure. For
example, such peptidomimetics may mimic the local topography about the cleavable amide
bonds (amide bond isosteres). Examples of backbone modifications are given in Figures 3A
and 3B (see also, Figures 4A-4B in WO 01/53331). These mimetics often match the peptide
backbone atom-for-atom, while retaining functionality that makes important contacts with
the binding sites. Amide bond rnimetics may also include the incorporation of unusual
amino acids or dipeptide surrogates. Examples of such unusual amino acids and dipeptide
surrogates are illustrated here, in Figure 4 (see also Figure 5 in WO 01/53331). Still other
examples are well known in the art (see, for example, in Gillespie et al.t Biopolymers 1997,
43:191-217). The conformationally rigid substructural elements found in these types of
mimetics are believed to result in binding with highly favorable entropic driving forces, as

compared to the more confonnationally flexible peptide linkages. Backbone modifications
can also impart metabolic stability towards peptidase cleavage relative to the parent peptide.
Other peptidomimetics may be secondary structure mimics. Such peptidomimetics generally
employ non-peptide structures to replace specific secondary structures, such as J3-turns, j3-
sheets and a-turns (see Figure 5).
To design a peptidomimetic, heuristic rules that have been developed through
experience may be used to systematically modify a cyclic peptide. Within such modification,
empirical data of various kinds are generally collected throughout an iterative refinement
process. As noted above, optimal efficiency in peptidomimetic design requires a three-
dimensional structure of the pharmacophore.
Pharmacophores as provided herein permit structure-based peptidomimetic design
through, for example, peptide scaffold modification as described above. Certain
peptidomimetics may be identified through visual inspection of one or more
•pharmacophores, as compared to the neurotrophin RGE conformation. Peptidomimetics can
also be designed based on a visual comparison of a cyclic peptide pharmacophore with a
three-dimensional structure of a candidate compound, using knowledge of the structure-
activity relationships of the cyclic peptide. Structure-activity studies have established .
important binding elements in the cyclic peptides, and have permitted the development of
pharmacophore models. Peptidomimetics designed in this manner should retain these binding
elements.
Peptidomimetics may also be designed around replacing the disulfide bond (-S--S-)
with a thioether (--S--CH2~C(0)--). The disulfide bond in general is not very stable as it can
readily be reduced under acidic conditions. Replacing the disulfide bond with a thioether
moiety (~S—CH2—C(0)—) can significantly improve the stability of the peptide and therefore
the oral availability.
As an alternative to design by visual inspection, libraries (e.g., containing hydantoin
and/or oxopiperazine compounds) may be made using combinatorial chemical techniques.
Combinatorial chemical technology enables the parallel synthesis of organic compounds
through the systematic addition of defined chemical components using highly reliable
chemical reactions and robotic instrumentation. Large libraries of compounds result from the
combination of all possible reactions that can be done at one site with all the possible

reactions that can be done at a second, third or greater number of sites. Combinatorial
chemical methods can potentially generate tens to hundreds of millions of new chemical
compounds as mixtures, attached to a solid support, or as individual compounds.
Pharmacophores can be used to facilitate the screening of such chemical libraries.
For example, instead of producing all possible members of every library (resulting in an
unwieldy number of compounds), library synthesis can focus on the library members with the
greatest probability of interacting with the target. The integrated application of structure-
based design and combinatorial chemical technologies can produce synergistic improvements
in the efficiency of drug discovery.
Further peptidomimetics are compounds that appear to be unrelated to the original
peptide, but contain functional groups positioned on a nonpeptide scaffold that serve as
topographical mimics. This type of peptidomimetic is referred to herein as a "non-peptidyl
analogue." Such peptidomimetics may be identified using library screens of large chemical
databases. Such screens use the three-dimensional conformation of a pharmacophore to
search such databases in three-dimensional space. A single three-dimensional structure may
be used as a pharmacophore model in such a search. Alternatively, a pharmacophore model
may be generated by considering the crucial chemical structural features present within
multiple three-dimensional structures.
Any of a variety of databases of three-dimensional structures may be used for such
searches. A database of three-dimensional structures may be prepared by generating three-
dimensional structures of a database of compounds, and storing the three-dimensional
structures in the form of data storage material encoded with machine-readable data. The
three-dimensional structures can be displayed on a machine capable of displaying a graphical
three-dimensional representation and programmed with instructions for using the data.
Within preferred embodiments, three-dimensional structures are supplied as a set of
coordinates that define the three-dimensional structure.
Preferably, theSD-database contains at least 100,000 compounds, with small, non-
peptidyl molecules having relatively simple chemical structures particularly preferred. It is
also important that the 3D co-ordinates of the compounds in the database be accurately and
correctly represented. The National Cancer Institute (NCI) 3D-database (Milne et al, J.
Chan. Inf. Comput. Sci. 1994,34:1219-1224) and the Available Chemicals Directory (ACD;

available from MDL Information Systems, San Leandro, Calif.) are two excellent databases
that can be used to generate a database of three-dimensional structures, using molecular
modeling, as discussed above. For flexible molecules, which can have several low-energy
conformations, it is desirable to store and search multiple conformations. The Chem-X
program (Oxford Molecular Group PLC; Oxford UK) is capable of searching thousands or
even millions of conformations for a flexible compound. This capability of Chem-X
provides a real advantage in dealing with compounds that can adopt multiple conformations.
Using this approach, although the NCI-3D database presently contains a total of 465,000
compounds, hundreds of millions of conformations can be searched in a 3D-pharmacophore
. searching process.
A pharmacophore search typically involves three steps. The first step is the
generation of a pharmacophore query. Such queries may be developed from an evaluation of
critical distances in the three dimensional structure of a cyclic peptide. Using the
pharmacophore query of interest, a distance bit screening is performed on the database to
identify compounds that fulfill the required geometrical constraints. In other words,
compounds that satisfy the specified critical pair-wise distances are identified. After a
compound passed the distance bit screening step, the program next checks whether the
compound meets the substructural requirements as specified in the pharmacophore query.
After a compound passes this sub-structural check, it is finally subjected to a conformational
analysis. In this step, conformations are generated and evaluated with regard to geometric
requirements specified in the pharmacophore query. Compounds that have at least one
conformation satisfying the geometric requirements, are considered as "hits' and are recorded
in a result database.
Other criteria, which will be apparent to those of ordinary skill in the art, may also be
considered when selecting specific compounds for particular applications, such as the
simplicity of the chemical structure, low molecular weight, chemical structure diversity and
water solubility. The application of such criteria is well understood by medicinal,
computational and structural chemists.
It will be apparent that a compound structure may be optimized using screens as
provided herein. Within such screens, the effect of specific alterations of a candidate
compound on three-dimensional structure may be evaluated, in order to optimize three-

dimensional similarity to a cyclic peptide. Such alterations include, for example, changes in
hydrophobicity, steric bulk, electrostatic properties, size and bond angle.
Biological testing of candidate compounds maybe used to confmii peptidomimetic
activity. In general, peptidomimetics should function in a substantially similar manner as a
structurally similar cyclic peptide. In other words, a peptidomimetic of the cyclic peptide N-
Ac-CSRRGEC-NH2 (SEQ ID NO:2) should bind to a TRK with an affinity that is at least
half the affinity of the cyclic peptide N-Ac-CSRRGEC-NH2 (SEQ ID NO:2), as measured
using standard binding assays. Further, a peptidomimetic of the cyclic peptide N-Ac-
CSRRGEC-NH2 (SEQ ID NO:2) should modulate a TRK-mediated function using a
representative assay provided herein at a level that is at least half the level of modulation
achieved using N-Ac-CSRRGEC-NH2 (SEQ ID NO:2).
Once an active peptidomimetic has been identified, related, analogues may be
identified using two-dimensional similarity searching. Such searching may be performed, for
example, using the program ISIS Base (Molecular Design Limited). Two-dimensional
similarity searching permits the identification of other available, closely related compounds,
which may be readily screened to optimize biological activity.
6.4. TRK MODULATING AGENTS
As noted above, the term "Trk modulator" is used here .to describe any molecule
comprising at least one cyclic peptide or peptidomimetic compound of the invention
containing the neurotrophin motif RGE {i.e., Arg-Gly-Glu). Multiple cyclic peptides and/or
peptidomimetics can be present in a modulating agent of the invention. Moreover, additional
RGE sequences (for example, tandem repeats of RGE sequences) maybe included in a
modulating agent.
Linkers may or may not be used to separate RGE sequences in a Trk modulator,
including tandem repeats of RGE sequences (such as in preferred Trk agonists of the
invention). Linkers can also be used to attach a modulating agent of the invention to a solid
support or material, as described below.
A linker may be any molecule (including peptide and/or non-peptide sequences as
well as single amino acids or other molecules), that does not contain a RGE sequence and
that can be covalently linked to at least two peptide sequences and/or peptidomimetics.

Using a linker, peptidomimetics and other peptide or protein sequences may be joined in a
variety of orientations.
Linkers preferably produce a distance between CAR sequences and/or
peptidomimetics between 0.1 to 10,000 run, more preferably about 0.1-400 nm. A separation
distance between recognition sites may generally be determined according to the desired
function of the modulating agent. For Trk antagonists, the linker distance should be small
(0.1-400 ran). For Trk agonists, the linker distance should be 400-10,000 nm. One linker that
can be used for such purposes is (H2N(CH2)nC02H)m, or derivatives thereof, where n ranges
from 1 to about 10 and m ranges from 1 to about 4000. For example, if glycine
(H2NCH2CO2H) or a multimer thereof is used as a linker, each glycine unit corresponds to a
linking distance of about 2.45 angstroms, or 0.245 nm, as determined by calculation of its
lowest energy conformation when linked to other amino acids using molecular modeling
techniques. Similarly, aminopropanoic acid corresponds to a linking distance of about 3.73 •
angstroms, aminobutanoic acid to about 4.96 angstroms, aminopentanoic acid to about 6.30
angstroms and amino hexanoic acid to about 6.12 angstroms. Other linkers that may be used
will be apparent to those of ordinary skill in the art and include, for example, linkers based
on repeat units of 2,3-diaminopropanoic acid, lysine and/or ormlhine. 2,3-Diaminopropanoic
acid can provide a linking distance of either 2.51 or 3.11 angstroms depending on whether
the side-chain amino or terminal amino is used in the linkage. Similarly, lysine can provide
linking distances of either 2.44 or 6.95 angstroms and ormthine 2.44 or 5.61 angstroms.
Peptide and non-peptide linkers may generally be incorporated into a modulating agent using
any appropriate method known in the art..
Modulating agents that are Trk antagonists may contain one or more
peptidomimetics. Preferably such peptidomimetics are adjacent to one another (i.e., without
intervening sequences) or are in close proximity (i.e., separated by peptide and/or non-
peptide linkers to give a distance between the peptidomimetics that ranges from about 0.1 to
400 nm). It will be apparent that other neurotrophin sequences, as discussed above, may also
be included.
As noted above, a modulating agent may consist entirely of one or more
peptidomimetics, or may contain additional peptide and/or non-peptide components. Peptide
portions may be synthesized as described above or may be prepared using recombinant

methods. Within such methods, all or part of a modulating agent can be synthesized in living
cells, using any of a variety of expression vectors known to those of ordinary skill in the art
to be appropriate for the particular host cell. Suitable host cells may include bacteria, yeast
cells, mammalian cells, insect cells, plant cells, algae and other animal cells (e.g., hybridoma,
CHO, myeloma). The DNA sequences expressed in this manner may encode portions of an
endogenous neurotrophin. Such sequences may be prepared based on known Cdna or
genomic sequences, or from sequences isolated by screening an appropriate library with
probes designed based on the sequences of known cadherins. Such screens may generally be
performed as described in Sambrook et al, Molecular Cloning: A Laboratory Manual, Cold
Spring Harbor Laboratories, Cold Spring Harbor, N.Y., 1989 (and references cited therein).
Polymerase chain reaction (PCR) may also be employed, using oligonucleotide primers in
methods well known in the art, to isolate nucleic acid molecules encoding all or a portion of
an endogenous neurotrophin. To generate a nucleic acid molecule encoding a peptide portion
of a modulating agent, an endogenous sequence may be modified using well known
techniques. Alternatively, portions of the desired nucleic acid sequences may be synthesized
using well known techniques, and then ligated together to form a sequence encoding a
portion of the modulating agent.
Trk modulating agents of the present invention may additionally comprise an
antibody, or antigen-binding fragment thereof, that specifically binds to a NT sequence or,
alternatively an antibody or antigen-binding fragment thereof that specifically binds to a Trk
receptor sequence. As used herein, an antibody, or antigen-binding fragment thereof, is said
to "specifically bind" to a NT or Trk sequence (with or without flanking amino acids) if it
reacts at a detectable level (within, for example, an ELISA, as described by Newton et ah,
Develop. Dynamics 1993,197:1-13) with a peptide containing that sequence, and does not
react detectably with peptides containing a different NT or Trk sequence, nor with a
sequence in which the order of amino acid residues in the NT (or Trk) and/or flanking
sequence is altered.
Antibodies and fragments thereof may be prepared using standard techniques. See,
e.g., Harlow & Lane, Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory,
1988. In one such technique, an immunogen comprising a NT or Trk sequence is initially
injected into any of a wide variety of mammals (e.g., mice, rats, rabbits, sheep or goats).

Small immimogens (i.e., less than about 20 amino acids) are preferably joined to a carrier
protein, such as bovine serum albumin or keyhole limpet hemocyanin. Following one or
more injections, the animals are bled periodically. Polyclonal antibodies specific for the NT
or Trk sequence may then be purified from such antisera by, for example, affinity
chromatography using the modulating agent or antigenic portion thereof coupled to a suitable
solid support.
Monoclonal antibodies specific for a NT (or Trk) sequence may be prepared, for
example, using the technique of Kohler & Milstein, (Eur. J. Immunol 1976,6:511-519) and
improvements thereto. Briefly, these methods involve the preparation of immortal cell lines
capable of producing antibodies having the desired specificity from spleen cells obtained
from an animal immunized as described above. The spleen cells are immortalized by, for
example, fusion with a myeloma cell fusion partner, preferably one that is syngeneic with the
immunized animal. Single colonies are selected and their culture supernatants tested for
binding activity against the modulating agent or antigenic portion thereof. Hybridomas
having high reactivity and specificity are preferred.
Monoclonal antibodies maybe isolated from the supernatants of growing hybridoma
colonies, with or without the use of various techniques known in the art to enhance the yield.
Contaminants may be removed from the antibodies by conventional techniques, such as
chromatography, gel filtration, precipitation, and extraction. Antibodies having the desired
activity may generally be identified using immunofluorescence analyses of tissue sections,
cell or other samples where the target cadherin is localized.
Within certain embodiments, monoclonal antibodies may be specific for particular
NTs or, alternatively, for particular Trk receptors. For example, the antibody may bind to
NGF, but do not bind to BNDF, or vice versa. As another example, a monoclonal antibody
may bind specifically to TrkB and not bind specifically to TrkA, or vice versa. Such
antibodies may be prepared as described above, using (to generate antibodies for a particular
NT) an immunogen that comprises the RGE sequence and also sufficient flanking sequence
to generate the desired specificity (e.g., 5 amino acids on each side is generally sufficient).
To evaluate the specificity of a particular antibody, representative assays as described herein
and/or conventional antigen-binding assays may be employed. Such antibodies may
generally be used for therapeutic, diagnostic and assay purposes, as described herein. For

example, such antibodies maybe linked to a drug and administered to a mammal to target the
drug to a particular Trk -expressing cell, such as a particular neuronal cell.
Within certain embodiments, the use of antigen-binding fragments of antibodies may
be preferred. Such fragments include Fab fragments, which may be prepared using standard
techniques. Briefly, immunoglobulins may be purified from rabbit serum by affinity
chromatography on Protein A bead columns (Harlow & Lane, Antibodies: A Laboratory
Manual, Cold Spring Harbor Laboratory, 1988; see especially page 309) and digested by
papain to yield Fab and Fc fragments. The Fab and Fc fragments may be separated by
affinity chromatography on protein A bead columns (Harlow & Lane, 1988, pages 628-29).
6.5. EVALUATION OF TRK MODULATOR ACTIVITY
As noted above, peptidomimetics, cyclic peptides and other Trk modulators of the
invention are capable of modulating (i.e., enhancing or inhibiting) Trk mediated activities
including, for example, neuronal survival, axonal growth and synaptic plasticity. Hence, the
ability of a modulating agent (or a suspected modulating agent) to modulate Trk mediated
activity can generally be evaluated either in vitro or in vivo by assaying one or more of these
effects. Generally speaking, a test compound is a Trk antagonist if, within such a
representative assay, contact of test cells with the .candidate results in a discernible disruption
of the Trk mediated activity being measured. A candidate compound is generally considered
a Trk agonist if, within such a representative assay, contact of test cells with the candidate
compound results in a discernible increase of the Trk mediated activity measured.
In particular, preferred embodiments of the invention, the activity of a Trk modulator
or candidate compound is evaluated in vivo in a neurite outgrowth assay. Within a
representative neurite outgrowth assay, which is demonstrated in the Examples, infra,
neurons may be cultured on a monolayer of cells (preferably 3T3 cells or cell lines derived
therefrom). As an example, monolayers of 3T3 fibroblasts can be established by overnight
culture of cells (preferably about 80,000) in individual wells of an 8-chamber well tissue
culture slide. Approximately 3,000 cerebellar neurons isolated from post natal day 3 (PND3)
mouse brains maybe cultured for 18 hours on the various monolayers in control media
(SATO/2% FCS) or in media supplemented with various concentrations of the candidate
modulating agent. Alternatively, the cells may be cultured in media supplemented with a

control peptide (for example, a non-cyclic, linear peptide having the same amino acid
sequence as a Trk modulator cyclic peptide) or with a neurotrophin (e.g, NGF, BDNF, NT-3,
NT-4,NT-5orNT-4/5).
The cell cultures may then be fixed and stained for GAP43 or with some other agent
that specifically binds neurons and their neurites. The length of the longest neurite on each
GAP43 positive neuron may then be measured, preferably using computer assisted
morphometry. A compound that is a Trk modulator will generally modulate (e.g., inhibit or
enhance) neurite outgrowth in such a cell culture assay.
6.6. USES OF TRK RECEPTOR MODULATORS
In general, modulating agents and compositions of the present invention can be used
for modulating (e.g., inhibiting or enhancing) activities that are mediated by a Trk receptor -
including activities mediated by TrkA, TrkB and/or TrkC. Trk receptors are implicated in
the growth and repair of the central nervous system (CNS) and mediate, at least in part, such
processes as neuronal survival, axonal growth, neurite outgrowth, synaptic plasticity and,
more generally, neurological growth. Hence, modulating agents and compositions of the
invention can be used to modulate any of these processes. Such uses include, inter alia,
therapeutic methods and pharmaceutical compositions for treating conditions, diseases and
disorders that are associated with such processes. Exemplary conditions, diseases and
disorders include Alzheimer's disease, Parkinson's disease, stroke, and head and spinal cord
injury to name a few.
In one embodiment of the invention, Trk agonist of the invention can be used to
increase or enhance activities that are mediated by a Trk receptor. Hence, Trk agonist of the
invention may be used, e.g., to increase or enhance the growth and/or repair of the CNS, for
example by increasing or enhancing such processes a neuronal growth, neuronal survival,
axonal growth, neurite outgrowth and synaptic plasticity. Trk agonists of the invention are
therefore useful, e.g., in therapeutic methods for treating diseases and disorders that involve
or are otherwise associated with damage to or impaired function of the central nervous
system. These include, inter alia, the disease and disorders listed above.
In other embodiments, Trk antagonists of the invention can be used to decrease or
inhibit activity mediated by a Trk receptor. Thus, Trk antagonists may inhibit processes such

as neuronal growth, neuronal survival, axanal growth, neurits outgrowth and synaptic
plasticity. Trk antagonists are also useful in therapeutic methods, for example to treat or
ameliorate diseases and disorders (for example, epilepsy) that are associated either with
increased Trk receptor activity, or with increased activity of a neurotrophin (for example,
BDNF) that binds to and activates a Trk receptor.
In still other embodiments, Trk agonists and antagonists of the invention can also be
used to modulate responses that inhibit CNS growth and repair (i. e., "CNS inhibitors"),
including responses that inhibit processes such as neuronal growth, neuronal survival, axonal
growth, neurite outgrowth and synaptic.plasticity. In particularly preferred embodiments,
Trk agonists of the invention (for example a BAG or other agonist polypeptide or peptide
mimetic) can be used to block or reduce a CNS inhibitor response, m other embodiments of
the invention, Trk agonists (for example, a BAG or other agonist polypeptide or peptide
mimetic) can be used to enhance and/or promote neuronal growth and recovery, even
administered in an inhibitory environment such as in the presence of one or more CNS
inhibitors.
As a particular example, it is understood that inhibitory factors, such as those
associated with myelin, exists which can inhibit or even prevent processes of CNS growth
and repair, including those recited above. Examples of such inhibitors include, but are not
limited to, the myelin associated glycoprotein (also referred to as "MAG")* NogOr-A and the
oligodendrocyte myelin glycoprotein. For a more complete description of such inhibitors,
see also Section 3,3 above. Trk agonists and antagonists of the invention can be used to
modulate responses that are produced by these and other CNS inhibitors.
Without being limited to any particular theory or mechanism of action, it is
understood that Trk receptors modulate CNS growth and repair at least in part by a
mechanism or mechanisms that involve protein kinase A (PKA) and phosphinositide 3-
kinase (PI3K). Accordingly, Trk agonists and antagonists of the invention can, in preferred
embodiments, modulate effects of inhibitory signals that are mediated by one or more
components which are themselves modulated by either PKA or PI3K. As an example, and
not by way of limitation, PKA is understood to activate Rho by direct phosphorylation on
I Serl88 of that molecule (Ellerbroek et al, J. Biol. Chen. 2003,278:19023-19031). Hence,
Trk agonists and antagonists of the present invention can be used to modulate signals

mediated by inhibitory cascades involving Rho. These include, inter alia, inhibitory signals
mediated by myelin inhibitors such as MAG (and MAG fusion constructs such as MAG-Fc),
Nogo-A, the oligodendrocyte myelin glycoprotein, NgR, GTlb and p75NTR. Other CNS
inhibitors involving Rho include signals mediated by chondroitin sulfate proteoglycans from
CNS glial scar (Moirriier et al, Neurosci. 2003,22:319-330) and, as such, these CNS
inhibitors can also be modulated by Trk agonists and antagonists of the invention. As
another non-limiting example, activation of PDK is expected to overcome inhibitory activity
of semaphorins (Eickholt et al, J. Cell Biol 2002,157:211-217). Hence, Trk agonists and
antagonists of the present invention can additionally be used to modulate these CNS
inhibitors.
In general, methods of the invention involve contacting a cell expressing a Trk
receptor (typically a neuronal cell) with a Trk modulating agent either in vivo or in vitro. The
amount of Trk modulating agent administered should be an "effective amount" - that is to
say, it should be an amount that effectively modulates a Trk mediated activity of interest or,
alternatively, an amount that effectively modulates a CNS inhibitor of interest. In
embodiments where the Trk modulator is administered as part of a therapeutic method,
amount administered should be an amount that effectively ameliorates (but does not
necessarily eliminate or cure) the condition, disease or disorder being treated. Alternatively,
the amount administered may be an amount effective to ameliorate (but not necessarily
eliminate) one or more symptoms associated with the condition, disease or disorder being
treated.
As a particular, non-limiting example, a Trk modulating agent of the invention can be
used to modulate {e.g., inhibit or enhance) neurological growth, such as neurite outgrowth.
In such methods, neurite outgrowth may be enhanced and/or directed by contacting a neuron
with one or more Trk agonists of the invention (e.g., the cyclic peptide N-Ac-
CSRRGELLAASRRGELC-NHA Alternatively, neurite outgrowth may be inhibited and/or
decreased by contacting a neuron with one or more Trk antagonists of the invention {e.g., the
cyclic peptide N-Ac-CSRRGEC-NH?). Preferred modulating agents for use within such
methods are preferably linked to a polymeric matrix or other support, and comprise a cyclic
peptide as described in Section 6.1, supra, or a pepn\Iomimetic thereof (as described in

Section 6.3). Modulating agents comprising 'antibodies, or fragments thereof may also be
used in such methods, with or without the use of linkers or support materials.
The method of achieving contact to the neuronal cell and the amount of Trk
modulating agent administered will depend upon the location of the neuron as well as the
extent and nature of desired outgrowth (or, where Trk antagonists are administered, the
extend and nature of desired inhibition). For example, a neuron may be contacted (e.g., via
implantation) with one or more Trk modulating agents linked to a support material such as a
suture, fiber nerve guide or other prosthetic device so that the neurite outgrowth is directed
along the support material. Alternatively, a tubular nerve guide maybe employed in which
the lumen of the nerve guide contains a composition comprising the modulating agent or
agents. In vivo, such nerve guides or other supported modulating agents may be implanted
using well known techniques to, for example, facilitate the growth of severed neuronal
connections and/or to treat spinal cord injuries. It will be apparent to those of ordinary skill
in the art that the structure and composition of the support should be appropriate for the
particular injury being treated. In vitro, a polymeric matrix may similarly be used to direct
the growth of neurons onto patterned surfaces as described, for example, in U.S. Patent No.
5,510,628.
6.7. TRK RECEPTOR MODULATORS: FORMULATIONS
In certain embodiments, a modulating agent as described herein may, but need not, be
linked to one or more additional molecules. For example, it may be beneficial for certain
applications to link multiple modulating agents (which may, but need not, be identical) to a
support molecule (e.g., keyhole limpet hemocyanin) or a solid support, such as a polymeric
matrix (which maybe formulated as a membrane or microstructure, such as an ultra thin
film), a container surface (e.g., the surface of a tissue culture plate or the interior surface of a
bioreactor), or a bead or other particle, which may be prepared.from a variety of materials
including glass, plastic or ceramics. For certain appHcations, biodegradable support
materials are preferred, such as cellulose and derivatives thereof, collagen, spider silk or any
of a variety of polyesters (e.g., those derived from hydroxy acids and/or lactones) or sutures
(see U.S. Pat. No. 5,245,012).

Suitable methods for linking a modulating agent to a support material will depend
upon the composition of the support and the intended use, and will be readily apparent to
those of ordinary skill in the art. Attachment may generally be achieved through noncovalent
association, such as adsorption or affinity or, preferably, via covalent attachment (which may
be a direct linkage between a modulating agent and functional groups on the support, or may
be a linkage by way of a cross-linking agent or linker). Attachment of a modulating agent by
adsorption may be achieved by contact, in a suitable buffer, with a solid support for a suitable
amount of time. The contact time varies with temperature, but is generally between about 5
seconds and 1 day, and typically between about 10 seconds and 1 hour.
Covalent attachment of a modulating agent to a molecule or solid support may
generally be achieved by first reacting the support material with a bifunctional reagent that
will also react with a functional group, such as a hydroxyl, thiol, carboxyl, ketone or amino
group, on the modulating agent. For example, a modulating agent may be bound to an
appropriate polymeric support or coating using benzoquinone, by condensation of an
aldehyde group on the support with an amine and an active hydrogen on the modulating
agent or by condensation of an amino group on the support with a carboxylic acid on the
modulating agent A preferred method of generating a linkage is via amino groups using
glutaraldehyde. A modulating agent may be linked to cellulose via ester linkages. Similarly, •
amide linkages may be suitable for linkage to other molecules such as keyhole limpet
hemocyanin or other support materials. Multiple modulating agents and/or molecules
comprising other NT and/or Trk receptor sequences may be attached, for example, by
random coupling, in which equimolar amounts of such molecules are mixed with a matrix
support and allowed to couple at random.
Although modulating agents as described herein may preferentially bind to specific
tissues or cells (i.e., neuronal cells and tissues), and thus may be sufficient to target a desired
site in vivo, it may be beneficial for certain applications to include an additional targeting
agent. Accordingly, a targeting agent may also, or alternatively, be linked to a modulating
agent to facilitate targeting to one or more specific tissues. As used herein, a "targeting
agent," may be any substance (such as a compound or cell) that, when linked to a modulating
agent enhances the transport of the modulating agent to a target tissue, thereby increasing the
local concentration of the modulating agent- Targeting agents include antibodies or

fragments thereof receptors, ligands and other molecules that bind to cells of, or in the
vicinity of, the target tissue. Known targeting agents include serum hormones, antibodies
against cell surface antigens, lectins, adhesion molecules, tumor cell surface binding ligands,
steroids, cholesterol, lymphokines, fibrinolytic enzymes and those drugs and proteins that
bind to a desired target site. An antibody targeting agent may be an intact (whole) molecule,
a fragment thereof, or a functional equivalent thereof. Examples of antibody fragments are
F(ab')2,-Fab\ Fab and F[v] fragments, which may be produced by conventional methods or
by genetic or protein engineering. Linkage is generally covalent and may be achieved by, for
example, direct condensation or other reactions, or by way of bi- or multi-functional linkers.
Within other embodiments, it may also be possible to target a polynucleotide encoding a
modulating agent to a target tissue, thereby increasing the local concentration of modulating
agent. Such targeting may be achieved using well known techniques, including retroviral and
adenoviral infection.
For certain embodiments, it may be beneficial to also, or alternatively, link a drug to a
modulating agent. As used herein, the term "drug" refers to any bioactive agent intended for
administration to a mammal to prevent or treat a disease or other undesirable condition.
Drugs include hormones, growth factors, proteins, peptides -and other compounds. The use
of certain specific drugs within the context of the present invention is discussed below.
Within certain aspects of the present invention, one or more modulating agents as
described herein may be present within a pharmaceutical composition. A pharmaceutical
composition comprises one or more modulating agents in combination with one or more
pharmaceutically or physiologically acceptable carriers, diluents or excipients. Such
compositions may comprise buffers (e.g., neutral buffered saline or phosphate buffered
saline), carbohydrates (e.g., glucose, mannose, sucrose or dextrans), mannitol, proteins,
polypeptides or amino acids such as glycine, antioxidants, chelating agents such as EDTA or
glutathione, adjuvants (e.g., aluminum hydroxide) and/or preservatives. Within yet other
embodiments, compositions of the present invention may be formulated as a lyophilizate. A
modulating agent (alone or in combination with a targeting agent and/or drug) may, but need
not, be encapsulated within liposomes using-well known technology. Compositions of the
present invention may be formulated for any appropriate manner of administration, including
for example, topical, oral, nasal, intravenous, intracranial, intraperitoneal, subcutaneous, or

intramuscular administration. For certain topical applications, formulation as a cream or
lotion, using well known components, is preferred.
Optionally, a pharmaceutical composition may also contain one or more drugs, which
may be linked to a modulating agent or may be free within the composition. Virtually any
drug may be administered in combination with a modulating agent as described herein, for a
variety of purposes as described below. Examples of types of drugs that may be
administered with a modulating agent include but are not limited to analgesics, anesthetics,
antianginals, antifungals, antibiotics, anticancer drags (e.g., taxol or mitomycin C),
antiinflammatories (e.g., ibuprofen and indomethacin), anthelmintics, antidepressants,
antidotes, antiemetics, antiMstarnines, antihypertensives, antimalarials, antimicrotubule
agents (e.g., colchicine or vinca alkaloids), antimigraine agents, antimicrobials,
antiphsychotics, antipyretics, antiseptics, anti-signaling agents (e.g., protein kinase C
inhibitors or inhibitors of intracellular calcium mobilization), antiarthritics, antithrombin
agents, antituberculotics, antitussives, antivirals, appetite suppressants, cardioactive drugs,
chemical dependency drugs, cathartics, chemotherapeutic agents, coronary, cerebral or
peripheral vasodilators, contraceptive agents, depressants, diuretics, expectorants, growth
factors, hormonal agents, hypnotics, immunosuppression agents, narcotic antagonists,
parasympathomimetics, sedativesj stimulants, sympathomimetics, toxins (e.g., cholera toxin),
tranquilizers and urinary antiinfectives.
For imaging purposes, any of a variety of diagnostic agents may be incorporated into
a pharmaceutical composition, either linked to a modulating agent or free within the
composition. Diagnostic agents include any. substance administered to iUuminate a
physiological function within a patient, while leaving other physiological functions generally
unaffected. Diagnostic agents include metals, radioactive isotopes and radioopaque agents
(e.g., gallium, technetium, indium, strontium, iodine, barium, bromine and phosphorus-
containing compounds), radiolucent agents, contrast agents, dyes (e.g., fluorescent dyes and
chromophores) and enzymes that catalyze a calorimetric or fluorometric reaction. In general,
such agents may be attached using a variety of techniques as described above, and may be
present in any orientation.
The compositions described herein may be aa'ministered as part of a sustained release
formulation (i.e., a formulation such as a capsule or sponge that effects a slow release of

modulating agent following adirunistration). Such formulations may generally be prepared
using well known technology and administered by, for example, oral, rectal or subcutaneous
implantation, or by implantation at the desired target site. Sustained-release formulations
may contain a modulating agent dispersed in a carrier matrix and/or contained within a
reservoir surrounded by a rate controlling membrane (see, e.g., European Patent Application
710,491A). Carriers for use within such formulations are biocompatible, and may also be
biodegradable; preferably the formulation provides a relatively constant level of modulating
agent release. The amount of modulating agent contained within a sustained release
formulation depends upon the site of implantation, the rate and expected duration of release
and the nature of the condition to be treated or prevented".
Pharmaceutical compositions of the present invention may be administered in a
manner appropriate to the disease to be treated (or prevented). Appropriate dosages and the
duration and frequency of administration will be determined by such factors as the condition
of the patient, the type and severity of the patient's disease, and the method of administration.
In general, an appropriate dosage and treatment regimen provides the modulating agent(s) in
an amount sufficient to provide therapeutic and/or prophylactic benefit. Within particularly
preferred embodiments of the invention, a modulating agent or pharmaceutical composition
as described herein may be administered at a dosage ranging from 0.001 to 50 mg/kg body
weight, preferably from 0.1 to 20 mg/kg, on a regimen.of single or multiple daily doses. For
topical administration, a cream typically comprises an amount of modulating agent ranging
from 0.00001% to 1%, preferably.0.0001% to 0.2%, and more preferably from 0.0001% to
0.002%. Fluid compositions typically contain about 10 ng/ml to 5 rng/ml, preferably from
about 10 .mu.g to 2 mg/Ml peptidomimetic. Appropriate dosages may generally be
determined using experimental models and/or clinical trials. In general, the use of the
minimum dosage that is sufficient to provide effective therapy is preferred. Patients may
generally be monitored for therapeutic effectiveness using assays suitable for the condition
being treated or prevented, which will be familiar to those of ordinary skill in the art.
Trk receptor modulators can also be formulated according to the description provided
in section 6.9, infra.
6.8. t)75 BINDING AGENTS

The term "p75 receptor binding agent" is used herein to describe a naturally-
occurring or synthetic (e.g., recombinant) molecule, which binds to a p75 receptor engaged
in an inhibitory complex, and interferes with p75 receptorrneurotrophin interaction but not
neurotrophin:Trk receptor interaction. Thus, a p75 receptor binding agent facilitates
neurotrophin-mediated neuron growth in an inhibitory environment. A p75 receptor is
engaged in an inhibitory complex when it interacts with a nogo receptor and any of the
myelin-associated proteins (e.g., MAG, Nogo-A, oligodendocyte myelin glycoprotein).
Examples of p75 receptor binding agents include, but are not limited to, neurotrophins, such
as NGF, and agents derived from neurotrophins, such as the NGF binding loop-derived N-
Ac-CTDIKGKEC-NH? (SEQ ID NO:43). A neurotrophin is a p75 receptor binding agent
according to the invention if it interferes with binding of another, different neurotrophin to
the p75 receptor and does not interact with the Trk receptor expressed on the injured neuron.
For example, in the case of neurons that express TrkB but not TrkA, the neurotrophin NGF is
a p75 receptor binding agent because NGF will compete (i.e., interfere) with a neurotrophin
that binds TrkB (e.g., BDNF) for p75 receptor binding but will not interfere with
neurotrophin binding (e.g., BDNF) to the TrkB receptor.
In a preferred embodiment, ap75 receptor binding agent comprises at least one cyclic
peptide or peph\ioniimetic compound containing the NGF motif TDIKGKE (i.e., Thr-Asp-
ne-Lys-Gly-Lys-Glu) (SEQ ID NO:42) within a cyclic ring ofthe cyclic peptide or
peptidomimetic compound. An especially preferred p75 receptor binding agent is N-Ac-
CTDIKGKEC-NH?. (SEQ ID NO:43). As noted previously, underlined peptide sequences
denote a peptide that has been cyclised by a covalent bond between the two last underlined
residues. In these examples, the p75 binding agents were cyclized by a disulfide bond
between two cysteine residues, acetylated and amide blocked. It is understood that preferred
peptides which bind to a p75 receptor will be constrained and, hence, are preferably cyclic
peptides. Methods for cyclization of peptides are described in section 6.1, supra.
Multiple cyclic peptides and/or peptidomimetics can be present in a p75 receptor
binding agent. Moreover, additional TDIKGKE sequences (for example, tandem repeats of
TDIKGKE sequences) may be included in ap75 receptor binding agent.
Linkers may or may not be used to separate p75 receptor binding sequences in a p75
receptor binding agent, including tandem repeats of p75 receptor binding sequences. A

linker may be any molecule (including peptide and/or non-peptide sequences as well as
single amino acids or other molecules), that can be covalently linked to at least two peptide
sequences and/or peptidomimetics, and does not contain a p75 receptor binding sequence.
Using a linker, peptidomimetics and other peptide or protein sequences may be joined in a
variety of orientations.
p75 receptor binding agents may contain one or more peptidomimetics. Preferably
such peptidomimetics are adjacent to one another (/.&, without intervening sequences) or are
in close proximity (i.e., separated by peptide and/or non-peptide linkers to give a distance
between the peptidomimetics that ranges from about 0.1 to 400 nm). A p75 receptor binding
agent may consist entirely of one or more peptidornimetics, or may contain additional peptide
and/or non-peptide components. Methods for making a peptidornimetic are described in
sections 6.2 and 63, supra.
All or part of a p75 receptor binding agent can be synthesized in living cells, using
any of a variety of expression vectors known to those of ordinary skill in the art to be
appropriate for the particular host cell Suitable host cells may include bacteria, yeast cells,
mammalian cells, insect cells, plant cells, algae and other animal cells (e.g., hybridoma,
CHO, myeloma). The DNA sequences expressed in this manner may encode portions of an
endogenous neurotrophin. Such sequences may be prepared based on known cDNA or
genomic sequences, or from sequences isolated by screening an appropriate library. Such
screens may generally be performed as described in Sambrook et al, Molecular Cloning: A
Laboratory Manual, Cold Spring Harbor Laboratories, Cold Spring Harbor, N.Y., 1989 (and
references cited therein). Polymerase chain reaction (PCR) may also be employed, using
oligonucleotide primers in methods well known in the art, to isolate nucleic acid molecules
encoding all or a portion of an endogenous neurotrophin. To generate a nucleic acid
molecule encoding a peptide portion of a modulating agent, an endogenous sequence may be
modified using well known techniques. Alternatively, portions of the desired nucleic acid
sequences may be synthesized using well known techniques, and then ligated together to
form a sequence encoding a portion of the p75 receptor binding agent.
6.9. p75 BINDING AGENTS: METHODS FOR PROMOTING CNS GROWTH

The present invention provides methods for promoting CNS. growth, which comprise
administering a p75 receptor binding agent. Trk receptors are implicated in the growth and
repair of the CNS and mediate such processes as neuronal survival, axonal growth, neurite
outgrowth, synaptic plasticity, and more generally, neurological growth. p75 receptors bind
neurotrophins with low affinity and this binding compromises the ability of neurotrophins to
activate Trk receptors in the situation where the p75 receptor is engaged in an inhibitory
complex. Hence, methods which interfere with the binding of neurotrophins to p75 receptor
allows neurotrophins to bind to and activate Trk receptors, and thus promote CNS neuron
growth in an inhibitory environment.
In an aspect of the present invention, a method is provided which comprises
administering to an individual a therapeutically effective amount of a p75 receptor binding
agent in combination with at least one neurotrophin. A preferred neurotrophin is NGF,
BDNF, NT-3, NT-4 or NT-5. In one embodiment, the p75 receptor binding agent is
administered in an amount about 10 to about 100 fold greater than that of the neurotrophin.
In another embodiment, the p75 receptor binding agent is NGF and the neurotrophin is
BDNF: The methods of the present invention can be used to treat conditions, diseases and
disorders that are associated with damage to or impaired function of the CNS. Exemplary
conditions include, but are not limited to, Alzheimer's disease,-.Parkinson's disease,
Huntington's disease, amyotrophia lateral sclerosis, stroke, traumatic brain injury, and spinal
cord injury.
According to the methods of the present invention, a neurotrophin is a p75 receptor
binding agent when the neurotrophin interferes with the binding of another, different
neurotrophin to a p75 receptor engaged in an inhibitory complex, but does not interfere with
the binding of the another, different neurotrophin to a Trk receptor expressed on an injured
CNS neuron.. For example, NGF is a p75 receptor binding agent according to the present
invention if it is co-administered with BDNF to an individual with neurons that express the
TrkB receptor because NGF competes with BDNF for binding to the p75 receptor but does
not compete with BDNF for binding to me TrkB receptor.
A p75 receptor agent as described herein can be present within a pharmaceutical
composition. A pharmaceutical composition comprises a p75 receptor binding agent in
combination with one or more pharmaceutically or physiologically acceptable carriers,

diluents or excipients. Such compositions may comprise butlers (e.g., neutral buffered saline
or phosphate buffered saline), carbohydrates (e.g., glucose, mannose, sucrose or dextrans),
manidtol, proteins, polypeptides or amino acids such as glycine, antioxidants, chelating
agents such as EDTA or glutathione, adjuvants (e.g., dirminum hydroxide) and/or
preservatives. Within yet other embodiments, compositions of the present invention may be
formulated as a lyophilizate. A p75 receptor binding agent (alone or in combination with a
targeting agent and/or drug) can be encapsulated within liposomes using well known
technology.
Compositions of the present invention may be formulated for any appropriate manner
of administration, including for example, topical, intravenous, intracranial, intraperitoneal,
subcutaneous, or intramuscular administration. Pharmaceutical compositions comprising a
p75 receptor binding agent can be administered by any means that allows the p75 receptor
binding agent to reach and bind with p75 receptors in the body of an individual.
Sterile injectable forms of pharmaceutical compositions comprising a p75 receptor
binding agent can be aqueous or oleaginous suspensions. These suspensions may be
formulated according to techniques known in the art using suitable dispersing or wetting
agents and suspending agents. A sterile injectable preparation can also be a sterile injectable
solution or suspension in a non-toxic parenterally-acceptable diluent or solvent. Among the
acceptable vehicles and solvents that can be employed are sterile water, lactated Ringer's
solution and isotonic sodium chloride solution. In addition, sterile, fixed oils are
conventionally employed as a solvent or suspending medium. For mis purpose, any bland
fixed oil may be employed including synthetic mono- or dirglycerides. Fatty acids, such as
oleic acid and its glyceride derivatives, are useful in the preparation of injectables, as are
natural pharmaceutically-acceptable oils, such as olive oil or castor oil, especially in their
polyoxyethylated versions. These oil solutions or suspensions may also contain a long-chain
alcohol diluent or dispersant. A p75 receptor binding agent can also be formulated according
to the description provided in section 6.7, supra.
p75 receptor binding agents can be administered topically: For example, a p75
receptor binding agent may be applied topically to the exposed spinal cord of an individual
following spinal cord injury or during surgery. For topical application, a pharmaceutical
composition can be formulated in a suitable ointment containing the p75 receptor binding

agent suspended or dissolved in one or more earners. Carriers for topical administration of
p75 receptor binding agents include, but are not limited to, mineral oil, liquid petrolatum,
white petrolatum, emulsifying wax, water, or absorbable materials, such as, for example,
Type I collagen gel or gelatin hemostasis sponge (Gelfoam®, Pharmacia & Upjohn,
Kalamazoo, MI).
Appropriate dosages and the duration and frequency of administration will be
determined by such factors as the condition of the patient, the type and severity of the
patient's disease and the method of administration, hi general, an appropriate dosage and
treatment regimen provides the p75 binding agent(s) in an amount sufficient to provide
therapeutic and/or prophylactic benefit. Various considerations for determining appropriate
dosages are described, e.g., in Gilman etal. (eds), The Pharmacological Bases of
Therapeutics, 8 Ed. (1990), Pergamon Press. Appropriate dosages may generally be
determined using experimental models and/or clinical trials. In general, the use of the
minimum dosage that is sufficient to provide effective therapy is preferred. Patients can be
monitored for therapeutic effectiveness using physical-examination, imaging studies, or
assays suitable for the condition being treated or prevented, which will be familiar to those of
ordinary skill in the art. Dose adjustments can be made based on the monitoring findings.
For example, an individual with a spinal cord injury associated with loss of sensation in an
arm can be monitored, following administration of a p75 receptor binding agent according to
the invention, for return of sensation to the arm by physical examination.
Compositions comprising a p75 receptor binding agent may be administered as part
of a sustained release formulation (i.e., a formulation such as a capsule or sponge that effects
a slow release of p75 receptor binding agent following administration). Such formulations
may generally be prepared using well known technology and administered by, for example,
subcutaneous implantation or implantation at the desired target site. Sustained-release
formulations may contain a p75 receptor binding agent dispersed in a carrier matrix and/or
contained within a reservoir surrounded by a rate-controlling membrane (see, e.g., European
Patent Application 710,491A). Carriers for use within such formulations are biocompatible,
and may also be biodegradable; preferably the formulation provides a relatively constant
level of binding agent release. The amount of binding agent contained within a sustained

release formulation depends upon the site of implantations the rate and expected duration of
release and the nature of the condition to be treated.
Although a p75 receptor binding agent as described herein may preferentially bind to
specific tissues or cells (i.e., neuronal cells and tissues), and thus may be sufficient to target a
desired site in vivo, it may be beneficial for certain applications to include an additional
targeting agent. Accordingly, a targeting agent can be linked to a p75 receptor binding agent
to facilitate targeting to one or more specific tissues. As used herein, a p75 receptor
"targeting agent" may be any substance (such as a compound or cell) that, when linked to a
p75 receptor binding agent, enhances the transport of the p75 receptor binding agent to a
target tissue (i.e., a damaged neuron), thereby increasing the local concentration of the p75
receptor binding agent.
Targeting agents can include antibodies or fragments thereof, receptors, ligands and
other molecules that bind to cells of, or in the vicinity of, the target tissue. Known targeting
agents include serum hormones, antibodies against cell surface antigens, lectins, adhesion
molecules, tumor cell surface binding ligands, steroids, cholesterol, lymphokines, fibrinolytic
enzymes and those drugs and proteins that bind to a desired target site. An antibody
targeting agent may be an intact (whole) molecule, a fragment thereof, or a functional
equivalent thereof. For example, a MAG, Nogo-A or myelin glycoprotein antibody can be a
targeting agent. Examples of antibody fragments are F(ab')2,-Fab', Fab and F[v] fragments,
which may be produced by conventional methods or by genetic or protein engineering.
Linkage is generally covalent and-may be achieved by, for example, direct condensation or
other reactions, or by way of bi- or multi-functional linkers. Within other embodiments, it
may also be possible to target a polynucleotide encoding a binding agent to a target tissue,
thereby increasing the local concentration of binding agent. Such targeting may be achieved
using well known techniques, including retroviral and adenoviral infection.
For certain embodiments, it may be beneficial to link a drug to a p75 receptor binding
agent. For a description of drugs suitable for linking to a p75 receptor binding agent, see
section 6.7, supra.

7. EXAMPLES
The present invention is also described and demonstrated by way of the following
examples. However, the use of these and other examples anywhere in the specification is
illustrative only and in no way limits the scope and meaning of the invention or of any
exemplified term. Likewise, the invention is not limited to any particular preferred
embodiments described here. Indeed, many modifications and variations of the invention
may be apparent to those skilled in the art upon reading this specification, and such variations
can be made without departing the invention in spirit or in scope. The invention is therefore
to be limited only by the terms of the appended claims along with the full scope of
equivalents to which those claims are entitled.
7.1. EXPERIMENTAL PROCEDURES
7.1.1 Neurite Outgrowth Assays
Co-cultures of cerebellar neurons on monolayers of either parental 3T3 cells or LK8
cells (an established transfected 3T3 cell line that expresses physiological levels of chick N-
cadherin; see Doherty et al, Neuron 1991, 6:247-258) were established as previously
described by Williams et al. (Neuron 1994,13:583-594). The cells were maintained in
Dulbecco's modified Eagle's medium supplemented with 10% foetal calf serum (FCS).
For establishment of the co-cultures, about 80,000 3T3 cells (or LK8 cells) were
plated into individual chambers of an eight-chamber tissue culture slide coated withpoly-L-
lysine and fibronectin. The plated cells were maintained overnight in Dulbecco's modified
Eagle's medium ("DMEM") supplemented with 10% FCS to allow for the formation of
confluent monolayers. The medium was removed and about 6,000 dissociated cerebellar
neurons (taken from post-natal day 9 rats) were plated into each well in SATO medium
supplemented with 2% FCS. Test reagents were added as indicated in the text and the co-
cultures maintained for 18 hours. The co-cultures were then fixed and stained for GAP-43
, immunoreactivity. The mean length of the longest neurite per cell was measured for between
about 120 and 150 neurons, again as previously described by Williams et al. (Neuron 1994,
13:583-594).

7.1.2 Molecular Modeling of Trk Receptor-Ligand Structures
X-ray crystallograpliy structures of the NGF/TrkA andNT-4/TrkB complexes were
used for molecular modeling. These structures have been previously described (See,
respectively, Wiesmann et al. Nature 1999,401:184-188; and Banfield et al. Structue
(Comb) 2001,9:1191-1199) and can be readily accessed, e.g., over the internet from the
Protein Data Bank (PDB) under the accession numbers lwww (for NGF/TrkA) and lhcf (for
iSTT-4/TrkB). Swiss PDB software packages were used to isolate the structures of various
motifs from the binding interfaces of the crystals, and Accelrys software-was used to
generate images.
Contact profiles were generated from various ligand/receptor interfaces at the NT-
1/TrkB crystal structure by measuring the average number of receptor contacts per ligand
residue against the ligand residue sequence number. The average was taken over a three
■esidue window and the contact number is the number of receptor residues that are within
ive Angstroms of the given ligand amino acid.
7.1.3 Reagents
Recombined human FGF2, BDNF and NT-4/5 were all obtained from R&D systems
Minneapolis, MN). The CB1 cannabinoid receptor agonist WJN55,2122-2 mesylate was
)btained and used as previously described (see, Williams et al, J. Cell. Biol. 2003,160:481-
1-86). The Trk receptor antagonist K252a was obtained from Calbiochem (San Diego, CA).
fne PD173074 FGFR antagonist (Mohammadi et al, Embo J. 1998,17:5986-5904) was
ynthesized and used as previously described by Skaper et al. (J. Neurochem. 2000,75:1520-
.527) and by Hamby et al, J. Med. Chern. 1997,40:2296-2303).
Synthetic peptides were all obtained from a commercial supplier (Multiple Peptide
Systems, San Diego, GA). All peptides were purified by reverse-phase high performance
iquid chromatography (RP-HPLC) according to routine methods, and obtained at the highest
evel of purity (i.e., greater than 95% pure). Where peptide sequences are underlined
throughout this specification) denotes a peptide that has been cyclised by a covalent bond
>etween the two last underlined residues. In these examples, peptides were cyclized by a
lisulfide bond between two cysteine residues, acetylated and amide blocked.

7.2. IDENTIFICATION OF A NT-4 LINEAR PEPTIDE SEQUENCE
THAT INTERACTS WITH TRKB RECEPTOR
This section describes molecular modeling experiments that identify a suitable linear
peptide (LIP) sequence from natural neurotrophin ligands that binds to a key site in a Trk
receptor.
Considerable evidence suggests that the membrane proximal immunoglobulin (Tg)
domain (D5) of the Trk receptor is directly involved in the binding of NTs. See, for example,
Perez ef at", AfoZ. CellNeiirosci. 1995,6:97-105; andUrfer et al., Embo J. 1995,14:2795-
2805. Crystal structures of the NGF/TrkA complex (Wiesmann et al, Nature 1999,401:184-
188) and of theNT-4/TrkB complex (Banfield et al, Structure (Comb) 2001,1191-1199)
have been solved which support this view. In both structures, a single NT dimer engages to
Trk receptor molecules with each NT molecule in the dimer interacting, in turn, with each
Trk receptor molecule.
Both crystal structures were analyzed with an algorithm designed to highlight linear
regions on the ligand that interact with the receptor (Doherty et al, Mol CellNeurosci 2000,
16:283-295). To illustrate this analysis, the NT-4/TrkB(D5) crystal structure is shown here
in Figure 6A. Within this complex, an individual NT monomer (labeled ai) makes linear
contacts with both TrkB receptor monomers (labeled bi and b2). Both of these two interfaces
were analyzed, and this analysis is illustrated in Figures 6B-6C. In these figures, interfaces
between NGF and the TrkA receptor are shown by dotted lines, whereas the interfaces
between NT-4 and TrkB are shown by a solid line. As can be seen from inspection of these •
two figures, the interfaces overlap considerably for these two ligand-receptor complexes.
The contact profile analysis indicates that the N-terminal of the NT ligand makes the
most intimate contact with the Trk receptor (Figure 6C). Moreover, a small linear motif
(SRRGE) situated at the dominant peak in the contact profile exists as half a helix and can be
considered a tight loop. The sequence is closely conserved in BDNF (ARRGE), which also
binds to the TrkB receptor, and is partially conserved in NGF (FHRGE) and NT-3 (SHRGE),
ligands for the TrkA and TrkC receptors, respectively. Interesting, this region of
neurotrophins is disordered in crystal structures of unbound NT (see, McDonald et al,
Nature 1991,354:411-414) and therefore has not previously been the subject of peptide
studies.

The a1/b2 interface contact profile is illustrated in Figure 6B. NT loops 1-4, winch
others have implicated in the NT/Trk interaction (LeSaateur et al., J. Biol. Chem. 1995,
270:6564-6569) are Wghlighted. However, none of these loops figure in the a1/b1 interface
and only loop 1 is involved in the a1/b2 interface.
7.3. DEVELOPMENT OF A TRKB ANTAGONIST PEPTIDE
Cerebellar neurons, isolated from rat pups at PND2, were cultured over monolayers
of 3T3 fibroblasts in either control media or in media supplemented with a range of BDNF
and/or NT-4 concentrations. After 16 hours, the co-cultures were fixed and mean neuxite
length determined as has been previously described (Williams et al, J. Biol. Chem.2000,
275:4007-4012). The results, which are illustrated in Figure 7A, show that both ligands
stimulate neurite outgrowth in a dose-dependent manner with a maximal response seen at
between about 1 and 10 ng/ml.
The results in Section 7.2, supra, suggest that a properly constrained peptide of the
small linear RGE motif that is present in all NTs might have a high structural overlap with
the native NT structure and thereby function as a Trk receptor antagonist. To test this
hypothesis, a cyclic version of the LIP was designed that was constrained by a disulfide bond
and has the amino acid sequence: N-Ac-CRGEC-NH2. The effect of this peptide on the
BDNF and NT-4 response was tested in the above-described neurite outgrowth assay, with
both NT ligands used at concentrations of 5 ng/ml. These results, which are illustrated in
Figure 7B show that the peptide antagonizes both the BDNF and the NT-4 response, with a
50% inhibition seen at 144 ± 23 μM for the BDNF response and 112 ± 22 μM for the NT-4
response. In contrast, the peptide has no effect on basal neurite outgrowth when tested at
concentrations greater than about 400 μM and in the absence of any natural NT ligaud.
These results suggest that the cyclic peptide itself has no specific effects on neuronal survival
and neurite outgrowth.
7.4. AN NT-4 LP IS A MOKE POTENT TRKB ANTAGONIST
THAN EQUIVALENT NT3 AND NGF LIPS
The RGE sequence motif is shared by all the NTs. However, the amino sequences
flanking this motif differ between the TrkA, TrkB and TrkC ligands. A series of

"equivalent" peptides, designed from sequences of different NT ligands, was therefore tested
to determine if an extended peptide from the TrkB ligand might be a more active. TrkB
receptor antagonist.
Molecular modeling studies suggest that the peptide N-Ac-CSRRGEC-NH2 shares
structural overlap with the natural SRRGE motif of NT-4. Accordingly, this peptide
sequence was tested alongside equivalent peptides derived from NGF (N-Ac-CFHRGEC-
NH2) and NT-3 (N-Ac-CSHRGEC-NH2) for their ability to inhibit the BDNF and NT-4
responses in the neurite outgrowth assay described in Section 7.3, supra. Surprisingly, the
N-Ac-CSRRGEC-NH2 derived from NT-4 was approximately 5-fold better than the NGF
and NT-3 derived peptides at inhibiting the BDNF response, with a 50% inhibition seen at
27 ±6 μM. These results are illustrated in Figure 7C.
As noted, above, the addition of as little as two flanking amino acid residues from
NT-4 increased the efficacy of the peptide up to five-fold against a TrkB response. Addition
of the equivalent amino acids from either NGF or NT-3 had no discernable effect on the
efficacy of the original cyclic RGE peptide, suggesting that the selectivity of NT binding
may be determined, at least in part, by the nature of amino acid residues that immediately
flank the RGE motif. Indeed, a considerable body of evidence suggests that specificity of
Trk receptor binding is encoded by the amino terminal sequences of the NTs. See, for
example, Urfer etal, Emba J. 1994,13:5896-5909; and Mclnnes & Sykes, Biopolymers
1997,43:339-366. These findings suggests that cyclic peptides and peptidomimetics of the
invention can be targeted to particular Trk receptors (i.e., to a TrkA, TrkB or TrkC receptor)
by selecting the RGE flanking amino acid sequences from an NT ligand that preferably binds
to the desired Trk receptor.
The same qualitative response described supra, was seen when the peptides were
tested against NT-4 (see, Figure 7D). However, whereas 50% inhibition of the BDNF
response could be obtained with only about 25 μM of the peptide N-Ac-CSRRGEC-NH2. the
same level of inhibition of the NT-4 response required the peptide to be used at about
55+4 μM.
As with the cyclic peptide N-Ac-CRGEC-NH2, neither the N-Ac-CSRRGEC-NH2
peptide nor its NGF or NT-3 equivalents had any effect on basal neurite outgrowth in control
cultures not supplemented with a NT ligand. The.peptides'were also tested for their ability to

inhibit the neurite outgrowth response stimulated by other agents, including N-cadherin,
FGF2 or a CB1 receptor agonist Cell growth responses produced by these agents have been
described elsewhere (Williams et al, J. Cell Biol 2003,160:481-486) and, in particular, are
not believed to involve Trk receptors. The results from those experiments are shown in
Figure 8. In particular, the cyclic peptides did not inhibit any of these responses, even when
administered at concentrations that fully inhibited the NT-4 and BDNF responses. These
data confirm that the cyclic peptides and peptide mimetics of this invention can fully inhibit
Trk receptor function without any non-specific effects on neurite outgrowth.
The effects of the linear peptide N-Ac-SRRGELA-NH2 were also evaluated in the
neurite outgrowth assay against NT-4, BDNF and the other agents mentioned, supra. These
results are also shown in Figure 8. As expected, whereas the cyclic version of this peptide
was a potent inhibitor of both the NT-4 and the BDNF responses, the linear peptide did not
inhibit any of the response even when tested at concentrations up to 125 μM. Hence,
peptides and peptidomimetics containing the RGE motif need to be constrained, e.g. by
disulfide bonds, to be intional Trk receptor antagonists.
7.5. DEVELOPMENT OF A TRKB AGONIST PEPTIDE
In crystal structures of the NT-4/TrkB receptor complex, the SRR.GE motif in NT-4
runs anti-parallel to itself in the NT-4 dimer. The corresponding, motif exhibits a similar anti-
parallel alignment in crystal structures of the NGF/TrkA receptor complex. Previously, a
"tandem-repeat" mimetic approach has been used to develop peptide agonist of N-cadherin.
See, Williams et al, J. Biol. Chem. 2002,277:4361-4367. The anti-parallel arrangement of
the RGE motif in neurotrophins suggests that the "tandem-repeat" approach might also be
used to develop Trk receptor agonist peptides.
Molecular modeling supports the hypothesis that a tandem repeat of the NT-4
SRRGEL sequence might be constrained in the cyclic peptide N-Ac-
CSRRGELAASRRGELC-NH2 (this peptide is also referred herein to as the "BAG" peptide)
in a manner that would allow for simultaneous engagement of two TrkB receptor monomers.
A modeled structure of the BAG peptide, which emphasizes this point, is shown here at
Figure 9. The effect of the BAG peptide on neurite outgrowth was therefore tested in the
assay described in Section 7.3 supra.

The results from such an experiment are illustrated in Figure 10A. The peptide can
be seen to stimulate neurite outgrowth in a dose-dependent manner, with an EC50 of about
300 nM and a near maximal response at about 600 Nm. As with the result to the natural
ligands BDNF and NT-4, the response of neurite outgrowth to the BAG peptide is biphasic
(compare Figures 7A and 10A). Next, the BAG peptide's ability to stimulate axonal growth
was compared to that of established growth promoting peptides. The data from these
experiments, which are illustrated in Figure 10B, demonstrate that at 6 uM the BAG peptide
promotes axonal growth by the same extent as maximally active concentrations of NT-R,
BDNF and FGF2.
7.6. TRKB ANTAGONISTS INHIBIT THE AGONIST PEPTIDE RESPONSE
To verify that the BAG peptide activates Trk receptor by binding to the same site as
the monomelic peptide antagonists (described in Sections 7.3-7.4, supra), experiments were
conducted to determine whether the peptide antagonist N-Ac-CSRRGEC-NH2, could inhibit
the BAG peptide's effects on neurite outgrowth. The results are shown in Figure 11.
At 125 u.M, the TrkB antagonist peptide can fully inhibit the activity of a maximally
active concentration of the BAG peptide. In contrast, the linear version of this peptide (i.e.,
the peptide N-AC-SRRGELA-NH2) had very little to no effect on BAG peptide activity in the
neurite outgrowth assay. K252a, a compound which is reported to be a relatively specific
Trk receptor antagonist (Tapley et al, Oncogene 1992,7:371-381), also fully inhibited the
response of neurite outgrowth to the BAG peptide. However,PD 17304, "a specific FGF
receptor antagonist, did not inhibit the response.
These data establish that "tandem-repeat" cyclic peptides and peptidomimetics, based
on the RGE motif, are specific and effective agonist of Trk. receptors.
7.7. TRK AGONISTS OVERCOME INHIBITORS OF NEURONAL GROWTH
This example describes additional experiments investigating the effect of Trk receptor
agonists under conditions that normally inhibit neuronal growth.. In particular, the
experiments demonstrate that, unlike the natural Trk receptor ligand, Trk receptor agonists of
the invention can counteract the activity of inhibitory molecules and/or their receptors.

7.7.1 Materials and Methods
Reagents and culture treatment. Unless otherwise noted here, reagents in the
experiments set forth in this Section were obtained and as set forth supra, in Section 7.1.3 et
seq. In particular, recombinant human FGF2 and BDNF were obtained from R&D systems
(Minneapolis, MN) and used at final concentrations of 5 ng/ml. The Trk receptor agonist
K252a was obtained from Calbiochem (San Diego, CA) and used at a final concentration of
100 nM. The Trk agonist peptide BAG (SEQ ID NO:18) was obtained from a commercial
supplier (Multiple Peptide Systems, San Diego CA). Recombinant MAG-Fc chimera was
obtained from R&D Systems (Minneapolis, MN) and used at a final concentration of 5-
25 μg/ml. Monoclonal antibody to GTlb (clone GMR5) was obtained from Seikagaku
America (Falmouth, MA) and used at a final concentration of 20 μg/ml. A p75NTR rabbit
polyclonal antibody was raised against the extracellular domain of that receptor as previously
described (see, Huber & Chao, Dev. Biol. 1995,167:227-238) and used at a 1:200 dilution of
serum. The known PKA inhibitors KT5720 and H-89 were obtained from Calbiochem (San
Diego, CA) and used at final concentrations of 200 and 400 nM, respectively. The known
PDK inhibitors Wortamannin and LY294002 were also obtained from Calbiochem (San
Diego, CA) and were both used at final concentrations of 10 uM. The Rho kinase inhibitor
Y27632 was obtained from Tocris (Bristol, UK) and used at 10 uM final concentration.
All reagents were diluted into the co-culture media and in general added to the
cultures just prior to plating of the neurons. The exception was- antiserum raised against the
p75NTR receptor. Instead, a high density neuronal suspension was treated with a 1:200
dilution of the serum for 60 minutes. The neurons were then diluted by a factor of about 20,
and seeded out for culture. The residual amount of p75NTR antibody in the cultures is
estimated to have been approximately a 1:5000 dilution of the serum. Separate control
experiments demonstrated that this antibody had no effect on neurite outgrowth at a dilution
of 1:1000, establishing that the 1:5000 dilution used in these experiments has, at most, a
negligible effect.
Neurite outgrowth assays. Neurite outgrowth assays were performed as described in
Section 7.1.1, supra.

7.7.2 Results
The Trk receptor agonist BAG block MAG inhibitory activity. Myelin associated
glycoprotein (MAG) has been previously shown to inhibit neurite outgrowth response from
post-natal day 2-3 rat cerebellar neurons when presented to those cells as either a transfected
molecule in the cellular substrate (Mukhopadhyay et al, Neuron 1994,13:757-767) or when
added as a soluble Fc chimeric protein (Tang et al, Mol Cell. Neurosci. 1997,9:333-346).
In furtherance of those studies, post-natal day 3 cerebellar neurons were cultured over
monolayers of LK8 cells, a 3T3 fibroblast cell line that express transfected N-cadherin and
have been previously shown to promote a robust neurite outgrowth response (Williams et al,
Neuron 1994,13:583-594). Cells were cultured with soluble MAG-Fc fusion protein present
in the culture medium at concentrations of 0, 5 or 25 μg/ml. As expected, MAG-Fc inhibited
neurite outgrowth in a dose-dependent manner, with an approximately 40% inhibition
response when present in the culture medium at a concentration of 25 μg/ml (see, Figure 12).
Previous reports have suggested that inhibitors such as MAG mediated their effect(s)
by activation of RhoA and/or its downstream effector Rho kinase. See, for example,
Dergham et al, J. Neurosci. 2002,22:6570-6570; Fournier et al, J. Neurosci. 2003,
23:1416-1423; and Lehmann et al, J. Neurosci. 1999,19:7537-7547. To confirm these
reports, cells were also cultured with the known Rho kinase inhibitor Y27632 (Narumiya et
al, Methods Enzymol 2000,325:273-284; Davies et al, Biochem. J. 2000, 351:95-105)
included in the culture medium at a concentration of 10 uM. As expected, MAG-Fc does not
inhibit neurite outgrowth under these conditions, even when present in the culture medium at
concentrations as high as 25 μg/ml (Figure 12).
Additional neurite outgrowth experiments were performed to investigate what effect,
if any, a Trk agonist might have on inhibitors such as MAG. In these experiments, neurons
were cultured with the Trk agonist polypeptide BAG (SEQ ID NO: 18, described in Section
7.5 above) present in the culture medium at a concentration of 6 uM. Surprisingly, MAG-Fc
failed to inhibit neurite outgrowth under these conditions, even when present in the culture
medium at concentrations as high as 25 μg/ml (Figure 12). By contrast, when the
neurotrophin BDNF was present in the culture medium at a concentration of 5 ng/ml there
was no measurable effect on the MAG response - i.e., MAG-Fc continued to inhibit neurite
outgrowth (Figure 12). This result is consistent with previous reports that neuron cells must

be "primed" with neurofrophins to circumvent the inhibitor activity of MAG and myelin (see,
Cai et al, Neuron 1999,22:89-101).
To further investigate the BAG polypeptide's ability to block MAG inhibitor activity,
the polypeptide was tested in neurite outgrowth assays at a variety of different concentrations
in the culture medium. Results from these experiments are depicted graphically in
Figure 13. These data show that the BAG polypeptide effectively blocks MAG inhibitor
activity when present in the culture medium at concentrations as low as 30 nM, with a half
maximal response when present at concentrations of between about 100 and 200 nM.
To confirm that the results from these experiments were not caused by any specific
MAG inhibition of the N-cadherin component in neurite outgrowth, experiments were also
performed with neurons cultured over monolayers of 3T3 fibroblasts cells that do not express
transfected N-cadherin. Bar graphs showing data from these experiments are shown in
Figure 14. Although the basal neurite outgrowth response is lower when cells are cultured
under these conditions, MAG-Fc nevertheless produces a measurable and substantial
inhibition of neurite outgrowth when present at 25 μg/ml. In the absence of MAG-Fc, basal
levels of neurite outgrowth are already robust, and the BAG polypeptide does not have a
substantial effect when present in the culture medium at a concentration of 6 μM. Inspection
of Figure 14, however, reveals that the Trk-receptor agonist at this concentration does
effectively block the MAG response, so that MAG-Fc fails to inhibit neurite outgrowth when
present at a final concentration of 25 μg/ml. As before, and again in contrast to the effect of
BAG, the neurotrophin BDNF has no apparent effect on the inhibitory response stimulated by
MAG-Fc when present at a concentration of 5 ng/ml.
These experiments demonstrate that Trk receptor agonists such as the BAG
polypeptide can be used to effectively prevent or reduce inhibitory responses produced by
signaling molecules such as MAG. The results from these experiments additionally show
that Trk receptor agonists {e.g., BAG) promote neuronal growth and recovery, even when
administered in an inhibitory environment, such as in the presence of the inhibitory signaling
molecule MAG.
BAG blocks inhibition by GTlb. The BAG polypeptide's ability to circumvent
inhibitory activity of GTlb was also treated in neurite outgrowth assays. Previous reports

have described multivalent IgM antibodies to GTlb that can inhibit neurite outgrowth from
cerebellar granule cells (Vinson et al, 1 Biol. chem. 2001,276:20280-20285). To confirm
these reports, cerebellar neurons were cultured over monolayers of N-cadherin expressing
3T3 cells in both control media and in media containing 20 μg/ml of monoclonal antibody
for GTlb. Data from these experiments are shown in the bar graph at Figure 15. Consistent
with previous reports, co-culturing the cells with 20 μg/ml of antibody robustly inhibits
neurite outgrowth under these conditions. Compare the column labeled (1) in Figure 15 to
column C in that same figure. Co-culturing cells with 10 μM of the Rho kinase inhibitor
Y27632 effectively abolishes this effect, confirming previous reports that the GTlb receptor
involves Rho kinase as a downstream effector in its signal cascade {see, Vinson et al, J. Biol
Chem. 2001,276:20280-20285). Surprisingly, when the Trk-receptor agonist BAG is present
in the culture medium (6 μM) with antibody to GTlb, the antibody's inhibitor effect is
effectively eliminated; i.e., a level of neurite outgrowth is observed which is comparable to
that seen when antibody is not present in the culture medium. Compare the column labeled
(2) in Figure 15 to column C in that same figure.
These results show that Trk receptor agonists such as the BAG polypeptide can be
used to effectively reduce or prevent inhibitory activity produced by such receptors such as
GTlb. The results from these experiments additionally show that Trk receptor agonists (e.g.,
BAG) promote neuronal growth and recovery, even when administered in an inhibitory
environment, such as in the presence of the inhibitory signaling by GTlb.
BAG Mocks inhibition byp 75N™. Because inhibitory molecules in myelin are
believed to signal either directly or indirectly via the p75NTR receptor, the BAG peptide's
ability to circumvent that receptor's inhibitory activity was also investigated. To verify, first,
that signaling from this receptor does inhibit neurite outgrowth, cerebellar neurons were
cultured over monolayers of N-cadherin expressing 3T3 cells in both control media and in
media containing polyclonal antibody to p75NTR (1:200 serum dilution).
Data from these experiments are presented in the bar graph at Figure 16. Pretreatment of the
cells with antibody for 60 minutes effectively inhibits the subsequent outgrowth of neurons,
as can be seen by visually comparing the columns labeled (1) and C in the bar graph at
Figure 16. As with MAG and GTlb, antibody to p75NTR does not elicit an inhibitory

response when the Rho kinase inhibitor Y27632 is added to neurons at a final concentration
of 10 uM immediately after antibody treatment (see column (2) in Figure 16). Likewise,
culturing the neurons with a final BAG polypeptide concentration of 6 uM also effectively
blocks the p75NTR antibody's inhibitory effect. However, culture the cells with the
neurotrophin BDNF (5 ng/ml final concentration) has no significant effect on the inhibitory
response elicited by p75NTR antibody.
Because the cell cultures may contain some residual amount of antibody (estimated to
be no more than approximately 1:5000 serum dilution) after treatment, control experiments
were performed in which cells were cultured with polyclonal antibody in the media at a
1:1000 serum dilution. The presence of antibody at this level had no measurable effect on
neurite outgrowth, demonstrating that the effects observed in these experiments are not
caused by the very low levels of residual antibody that may remain after treatment.
The results from these experiments demonstrate that Trk receptor agonists such as the
BAG polypeptide can be used to effectively reduce or prevent inhibitory responses produced
by the p75NTR pathway, he results additionally show that Trk receptor agonists (e.g., BAG)
promote neuronal growth and recovery, even when administered in an inhibitory
environment, such as in the presence of the inhibitory signaling by p75NTR.
BAG signaling is mediated by PEA and PI3K To further investigate mechanisms by
which a Trk receptor agonists may block inhibitory signals, cerebellar neurons were cultured
for 18 hours over 3T3 monolayers in control media or in media supplemented with what hade
been determined to be maximally active concentrations of either the BAG polypeptide (6 μM
final concentration), the neurotrophin BDNF (5 ng/ml final concentration) or FGF2 (5 ng/ml
final concentration). Findings from these experiments are depicted in the bar graph at
Figure 17. Under these conditions, each of the three factors (BAG, BDNF and FGF2)
enhances neurite length by about 60-70% compared to the control culture. When K252a, a
compound which is reported to be a relatively specific Trk receptor antagonist (Tapley et al,
Oncogene 1992, 7:371-381), was included in the culture media at a final concentration of
100 nM, the outgrowth response produced by both BAG and BDNF were essentially
abolished. However, the outgrowth response produced by FGF2 was unaffected, confirming
reports suggesting that FGF2 promotes neurite outgrowth by a signaling cascade that is

distinct from that of Trk receptors and, in particular, does not involve either PKA or PBK
(see, Williams era/., Cell Biol. 2003,160:481-486).
In similar experiments, neuronal cells were cultured either with the protein kinase A
(PKA) inhibitor KT5720 (200 nM final concentration) or H-89 (400 nM final concentration),
or with the phosphoinositide 3-kinase (PI3K) inhibitor Worthmannin (10 uM final
concentration) of LY294002 (10 uM final concentration) in the culture media. As with the
Trk receptor antagonist, me neurite outgrowth response to both BAG and BDNF was
essentially abolished by these kinase inhibitors. As expected, the neurite outgrowth response
to FGF2 was unaffected.
These results demonstrate that the activated Trk receptor stimulates neurite growth by
a mechanism or mechanisms that involve activation of both PKA and PI3K. Hence, Trk
agonists of this invention (e.g., the BAG polypeptide) can be effective at blocking or reducing
a wide variety of inhibitory signals. In particular, Trk agonists of the invention can be
effective at blocking inhibitory signals mediated by signal cascades with one or more
components that are inhibited or inactivated by either PKA or PI3K.
As an example, and not be way of limitation, PKA is reported to inactivate Rho by
direct phosphorylation on Serl88 of that molecule (Ellerbroek et al, J. Biol Chem. 2003,
278:19023-19031). Hence, Trk agonists of the present invention can be used to block or
reduce signals mediated by inhibitory cascades involving Rho. These include, inter alia,
inhibitory signals mediated by myelin inhibitors such as MAG (or by MAG fusion constructs
such as an MAG-Fc), Nogo-A, the oligodendrocyte myelin glycoprotein, NgR, GTlb and
p75NTR as well as signals mediated by chondroitin sulfate proteoglycans from the CNS glial
scar (Monnier et al Neurosci. 2003,22:319-330). As another non-limiting example,
activation of PI3K is expected to overcome inhibitory activity of semaphorins (Eickholt et
al, J. Cell Biol 2002,157:211-217). Indeed, neurolrophins are reported to overcome such
inhibitor signaling by activating a Trk-PBK cascade in neurons (Atwal et al, J. Neurosci.
2003,23:7602-7609). Hence, Trk agonists of the present invention can be used to block or
reduce these inhibitory signals as well.

7.8. ADDITIONAL TRK AGONIST COMPOUNDS
This example describes additional peptides and peptidomimetic compounds that are
either based on or derived from the BAG polypeptide described in the preceding examples.
Data from biological assays are also presented, demonstrating that these novel compounds
also exhibit activity as Trk receptor agonists.
7.8.1 Novel Trk receptor agonists
The following peptides and peptidomimetics were designed based on the amino acid
sequence of the BAG polypeptide described supra - i.e., CSRRGELAASRRGELC (SEQ ID
NO:17). These novel compounds, which are referred to here as 1LBAG2J HBAGI and hriBAG2,
are set forth in Table 1, below.

In Table L above, the lowercase "c" is used to denote a cyclization by a peptide or
amide bond joining the ainmo-tenninal amino acid residue to the carboxy-terminal amino
acid residue. Hence, the hBAGi polypeptide (SEQ ID NO:39) preferably comprises an amide
bond j oining the N-terminal serine residue to the C-terminal leucine residue. Similarly, the ..
peptide hriBAG2 (SEQ ID NO:41) preferably comprises an amid bond joining the N-terminal
leucine residue to the C-terminal serine residue.
The lowercase "d" in front of an amino acid residue in Table I denotes that the
residue is a D-amino acid residue (as opposed to an L-amino acid residue). Hence, the
polypeptides riBAGi and hriBAG2 preferably comprise D-amino acid residues. Indeed, all of

the amino acid residues in these polypeptides (with the exception of the glycine residues,
which are neither L nor D amino acid residues) are preferably D-amino acid residues.
It is readily apparent, upon visual inspection of the riBAGi and hriBAG2 amino acid
sequences (SEQ ID NOS:40 and 41), that these sequences are reverse sequences of the BAG
polypeptide sequence (SEQ ID NO: 17). 3h particular, and as will be appreciated by those of
skill in the art, polypeptides of the invention which comprise a sequence of D-amino acid
residues are expected to adopt three dimensional structures (z.e., "conformations") that are
substantially similar or identical to the three dimensional conformation of a polypeptide
comprising the reverse sequence of L-amino acid. Hence, in addition to the polypeptides of
L-amino acid residues described supra, the present invention also contemplates polypeptides
having the reverse sequence of D-amino acid residues. Hence, in one preferred embodiment
peptides andpepn^ornimetics of the present invention comprise sequences of L-amino acid
residues including the Arg-Gly-Glu (i.e., "RGE") motif described, supra. Accordingly, the
invention also provides, in an alternative embodiment) peptides and peptidomimetics
comprising sequences of D-amino acid residues including the short linear sequence motif
dGlu-Gly-dArg (z.e., "dEGdR").
Peptides and peptidomimetics of the invention that comprise such D-amino acid
residues are expected to be more stable and less readily degraded in vivo, e.g., by proteolytic
enzymes. Similarly, cyclic amide bonds, such as those used in the hBAG2 and hriBAG2
polypeptides, are also expected to-be less readily degraded in vivo. Shortened peptides (e.g.,
KBAG, which lacks two terminal cysteines and two central alanines compared to BAG) are
more likely to cross the blood-brain barrier. Accordingly, such peptides may be preferred,
e.g., for use in pharmaceutical compositions and administration to an individual.
7.8.2 Biological Activity
The hBAG2, riBAGi and briBAG2 polypeptides were tested in a substrate based assay, to
evaluate their ability to promote neurite outgrowth in an inhibitory environment. In
particular, and as discussed above, the myelin associated glycoprotein (MAG) has been
previously shown to inhibit neurite outgrowth response. See, for example, Mukhopadhyay et
al, Neuron 1994,13:757-767; and Tang e* al., Mot Cell. Neurosci. 1997,9:333-346. As
demonstrated in the examples, supra, Trk receptor agonists such as the BAG polypeptide are

able to block MAG inhibitory activity, and promote neurite outgrowth in that inhibitory
environment (i.e., in the presence of MAG). The data presented in the experiments described
here, demonstrate that the hBAG2> riBAGi and hriBAG2 polypeptides also block MAG
inhibitory activity and promote neurite outgrowth.
Materials and Methods.
Briefly, standard plastic 8 chamber tissue-culture slides were coated as follows with
either: (a) polylysine; (b) polylysine and a mixture of goat anti-human IgG and fibronectin;
or (c) polylysine, a mixture of goat anti-human IgG (Fc-speciflc) and fibronectin and MAG-
Fc. First, slides are coated with polylysine at 17 μg/ml in distilled water ("dEkO) for thirty
(30) minutes at room temperature. After aspirating the wells, a mixture of anti-human IgG
and/or fibronectin (both at 10 μg/ml in DMEM) is added to wells to be coated with those
compounds, and incubated for 120 minutes. The wells are again aspirated and (for wells
coated with MAG-Fc) incubated for sixty (60) minutes with MAG-Fc (0.25 μg/ml in
DMEM and 10% FCS). PND2/3 rat cerebellar neurons are then added at 15K to each well in
DMEM, 10% FCS, 25 mM KC1 and 5 ng/ml FGF2, bringing the final media volume to 300
ul in each well. The cerebellar neurons are cultured for 27 hours before fixing and staining
for GAP-43. Polylysine, goat anti-human IgG (Fc-specific) and fibronectin are available
from SIGMA (St. Louis, Missouri).
Results. The mean length of the longest neurite per neuron was determined. Basal
neurite growth of about 9 urn was observed on the polylysine substrate. Neurite growth
increased to about 24 urn on the polylysine/fibronectin substrate. Neurite growth decreased
to about 15 urn in the wells that had the additional MAG-Fc coating. Figure 18A shows a
dose response curve for the three peptidomimetics. Peptidomimetic hriBAG2 (SEQ ID
NO:40) promoted substantial dose-dependent neurite growth in the inhibitory environment
A neurite growth response can be observed at a dose of about 10 μg/ml and is almost double
the value seen in the inhibitory environment without the peptidornimetic at a dose of 33
μg/ml (the highest concentration tested). 1IBAG2 (SEQ ID NO:39) promotes neurite growth at
l a dose of 33 μg/ml. riBAGi (SEQ ID NO: 41) does not promote growth at the same
concentration.

Figure 18B shows a bar graph depicting neurite growth in the inhibitory environment
in the presence of BDNF, BAG, hriBAG2, hBAG2 or riBAG. BDNF has no effect on neurite
growth at concentrations of 10 μg/ ml and 100 μg/ml. BAG peptide promotes neurite growth
at concentrations of both 10 μg/ml and 100 u.g/ml. hriBAG2 at a concentration of 33 u.g/ml
promoted neurite growth to a substantially greater extent than BDNF, BAG peptide, and the
KBAG2 and riBAGi at any concentration. hBAG2 at a concentration of 33 μg/ml promoted
neurite growth to an extent comparable to BAG polypeptide at a concentration of 10 jig/ml.
These results show that other peptides and peptidomimetics, such as the BAG peptide
derivatives of this application, can promote neurite growth in an inhibitory environment and
to an extent that is comparable or even superior to that of the BAG polypeptide.
IJU p75 RECEPTOR BINDING AGENTS OVERCOME INHIBITION OF
NEURONAL GROWTH
This example describes experiments investigating the effect of p75 receptor binding'
agents under conditions that normally inhibit neuronal growth. In particular, the experiments
demonstrate that p75 receptor binding agents counteract the activity of inhibitory molecules
and/or their receptors.
Neurite Outgrowth Assays
Co-cultures of cerebellar neurons on monolayers of either parental 3T3 cells or LK8
cells (an established transfected 3T3 cell line that expresses physiological levels of chick N-
cadherin; see Doherty et al, Neuron 1991,6:247-258) were established as previously
described by Williams et al. {Neuron 1994,13:583-594). The cells were maintained in
Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (FCS). These
cerebellar neurons express the TrkB receptor and do not express functional levels of the
TrkA receptor.
For establishment of the co-cultures, about 80,000 3T3 cells (or LK8 cells) were
plated into individual chambers of an eight-chamber tissue culture slide coated withpoly-L-
lysine and fibronectin. The plated cells were maintained overnight in Dulbecco's modified
Eagle's medium ("DMEM") supplemented with 10% FCS to allow for the formation of
confluent monolayers. The medium was removed and about 6,000 dissociated cerebellar
neurons (taken from post-natal day 2/3 rats) were plated into each well in SATO medium

supplemented with 2%FCS. Test reagents were added as indicated in the text and the co-
eultures maintained for 23 hours. The co-cultures were then fixed and stained for GAP-43
irnmunoreactivity. The mean length of the longest neurite per cell was measured for between
about 120 and 150 neurons, again as previously described by Williams et al. (Neuron 1994,
13:583-594).
Reagents
Recombined human NGF and BDNF were obtained from R&D systems
(Minneapolis, MN). Synthetic peptides were all obtained from a commercial supplier
(Multiple Peptide Systems, San Diego, CA). All peptides were purified by reverse-phase
high performance liquid chromatography (RP-HPLC) according to routine methods, and
obtained at the highest level of purity (i.e., greater than 95% pure).
T!iep75 receptor binding agents promote neurotrophin-mediated neuron growth in
an inhibitory environment Myelin associated glycoprotein (MAG) has been previously
shown to inhibit neurite outgrowth response from post-natal day 2-3 rat cerebellar neurons
when presented to those cells as either a transfected molecule in the cellular substrate
(Mukhopadhyay et al, Neuron 1994,13:757-767) or when added as a soluble Fc chimeric
protein (Tang et al, Mol Cell. Neurosci. 1997, 9:333-346). In furtherance of those studies,
post-natal day 3 cerebellar neurons were cultured over monolayers of LK8 cells, a 3T3
fibroblast cell line that express transfected N-cadherin and have been previously shown to
promote a robust neurite outgrowth response (Williams et ah, Neuron 1994,13:583-594).
Cells were cultured with soluble MAG-Fc fusion protein present in the culture medium at
concentrations of 0, 5 or 25 μg/ml. As expected, MAG-Fc inhibited neurite outgrowth in a
dose-dependent manner, with an approximately 40% inhibition response when present in the
culture medium at a concentration of 25 μg/ml. Thus, this culture medium, containing
soluble MAG-Fc fusion protein, is an inhibitory culture medium.
The inhibitory culture medium was further supplemented with BDNF at 1 ng/ml,
NGF at 10 ng/ml or 100 ng/ml, BDNF (at 1 ng/ml) in combination with NGF (at 10 ng/ml or
100 ng/ml), a constrained monomer of the loop 1 motif in NGF which binds to the p75
receptor (N-Ac-CTDKGKEC-NH2) at 100 μg/ml, or the NGF loop 1 peptide (at 100 μg/ml)

in combination with BDNF (at 1 ng/ml). The growth media containing MAG-Fc alone was
the control. When the neurotrophin BDNF was present in the inhibitory culture medium at a
concentration of 1 ng/ml there was no measurable effect on the MAG response - i.e.t MAG-
Fc continued to inhibit neurite outgrowth. Like BDNF, NGF at concentrations of either 10
ng/ml or 100 ng/ml did not stimulate neurite outgrowth in the presence of MAG-Fc (Figure
19). In data from individual experiments, the results obtained with NGF at 10 ng/ml and 100
ng/ml were not obviously different, and these data were therefore pooled. These results are
consistent with previous reports that neuron cells must be "primed" with neurotrophins to
circumvent the inhibitor activity of MAG and myelin {see, Cai et al, Neuron 1999,22:89-
101).
The constrained monomer of the NGF loop 1 binding motif alone also had no effect
on neurite outgrowth in the presence of MAG-Fc (Figure 19). However, the NGF loop 1
peptide in combination with BDNF produced a significant neurite outgrowth response.
Additionally, a significant neurite outgrowth response was also observed when BDNF and
NGF were added together (Figure 19) in a ratio of 1:10 or 1:100 (BDNF to NGF).
The results from these experiments suggest that, when NGF and BDNF are
administered to an inhibitory environment, NGF allows BDNF to promote neurite outgrowth.
The results additionally show that administration of a constrained monomer of the first β
hairpin loop in NGF allows BDNF to promote CNS neuron growth in an inhibitory
environment.
7.10. TREATMENT OF A PATIENT WITH SPINAL CORD INJURY
A patient is diagnosed with a thoracic spinal cord injury and has loss of sensation and
motor activity in his legs. The patient undergoes surgery to stabilize the thoracic spine.
Following debridement of soft tissue and bone, the damaged spinal cord is exposed. A sterile
• pharmaceutical powder comprising a p75 receptor binding agent is mixed with sterile normal
saline to form a gel. The surgeon topically applies the p75 receptor binding agent gel to the
exposed surface of the cord. The stabilization procedure is completed in the usual fashion.
Post-operatively, the patient is monitored for improvement in sensation and/or motor activity
in the lower extremities.

8. REFERENCES CITED
Numerous references, including patents, patent applications and various publications,
are cited and discussed in the description of this invention. The citation and/or discussion of
such references is provided merely to clarify the description of the present invention and is
not an admission that any such reference is "prior art" to the invention described here. All
references cited and/or discussed in this specification (including references, e.g., to biological
sequences or structures in the GenBank, PDB or other public databases) are incorporated
herein by reference in their entirety and to the same extent as if each reference was
individually incorporated by reference.

WE CLAIM:
1. A cyclic peptide comprising, within a cyclic ring of the cyclic peptide, the
amino acid sequence:
Arg-Gly-Gln
wherein the cyclic peptide modulates Trk receptor mediated activity.
2. A cyclic peptide according to claim 1, wherein the Trk mediated activity is
selected from the group consisting of: neuronal growth, neuronal survival, axonal growth,
synaptic plasticity, and neurite outgrowth.
3. A cyclic peptide according to claim 1 mat modulates neurite outgrowth.
4. A cyclic peptide according to claim 1 that inhibits Trk mediated activity.
5. A cyclic peptide according to claim 1 that enhances Trk mediated activity.
6. A cyclic peptide according to claim 1, the cyclic peptide comprising the
formula-
wherein:
(a) Y1 and Y2 are independently selected amino acids with a covalent
bond formed between Yl and Y2; and
(b) X1 and X2 are optional and, if present, are independently selected
amino acids or sequences of amino acids joined by peptide bonds.
7. A cyclic peptide according to claim 6 wherein the size of the cyclic peptide
ring ranges from 5 to 15 amino acids.

8. A cyclic peptide according to claim 6, said cyclic peptide having the formula:

9. A cyclic peptide according to claim 8, said cyclic peptide having the amino
acid sequence:
wherein a covalent bond joins the N-teiminal and C-tenninal cysteines in said amino acid
sequence.
10. A cyclic peptide according to claim 6 that inhibits Trk mediated activity.
11. A cyclic peptide according to claim 6, the cyclic peptide comprising the
formula:


wherein:
(a) Y1 and Y2 are independently selected amino adds with a covalent
bond formed between Y1 and Y2; and
(b) Z1, Z2 and Z0 are optional and, if present, are independently selected
amino acids or sequences of amino acids joined by peptide bonds.
12. A cyclic peptide according to claim 11 wherein the size of the cyclic peptide
ring ranges from about 8-50 amino acid residues.
13. A cyclic peptide according to claim 11, said cycEc peptide having the
' formula:
14. Acyclic peptide accordimg to claim 11,said cyclic peptide having the amino
acid sequence:
Cys-Ser-Arg-Arg-Gly-Glu-Leu-Ala-Ala-Ser-Arg-Arg-Gly-Glu-Leu-Cys (SEQ ID
NO:17),
wherein a covalent bond joins the N-terminal and C-terminal cysteines in said amino acid -.
sequence.
15. Acyclicpeptide according to claim 11 that enhances Trk mediated activity.
16. A cyclic peptide according to any one of claims 1, 6,8-9,11 and 13-14,
which cyclic peptide further comprises, on the amino terminal residue, an N-acetyl group, an
N-formyl group or an N-mesyl group.

17. A cyclic peptide accordimg to any one claims 1,6, 8-9,11,13-14 and 16, .
which, cyclic peptide further comprises, on the C-terminal residua, a C-amide group.
18. A cyclic peptide according to any one of claims 6,8,11 and 13, wherein Y1
and Y2 are covalently linked by disulfide bonds.
19. A cyclic peptide according to claim 18, wherein Yl and Y2 axe independently
selected from the group consisting of: penicillamine; β, β-tetramethylene cysteine; β β-
pentamethylene cysteine; β-mercaptopropionic acid; β,β-pentamethylene-β-
raercaptopropionic acid; 2-mercaptobenzene; 2-raercaptoatdline; 2-mercaptoproline; and .
derivatives thereof.
20. A cyclic peptide according to claim 19, wherein Y1 and Y2 are each cysteine
or a derivative thereof
21. A cyclic peptide according to any one of claims 6, 8,11 and 13 wherein Y1
and Y2 are covalently linked by an amide bond.
22. A cyclic peptide according to claim 21 wherein the amide bond is formed
between terminal functional groups.
23. A cyclic peptide according to claim 21 wherein the amide bond is formed
between one terminal functional group and one residue side chain.
24. A cyclic peptide according to claim 22 wherein:

(a) Y1 is selected from the group consisting of lysine, ornamine and
derivatives thereof; and
(b) Y2 is selected from the group consisting of aspartate, glutamine and
derivatives thereof.
25. A cyclic peptide according to claim 22 wherein:

(a) . Y1 is selected from the group consisting of aspartate, glutamine and
derivatives there of and
(b) Y2 is selected from the group consisting of lysine, omathine and
derivatives thereof
25. A cyclic peptide according to any one of claim 6, 8,11 and 13 wherein Yl
and Y2 are covalently linked by a thioether bond.
27. A cyclic peptide according to any one of claims 6,8,11 and 13 wherein:
(a) Y1 and Y2 are each tryptophan or derivatives mereof and
(b) the covalent bond between Y1 and Y2 forms a 8161-ditryptophan or a
derivative mere of.
28. A method for screening a candidate compound for the ability to modulate Trk
Receptor mediated activity, which method comprises comparing a three-dimensional structure
of the candidate compound to a three-dimensional structure of a cyclic peptide that
modulates Trk receptor mediated activity, wherein:
(a) said cyclic peptide comprises, within a cyclic ring thereof, the amino
acid sequence Arg-Gly-Gln, and
(b) similarity between the structure of the candidate compound and the
structure of the cyclic peptide is indicative of the candidate
compound's ability to modulate Trk receptor mediated activity.
29. A method according to claim 28 wherein the cyclic peptide comprises the
formula:
wherein:
(a) Yi and Y2 are independently selected amino acids with a covalent
bond formed between Yl and Y2; and

(b) X1 and X2 arc optional and, if present, arc independently selected
amino acids or sequences of amino acids joined by peptide bonds.
30. A method according to claim 28 wherein the cyclic peptide comprises the
formula:

wherein:
(a) Y1 and Y2 are independently selected amino acids with a covalent
bond formed between Y1 and Y2; and
(b) Z1, Z2 and Zo are optional and, if present, are independently selected
amino acids or sequences of amino acids joined by peptide bonds.
31. A method according to claim 28 wherein the cyclic peptide comprises an
amino acid sequence selected from the group consisting of:

Cys-er-Arg-Arg-Gly-Glu-Leu-Ala-Alk-Ser-Arg-Arg-Gly-Glu-Leu-Cys (SEQ ID NO:17);
wherein a covalent bond joins the N-terminal and C-termmal cysteines in said amino acid
sequence.
32. A method according to claim 28 wherein:
(a) the cyclic peptide enhances Trk receptor mediated activity, and
(b) similarity between the structure of the candidate compound and the
structure of the cyclic peptide is indicative of the candidate
compound's ability to enhance Trk receptor mediated activity.

33. A method according to claim 28 wherein:
(a) the cyclic peptide inhibits Trk receptor mediated activity, and
(b) similarity between the structure of the candidate compound and the
structure of the cyclic peptide is indicative of the candidate
compound's ability to inhibit Trk receptor mediated activity.

34. A peptidornimetic that modulates Trk receptor mediated activity, wherein the
peptidornimetic has a three-dimensional structure that is substantially similar to a three
dimensional structure of a cyclic peptide that modulates Trk mediated activity, said cyclic
peptide comprising, within a cyclic ring there of, the amino acid sequence Arg-Gly-Glu.
35. A peptidornimetic according to claim 34 wherein the cyclic peptide comprises
me formula:
wherein;
(a) Y1 and Y2 are independently selected amino acids with a covalent
bond formed between Yl and Y2; and
(b) X1 and X2 are optional and, if present, are independently selected
amino acids or sequences of amino acids joined by peptide bonds.
36. A peptidornimetic according to claim 34 wherein the cyclic peptide comprises
the formula:
wherein:
(a) Y1 and Y2 are independently selected amino acids with a covalent
bond formed between Y1 and Y2; and
(b) Z1, Z2 and Zo are optional and, if present, are independently selected
amino acids or sequences of amino acids joined by peptide bonds.

37. Apeptidomimetic according to claim 34 wherein the cyclic peptide comprises
aa amino acid sequence selected from the group consisting of
Cys-Arg-Gly-Glu-Cys (SEQ ID NO:9);
Cys-ser-Arg-Arg-Gly-Glu-Cys (SEQ ID NO: 1);
Cys-Ala-Arg-Arg-Gly-Glu-Cys (SEQ ID N03);
Cys-Phe-His-Arg-Gly-Glu-Cys (SEQ ID NO:5);
Cys-Ser-His-Arg-GIy-Gla-Cys (SEQ ID NO:7); and
Cys-Ser-Arg-Arg-Gly-Glu-Leu-Ala-AIa ser-Arg-Arg-gly-Glu-Leu-Cys (SEQ ID NO:17);
wherein a covaleat bond joins the N-termmal and C-terminal cysteines in said amino acid
sequence.
38. • A peptidomimetic according to claim 34 that enhances Trk receptor mediated-
activity.
39. A peptidomietic according to claim 34 that inhibits Trk receptor mediated
activity.
40. A pharmaceutical composition comprising:
(a) an amount of a cyclic peptide according to claim 1 or a
peptidomimetic according to claim 34, which amount is effective for
modulating a Trk receptor mediated activity; and
(b) one or more phatmacentically or physiologically acceptable carriers,
diluents or excipients.
42. A pharmaceutical composition according to claim 40, wherein the cyclic
peptide or peptidomimetic enhances a Trk receptor mediated activity.
42. A pharmaceutical composition according to claim 40, wherein the cyclic
peptide or peptidomimetic inhibits a Trk receptor mediated activity.
43. A cyclic peptide according to claim 8, said cyclic peptide having the amino acid
sequence:
Ser-Arg-Arg-Gly-Glu-Leu-Ser-Arg-Arg-Gly-Glu-Leu (SEQ ID N0 39);
whereia the terminal serine and the terminal leucine are covalently linked by a
peptide bond.

Cyclic peptides and peptidomimetics are provided that bind to and/or modulate activities associated with Trk recep
tors, including processes associated with the growth and repair of the central nervous system (e.g., neuronal growth and survival,
axonal growth, neurite outgrowth and synaptic plasticity). Cyclic peptides and peptidomimetics are also provided that block or reduce
the effect of other factors mat inhibit growth and/or repair of the central nervous system. Pharmaceutical compositions and other
formulations comprising these compounds are provided. In addition, the invention provides methods for using the cyclic peptides
and peptidomimetics to modulate Trk mediated activities, including processes such as neuronal growth, survival and recover, axonal
growth, nenrite outgrowth, and synaptic plasticity. Further, the invention provides methods for promoting central nervous system
(CNS) neuron growth by administering a p75 receptor binding agent.