Background Art Alzheimer's Disease is the most common form of dementia deriving from neurodegeneration of the central nervous system (Selkoe, 2001). Recent studies show that Alzheimer's Disease is a heterogeneous syndrome due to the presence of different phenotypes and genotypes. However, Alzheimer's Disease exhibits a single clinical condition characterised by a progressive decrease in cognitive abilities in elderly persons, whose initial symptoms are detectable in a loss of memory and language (Marin et al., 2002). The neuropathological diagnosis of Alzheimer's Disease is performed by the histopathological examination of the brain. The main characteristics are given by the presence of plaques constituted by extracellular deposits of p-amyloid protein, intracellular tangles constituted by tau protein phosphorylated in anomalous fashion, gliosis and inflammation, as well as neuronal death and decrease in neuronal synapses (Geula, 1998; Lee et al. , 2001; McGeer and McGeer, 2001; Schubert et al. , 2001; Selkoe, 2001).

Although these neuropathological characteristics are present in the worst hit areas during the progress of the disease (the hippocampus, the parietal, temporal and entorhinal cortexes), one of the first signs linked to the onset of the disease is the neurodegeneration of cholinergic neurons of the basal forebrain (Whitehouse et al. , 1982). Evidence that the cholinergic system of the basal forebrain is implicated in Alzheimer's Disease comes from numerous studies which demonstrate, in autoptic findings, a decrease in the activity of the acetylcholinesterase enzyme (ChAT) (Whitehouse et al. , 1982; DeKosky et al., 1992; Lehericy et al. , 1993), of uptake (Rylett et al., 1983) and of the release of acetylcholine (Nilsson et al. , 1986) and a decreased expression of the nicotinic and muscarinic receptors (Nordberg et al. , 1992).

An indirect demonstration of basal forebrain involvement in Alzheimer's Disease derives from pharmacological studies, in which the administration of blockers of the cholinergic activity to rodents and non human primates determines a decrease in cognitive abilities (Fibiger, 1991; Aigner et al. , 1991). In Alzheimer's Disease patients, a direct correlation has been identified between a decreased activity of the ChAT enzyme, tests to verify the patients'mnemonic state and the severity of neurodegeneration (Wilcock et al. , 1982).

Moreover, there is a definite regional correlation between decreased cholinergic innervation and the formation of ß-amyloid plaques (Arendt et al. , 1985; Etienne et al., 1986).

In spite of the numerous data collected during neuropsychological studies and the analysis of autopsy findings, it is crucially important to have available an animal model which allows to study in detail the relationship between the cholinergic system and the rest of the neuropathological alterations that characterise Alzheimer's Disease, to determine whether there are any neurodegeneration-resistant cholinergic cells and which factors are involved in the establishment of this resistance. The first animal models used in the study of Alzheimer's Disease were produced by means of pharmacological treatments aimed at altering the animals'mnemonic processes (Drachmann & Leavitt, 1974). Several drugs were used, aimed at blocking the activity of muscarinic receptors in rodents and non human primates. However, as a consequence of this treatment, all muscarinic receptors were blocked, not just those expressed in basal forebrain cells. The cellular response at the basal forebrain level, therefore, is generically altered and non specific. Another group of animal models includes the use of animals in which a lesion of the basal forebrain is obtained by the mechanical interruption of innervation, or by means of pharmacological agents and toxins which cause cellular death in the basal forebrain (Wenk, 1997).

However, as in the case of the administration of immunotoxins, degenerated cells in the basal forebrain may include both cholinergic and non cholinergic cells. Moreover, these are always models in which one does not induce resistance to cell death but rather causes it, hence they are models which can be used to study basal forebrain degeneration and not resistance to neurodegeneration.

A group of animal models is based on the fact that Nerve Growth Factor is the main factor that contributes to the development, to the differentiation and to the function of the cholinergic neurons of the basal forebrain. NGF influences the activity of the ChAT enzyme, the release of acetylcholine and the morphology of cholinergic neurons (Vantini

et al. , 1989; Li et al. , 1995; Fagan et al. , 1997; Debeir et al. , 1999; Ruberti et al. , 2000).

Moreover, NGF assumes a particular importance in the study of Alzheimer's Disease: in the hippocampus of these patients, there is an increase in the expression of NGF, whilst a decrease in NGF levels is observed in the basal forebrain (Crutcher et al. , 1993; Scott et al. , 1995; Fahnestock et al. , 1996). This results depends on an alteration in the retrograde transport of NGF from the post-synaptic cortical territory to the cellular body of the basal forebrain. Moreover, a decreased expression of the NGF receptors, TrkA and p75 has been reported (Chao et al. , 1986; Kaplan et al. , 1991; Klein et al. , 1991), in the basal forebrain of Alzheimer's Disease patients (Mufson et al. , 1996,1997 ; Salesi et al. , 2000).

Another study shows that chronic deprivation of NGF, obtained with an ectopic expression of highly specific antibodies neutralising NGF, determines in a transgenic murine model the appearance of a neurodegeneration characterised not only by cognitive deficits and by decreased ChAT expression in the basal forebrain, but also by ß-amyloid plaques in the hippocampus and the presence of tau protein tangles (Capsoni et al. , 2000; Capsoni et al. , 2002a, b). This animal model is particularly well suited for the study of alterations of intracellular signalling deriving from the reduced number of cholinergic neurons in the basal forebrain and from the reduced cholinergic innervation of the hippocampus and of the cortex. hi anti-NGF mice the expression of the antibody is absent in the immediate postnatal period, thereby allowing a completely normal development of the animal. The expression of the antibodies starts 45 days after birth and remains constant over time. Therefore, NGF is neutralised only in the adult.

NGF exercises its biological function through the binding to a specific receptor, TrkA, belonging to the super-family of tyrosine kinase receptors. NGF also binds, though with low affinity, the p75NTR receptor, member of the TNF (Tumour Necrosis Factor) receptor family. The block of the NGF activity in the SNC induces cell death and neurodegeneration.

Transgenic mice which express a neutralising NGF antibody are characterised by the insurgence of a neurodegeneration similar to the one observed in Alzheimer's Disease (Capsoni et al. , 2000).

NGF is generally expressed as a pre-pro-protein of 27 kDa (pro-NGF). The pre-peptide (signal peptide) is removed during the transfer into the endoplasmatic reticule, whilst the pro-peptide is removed by enzymes of the furin type in the trans-Golgi network, thereby generating mature NGF. Mature NGF in its active form is a homodimer of 26 kDa.

Until recently, no biological role had been attributed to pro-peptide, with the exception of the demonstration that it is able to regulate the secretion of neurotrophins (Suter et al., 1991) and the current hypothesis is that it has a chaperone role, to assist with NGF folding.

Recently, some researchers (Rattenholl et al., 2001 a and b) have shown that pro-peptide facilitates folding of human NGF expressed in E. Coli.

More recently, numerous publications have shed new light on the function of proNGF in vivo. It has been discovered that proNGF is the predominant form of NGF in the brain and that it is present in greater quantity in patients affected by Alzheimer's Disease (AD) (Fahnestock et al., 2001). Lastly, some data have shown that the proNGF, through the activation of the high NGF affinity receptor p75NTR, induces apoptosis in neurons in culture (Lee et al., 2001).

The most recent results have also demonstrated (Nykjaer et al., 2004) that there is a specific receptor for proNGF, sortilin, which belongs to the family of VpslOp-domain receptors, recently discovered. The interaction of this receptor with p75NTR enhances its pro-apoptotic action. Therefore the system of NGF (in its processing forms) and of its receptors can activate both trophic ways, of neuronal survival, and pro-apoptotic ways, of death. The global functionality of the NGF and its receptors depends in delicate fashion by its fine regulation of the relationships between the levels of its various forms of processing of the NGF and of its receptors.

Therefore an imbalance in the relationship between the proNGF, which induces apoptosis and cell death, and the NGF, which determines cell survival, could be determining in the insurgence of some neurodegenerative diseases.

Description of Invention In light of this hypothesis and of the fact that the anti-NGF antibody oDll induces neurodegeneration, the authors have studied the affinity of the binding of this antibody for pro-NGF and NGF. Moreover, they have worked on the research of antagonist molecules and of antibodies which would allow to block the function of pro-NGF to prevent its toxicity, but not that of the mature NGF. For this purpose, inhibitors of the proNGF-p75 receptor and associated co-receptor, such as sortilin, are found advantageous, both whether they act on the pro-NGF and on its receptor and specific co-receptors, such as p75 and sortilin.

The authors of the present invention have already shown that the monoclonal antibody aDl1 (Cattaneo et al. 1988) is able selectively to recognise the mature form of NGF, with respect to proNGF. aDll binds NGF at an epitope responsible for the bind with its high affinity receptor, TrkA, acting as a neutralising antibody. The epitope of NGF recognised by aDl l comprises aa. 41-49.

This antibody was used for the production of transgenic mice (W0001/10203). AD11, anti-NGF mice express the recombining form of oDll, have a phenotype similar to Alzheimer's Disease and hence represent an ideal animal model to study the effects of the proNGF block on the Alzheimer phenotype. Consequently, the ad 11 antibody can be used to amplify in unexpected and unforeseen fashion the negative and neurodegenerative effects of proNGF, thereby allowing to definition of assays for isolating inhibitors of the activity of proNGF, of therapeutic interest.

The authors have also selected intracellular antibodies against the precursor protein of Nerve Growth Factor (NGF), proNGF.

An object of the invention is to provide a pharmaceutical composition comprising at least one molecule able to prevent and/or to inhibit the binding and/or the interaction between pro-NGF or NGF and receptor p75 or its co-receptors, preferably sortilin, for the therapy of neurodegenerative diseases and solvents and/or adjuvants a-nd/or excipients in appropriate form and dosage.

In a preferred embodiment, the composition is for the preventive therapy of neurodegenerative diseases.

Preferably, the molecule of the composition is included in the following group: monoclonal or recombining antibodies or synthetic fragments thereof or molecules of natural derivation or synthetic, able to prevent and/or to inhibit the binding and/or the interaction between pro-NGF or NGF and receptor p75 or its co- : receptors, preferably sortilin, without having undesired side effects.

Alternatively, the molecule of the composition is included in the following group: IAP activator molecules, CREB activators and ARMS, TRAF-2 and/or cj un inhibitors.

In a further preferred form, the monoclonal or recombining antibody or synthetic fragment thereof is an anti-NGF or anti pro-NGF blocking p75. Alternatively, the monoclonal or recombining antibody or synthetic fragment thereof is an anti-p75.

Alternatively, the monoclonal or recombining antibody or synthetic fragment thereof is an

anti-sortilin. Alternatively, the naturally derived or synthetic molecule belongs to the family of immunoadhesins for p75.

A further object of the invention is to provide a method for assaying a molecule able to prevent and/or to inhibit the binding and/or the interaction between pro-NGF or NGF and receptor p75 or its co-receptors, preferably sortilin, comprising the steps of : a) treating a non human transgenic animal, expressing anti-mature NGF antibodies and having a phenotype mimicking a neurodegenerative disease, with said molecule with appropriate doses, modes, vital phase and times, and b) measuring the effect of the treatment on cholinergic cells.

Preferably, the neurodegenerative disease is Alzheimer's Disease.

Preferably, the doses are between 0.5 and 15. 0, ug/kg of body weight.

Preferably, exposure takes place for at least 15 days.

Preferably, the measuring on cholinergic cells takes place by means of reaction of said cells with specific reactants for cholin acetyltransferase and revelation of positive cells.

Preferably, the non human transgenic animal belongs to the Mus musculus species. More preferably, the non human transgenic animal belonging to the Mus musculus species is transgenic for the oD 1 anti-NGF antibody.

A further object of the invention is to provide a molecule selectable with the described method. In a preferred embodiment, the molecule is an anti pro-NFG antibody.

A further object of the invention is to provide a method for identifying a selective inhibitor of pro-NGF comprising the steps of : a) treating a non human transgenic animal, expressing anti-mature NGF antibodies and having a phenotype mimicking a neurodegenerative disease, with said inhibitor with appropriate doses, modes, vital phase and times, and b) verifying that the inhibitor alleviates at least one of the characteristics of the phenotype mimicking the neurodegenerative disease.

Preferably, the neurodegenerative disease is Alzheimer's Disease.

Preferably, the non human transgenic animal belongs to the Mus musculus species. More preferably, the non human transgenic animal belonging to the Mus musculus species is transgenic for the aDl l anti-NGF antibody.

Another object of the invention is to provide a selective inhibitor of proNGF selectable according to the described method. In a preferred embodiment, the inhibitor is an anti pro-NFG antibody.

Detailed description of the invention The invention derives from the result obtained after administering anti-NGF antibodies to anti-NGT transgenic mice (patent application WO 01/10203) at a precocious age. Anti- NGF transgenic mice are characterised by a complex pathological condition (Ruberti et al. , 2000; Capsoni et al. , 2000,2002a, 2002b) that is wholly similar to the one exhibited at macroscopic, histological and molecular in Alzheimer's Disease patients. In particular, anti-NGF mice are the sole animal model for Alzheimer's Disease which shows lesions at the level of the cholinergic neurons of the basal forebrain and of the Meynert basal nucleus (Ruberti et al. , 2000; Capsoni et al. , 2000,2002a). The anti-NGF (aDll) antibody produced by hybridomas binds the NGF molecule at the epitope responsible for its interaction with the high affinity TrkA receptor and blocks interaction with the p75 receptor. Consequently, blocking this binding, it is an antibody characterised by neutralising activity on both the TrkA and p75 receptors.

It was expected that the administration of said molecules would anticipate and/or enhance the degree of neurodegeneration in anti-NGF transgenic mice. Surprisingly, the administration of anti-NGF antibodies which block the binding of NGF to TrkA and to p75, has determined the prevention of neurodegeneration at the level of the cholinergic system, observed in anti-NGF mice from the age of 2 months onwards. To understand whether this effect was due to the inhibition of the interaction between NGF and TrkA, antibodies were used which block the binding of NGF to TrkA (Cattaneo et al. , 1999; patent EP 1.181. 318) and it was observed that these antibodies do not protect from neurodegeneration.

Further investigations were then conducted to determine whether the preventive effect on degeneration depended on the blocking of the NGF action through p75, or rather on the blocking of the interaction between NGF and TrkA. The hypothesis was verified by means of two types of experiments: 1) administration of anti-NGF antibodies which inhibit only interaction with the p75 receptor; 2) administration of anti-p75 antibodies which inhibit the binding of NGF to p75 and to receptors in soluble form (immunoadhesins) which, binding to NFG, prevent its binding to the endogenous p75 receptor.

Consistently with the hypothesis of the authors, both the antibodies against p75 and the immunoadhesins are able to prevent neurodegeneration.

It is concluded that the preventive blocking of the NGF/p75 interaction is able to inhibit the action mechanism which leads to neurodegeneration in anti-NGF mice.

The authors have also researched"markers"which are correlated to the neurodegeneration prevention mechanisms, and for this purpose they have executed a screening of known proteins to be i) involved in the transduction of the NGF signal, ii) involved in the neuronal survival process modulated by NGF and iii) expressed in basal forebrain and cortical/hippocampal neurons.

Of particular interest is the study of a family of phylogenetically preserved proteins.

Apoptosis inhibitor proteins (IAP) are natural inhibitors of caspasis, the proteolytic enzymes effecting apoptosis which, when activated, determine cell death, influencing the cell cycle and the expression of different molecules (Salvesen & Duckett, 2002; LeBlanc, 2003). IAP are a family of proteins whereto belong X-IAP, c-IAP-1 e C-IAP-2 (Salvesen & Duckett, 2002; LeBlanc, 2003). IAP proteins prevent the activation of caspasis by means of a direct binding with these enzymes or by means of induction of their degradation by the proteosoma (Salvesen & Duckett, 2002; LeBlanc, 2003).

The expression of LAP is regulated by different factors, such as"tumour necrosis factor a" (Chu et al. , 1997), the growth factor of vascular endothelium (Tran et al. , 1999), and NGF (Wiese et al. 1999). In the sensorial and sympathetic ganglia of chickens, NGF determines an increase in the expression of ITA, the homologue of c-IAP-2 in chickens (Wiese et al. , 1999). Since, in chickens, both NGF and ITA determine the survival of sympathetic and sensorial neurons, it is presumed that NGF is able to promote the survival of these neuronal populations by the induction of high levels of ITA. It is therefore supposed that also in mammal cells, IAP is one of the mediators of the survivals of those cells which depend on NGF (Wiese et al. , 1999). In particular, c-IAP-1 and c- IAP-2 are expressed in the hippocampus cells of the areas CA1 and CA3 of the hippocampus (Korhonen et al. , 2001) which are innervated by the cholinergic fibres originating from the basal forebrain and which also express the receptor of NGF p75 (Mrzljak & Goldman-Rakic, 1993).

In the basal forebrain, in the cerebral cortex and in the hippocampus of anti-NGF mice, a decrease is observed in the expression of c-IAP-1 and c-IAP-2 from the age of 2 months onwards. The administration of anti-NFG antibodies (blocking the binding of NGF to Trka and p75 or solely the binding to p75) and anti-p75 antibodies determines the restoration of the number of cells which express c-IAP-1 and c-IAP-2 in the basal

forebrain, in the cerebral cortex and in the hippocampus. The administration of anti-TrkA antibodies does not return the number of the cells which express c-IAP-1 and c-IAP-2 to normal values.

IAP proteins therefore represent i) an example of proteins involved in the neurodegeneration mechanism which can be modulated by the administration of anti- NGF antibodies and anti-p75 ii) target candidates for pharmacological intervention. IAP activators could be used to prevent neurodegeneration, as well as CREB activators and ARMS, TRAF-2 and/or c-jun inhibitors (see Table 1).

The present invention shall now be described in non limiting examples.

Description of the figures Figure 1. Schematic representation of the interaction of anti-NGF oDll antibodies (Cattaneo et al. , 1988), anti-NGF 27/21 antibodies (Nanduri et al. , 1994), anti-TrkA MNAC13 antibodies (Cattaneo et al. , 1999), anti-p75 LLG 17 antibodies; and immunoadhesin LLG 65 with the receptors of NGF TrkA and p75.

Figure 2. Experimental scheme. The antibodies were administered to 8-day old anti-NGF mice, in the period when there is no expression of the transgenic antibody. Histological analysis was performed at the age of 2 months.

Figure 3. Coloration for cholin acetyltransferase (ChAT) of the basal forebrain of control mice (A), anti-NGF transgenic mice (B), anti-NGF transgenic mice after implant with hybridoma P3U (C), anti-NGF transgenic mice after implant with hybridoma secreting aD 11 (D).

Figure 4. Quantification of the cholinergic deficit in anti-NGF mice treated with hybridomas secreting aDll, MNAC13, 27/21, LLG 17 and LLG 65. The chart shows the prevention of the reduction in the number of cholinergic neurons of the basal forebrain in anti-NGF mice treated with the oDll, 27/21, LLG 17 antibodies, and with the immunoadhesin LLG 65. The anti-TrkA antibody MNAC13 is not able to prevent cholinergic deficit in the basal forebrain.

Figure 5. Marking of the hippocampus with anti c-IAP-1 antibodies. The anti c-IAP-1 antibody, directed against a recombining protein corresponding to the amino acids 111- 193 of the inner region of c-IAP-1, (Santa Cruz, Santa Cruz, CA) marks the neurons in the control mouse (A) and in anti-NGF mice treated with aDl l hybridomas (D), but not in anti-NGF mice (B) and in anti-NGF mice treated with the P3U hybridoma (C).

Figure 6. Marking of the hippocampus with anti c-IAP-2 antibodies. The anti c-IAP-2, directed against a recombining protein corresponding to the amino acids 94-178 of the inner region of c-IAP-2, (Santa Cruz, Santa Cruz, CA) marks the neurons in the control mouse (A) and in anti-NGF mice treated with aDl1 hybridomas (D), but not in anti-NGF mice (B) and in anti-NGF mice treated with the P3U hybridoma (C).

Figure 7. Quantification of the number of positive c-IAP-1 neurons in the basal forebrain after treatment with hybridomas secreting aDll, MNAC13, 27/21, LLG 17 and LLG 65.

The aDll, 27/21, LLG 17 antibodies and LLG 65 immunoadhesin determine the prevention of the reduced expression of c-IAP-1 in the basal forebrain of the anti-NGF mice. The administration of MNAC13 antibodies is not able to prevent the reduction in positive c-IAP-1 cells in the basal forebrain.

Figure 8. Quantification of number of positive c-IAP-2 neurons in the basa1 forebrain after treatment with hybridomas secreting aDl l, MNAC13, 27/21, LLG 17 and LLG 65.

The oDll, 27/21, LLG 17 antibodies and LLG 65 immunoadhesin determine the prevention of the reduced expression of c-IAP-2 in the basal forebrain of the anti-NGF mice. The administration of anti-TrkA (MNAC13) antibodies is not able to prevent the reduction in positive c-IAP-2 cells in the basal forebrain.

Figure 9. Quantification of the number of positive c-IAP-1 neurons in the entorhinal cortex after treatment with hybridomas secreting anti-NGF aDll and 27/21, MNAC13, LLG 17 and of LLG 65 immunoadhesin. The aDl l, 27/21, LLG 17 antibodies and LLG 65 immunoadhesin determine the prevention of the reduced expression of c-IAP-1 in the entorhinal cortex of the anti-NGF mice. The administration of MNAC13 antibodies is not able to prevent the reduction in positive c-IAP-1 cells in the entorhinal cortex.

Figure 10. Quantification of the number of positive c-IAP-2 neurons in the entorhinal cortex after treatment with hybridomas secreting aDll, 27/21, MNAC13, LZ, G 93 and LLG 65. The aD 11 and 27/21 which block the interaction of NGF with TrkA and p75 or exclusively with p75, and LLG17 and the LLG 65 immunoadhesin, determine the prevention of the reduced expression of c-IAP-2 in the entorhinal cortex of the anti-NGF mice. The administration of MNAC13 anti-TrkA antibodies is not able to prevent the reduction in positive c-IAP-2 cells in the entorhinal cortex.

Figure 11. Diagram of pre-proNGF.

Figure 12. a) Series of successive injections of h-NGF on the channel with ADIl at low concentration. Increasing concentration of antigen, from 100 to 500 nm. The white was

removed from the curves (fc2-fcl). B) Series of successive injections of h-NGF on the channel with AD 11 at high concentration. Increasing concentration of antigen, from 100 to 500 nm. The white was removed from the curves (fc4-fcl). c) Series of successive injections of h-NGF on the channel with asv5. Increasing concentration of antigen, from 100 to 500 nm. The white was removed from the curves (Fc3-Fcl).

Figure 13. a) Series of successive injections of m-NGF on the channel with old l at low concentration. Increasing concentration of antigen, from 100 to 500 nm. The white was removed from the curves (fc2-fcl). B) Series of successive injections of m-ngf on the channel with aD 11 at high concentration. Increasing concentration of antigen, from 100 to 500 nm. The white was removed from the curves (fc4-fcl). C) Series of successive injections of m-ngf on the channel with asv5. Increasing concentration of antigen, from 100 to 500 lim. The white was removed from the curves (fc3-fcl).

Figure 14. A) Series of successive injections of h-prongf on the channel with AD 11 at low concentration. Increasing concentration of antigen, from 100 to 500 nm. The white was removed from the curves (fc2-fcl). B) Series of successive injections of h-proNGF on the channel with AD11 at high concentration. Increasing concentration of antigen, from 100 to 500 nm. The white was removed from the curves (fc4-fcl). C) Series of successive injections of h-prongf on the channel with asv5. Increasing concentration of antigen, from 100 to 500 nm. The white was removed from the curves (fc3-fcl).

Figure 15. a) Series of successive injections of m-proNGF on the channel with aDIl at low concentration. Increasing concentration of antigen, from 100 to 500 nm. The white was removed from the curves (fc2-fcl). B) Series of successive injections of m-proNGF on the channel with aDl l at low concentration. Increasing concentration of antigen, from 100 to 500 nm. The white was removed from the curves (fc4-fcl). C) Series of successive injections of m-proNGF on the channel with aSV5. Increasing concentration of antigen, from 100 to 500 nM. The white was removed from the curves (Fc3-Fcl).

Figure 16. a) Series of successive injections of GSTpro on the channel with aDl1 at low concentration. Increasing concentration of antigen, from 100 to 500 nM. The white was removed from the curves (Fc2-Fcl). b) Series of successive injections of GSTpro on the channel with aD 11 at high concentration. Increasing concentration of antigen, from 100 to 500 nM. The white was removed from the curves (Fc4-Fcl). c) Series of successive injections of GSTpro on the channel with aSV5. Increasing concentration of antigen, from 100 to 500 nM. The white was removed from the curves (Fc3-Fcl).

Figure 17. Diagram of the deletion mutants of the pro-NGF protein.

Figure 18. Xgal analysis for verifying the transactivation of the proNGF bait and of its deletion mutants.

Figure 19. Specificity of the scFv-anti-proNGF-10 fragment with respect to the different deletion mutants in vivo. The epitope-scFv binding is highlighted by means of IACT technology measuring the transactivation of the HIS3 and lacZ genes and therefore the cells'ability to grow in a medium without histidine and to synthesise ß-galactoxidase.

Figure 20. SDS-PAGE of the anti-proNGF coloured with Coomassie blue.

Figure 21. Purification of ScFv on steric exclusion column, Superdex 75 (Amersham).

Run buffer: PBS ; Flow: 0.7 mL/min.

Figure 22. a) and b) Charts showing the data of the experiment 1 (see Tab. 5). Both preparations of anti-proNGF ScFv (anti-pro20 and anti-pro21) specifically recognise proNGF when applied in incubation: the data obtained for NFG, GSTpro, proGST and GST are negative, as is the data obtained with the control ScFv, which does not recognise any protein in specific fashion.

Figure 23. ELISA sandwich to test the specificity of the anti-proNGF antibody. a) and b) are two charts which show the data of the experiment 2 (see Table 5). It is readily apparent that even in the case in which the anti-pro ScFv adheres to the plastic of the plate, and the antigens are applied in incubation, there is an anti-pro recognition effect for the proNGF. If to the same ScFv are applied antigens other than the proNGF, the signal is wholly absent.

Figure 24. Immunohistochemistry for ß-amyloid in (A) non transgenic mice, (B) p75NTR'" (-/-) mice, (C) AD11 VK mice and (D) AD12 mice. In (E) quantification of the number of deposits for ß-amyloid in the four experimental groups.

Example 1. Prevention of the cholinergic deficit in anti-NGF mice by implanting hybridomas secreting antibodies neutralising NGF, TrkA and p75 Immunoadhesin LLG65 was prepared by fusion of the extracellular portion of the human p75 receptor with the Fc portion of camel IgG2 comprising the 35 amino acids of the hinge region of the antibody followed by the CH2 and CH3 domains according to the protocol published by Hamers-Casterman (1993). The resulting coding sequence for immunoadhesin p75 was inserted in the genome of Baculovirus (Autographa califorhica virus) using the pAcGP67B vector for expression in insect cells. Sf9 insect cells were

used to amplify the viruses. H5 cells were infected with amplified recombining Baculovirus, and the immunoadhesins were purified from the culture medium by high affinity chromatography on a Protein A-Sepharose column. Sf9 cells were grown in the TNM-FH medium with 10% bovine foetal serum, whilst H5 cells were grown in the Ex- Cell 400 medium. Both Sf6 and H5 cells were grown in an incubator at 27°C.

To obtain the anti-p75 neutralising antibody LLG 17, the extracellular portion of p75 was purified and injected subcutaneously into Balc/c mice. The mice were injected once with 300, ug of p75 dissolved in complete Freund adjuvant and 10 times with 150, ug of p75 dissolved in incomplete Freund adjuvant, with monthly frequency.

To obtain the hybridomas secreting the LLG 17 antibody, the mice cells (5.6 x 108) were fused with 7 x 107 cells of murine myeloma P3-X 63/AG8a (Galfre and Milstein, 1981) 8 days after a recall injection and cultivated in 24-well plates (Sterilin) at such a concentration that the growth of the hybridomas would be visible in all wells (5-15 colonies/well).

Cells were grown in Dulbecco's modified Eagle's medium (DMEM) (GIBCO) containing 10% of bovine foetal serum, hypoxanthin, aminopterin and thymidin.

The aD 11 hybridomas (secreting anti-NGF monoclonal antibodies which block the interaction between NGF and its receptors TrkA and p75, Fig. 1), 27/21 (secreting antibodies which inhibit the binding of NGF to p75, Fig. 1), MNAC13 (secreting anti- TrkA antibodies; Fig. 1), LLG17 (secreting p75 neutralising antibodies, Fig. 1) and LLG 65 (secreting immunoadhesin which prevents the binding of NGF to p75; Fig. 1) and myeloma P3X63Ag8 (P3U, cellular control line) were grown in the Dulbecco's modified Eagle's culture medium supplemented with 10% of bovine foetal serum. Before injection, cells were washed four times in the Hank saline solution (HBSS) and re-suspended in HBSS at the concentration of 2 x 105 cellslal.

The P3U cellular control line and the cells secreting the different antibodies and the immunoadhesins were injected into the right lateral ventricle of anti-NGF transgenic mice. All injections were performed 8 days after birth (Fig. 2). For this purpose, the mice were anaesthetised in icy water. The cells were injected using a 17 gauge diameter needle connected, through a polyethylene cannula, to a 25 lil Hamilton syringe. The injection point was identified by the following reference points: 1.5 mm lateral to median suture and 0.5 mm at Bregma. One microlitre of the cell suspension was injected into each mouse. After the injection, the mice were treated with cyclosporin A (Roche, Basel,

Switzerland, 15 mg/kg) every second day, to prevent rejection of the implant. The injection of hybridomas allows a constant supply for about 15-20 days in the basal forebrain (Molnar et al. , 1998). After this period, the levels of the antibody decrease until they become insufficient for any neutralising action.

The mice injected with the hybridomas were allowed to grow until reaching 2 months and 6 months of age. They were then sacrificed by injection of Chloralium hydrate (8 ttL of a 10.5% solution per animal) and the tissues fixed by means of transcardiac perfusion of a cold 4% paraformadehyde solution in phosphate buffer.

The brain was removed and post-fixed in the same fixative. After cryoprotection in 30% saccharose, the brains were sectioned using a refrigerated slide microtome.

The sections containing the basal forebrain were processed to highlight the number of cholinergic cells. For this purpose, the sections were washed twice in Tris/HCl 0.1 M buffer to which were added 0.15 M of NaCl and 0.3% of Triton-X 100 (TBST). After blocking the endogenous peroxidase in a 3% solution of hydrogen peroxide, non specific bindings were blocked in 10% of bovine foetal serum dissolved in (TBST). After half an hour, the sections were placed in incubation with the primary antibody directed against choline acetyltransferase (antibody made in anti-ChAT goat; 1: 500, Chemicon, Temecula). Incubation took place at 4°C for 12 hours.

After washing the sections in Tris/HCl 0.1 M buffer to which were added 0.15 M of NaCl (TBS), the sections were incubated for 3 hours in the biotylinated secondary antibody directed against goat IgG (1: 200, Vector Labtek).

The signal deriving from the reaction between primary and secondary antibody was amplified by means of incubation with a peroxidase (Vector Labtek). The reaction was then developed by incubation with the diaminobenzidine chromogen (Sigma, St. Louis, MO).

After collecting the sections on gelled slides, the number of ChAT-positive cholinergic cells was quantified with neurostereology techniques. The basal forebrain is delimited dorsally by the corpus callosum. The anatomical boundaries that define the basal forebrain are: dorsally, the corpus callosum; centrally, the ventral surface of the brain; laterally, the medial line that passes through the rostral branch of the front commissure.

The rostral boundary is determined by the plane that passes through the knee of the corpus callosum, whilst the caudal boundary is constituted by the plane that passes

through the first section that contains the front commissure. All anatomical references were taken from the Franklin and Paxinos (1997) mouse Atlas.

The volume of the basal forebrain was calculated using Cavalieri's method (Michel and Cruz-Orive, 1988; Peterson et al. , 1999). A grid was overlaid on the monitor of the computer on which the images of the basal forebrain were projected using a 2. SX lens. The volume of each square of the grid was calculated multiplying the thickness of the sections (T) times the area of each square that formed the grid (a). The total volume (Vref) was obtained counting the squares superposed on the basal forebrain in each section (Pi) and multiplying the sum thereby obtained times the volume of each square of the grid, according to the formula Vref= (SPi) (a) (T).

The estimate of the total number of the cholinergic neurons of the basal forebrain was obtained with the optical fractionating method (West, 1993; Peterson et al., 1999). The images, acquired at 60X enlargement with a Zeiss transmitted light microscope connected to a CCD television camera, were projected on the monitor of a computer. Numerical density (Nv) was determined using a grid superposed on the monitor screen. The number of neurons for each grid was counted by focusing in tissue at a known distance (20 Itm), as described by Sterio (1984). The number of cells calculated in this volume provides the average number of cells/volume (Nv). The total number of neurons was calculated multiplying Nv times the total volume of the basal forebrain (calculated according to Cavalieri's procedure). The volume and the total number of neurons was calculated for n=4 for each experimental group. Statistical analysis was conducted according to the T- tail test method.

The quantification of the total number of cholinergic neurons in the basal foreb : rain has demonstrated that the injection of hybridomas secreting the antibodies neutralising aDIl (Fig. 3,4), 27/21, LLG 17 and LLG 65 immunoadhesin at a precocious age determines a prevention of the cholinergic deficit that will take place at a late age (Fig 4). Prevention does not take place after administration of the anti-TrkA antibody MNAC13 (Fig. 4).

Example 2 Screening of molecules to verify the altered phenotype and the effect of the treatment with hybridomas secreting anti-NGF, anti-TrkA and anti-p75 antibodies To determine whether treatment with hybridomas prevents alterations in expression of some molecules involved in the translation of the NGF signal, the sections of the brain of the anti-NFG mice obtained according to Example 1 were incubated with the following

primary antibodies according to the protocol for immunohistochemistry described in example 1 : anti-phospho CREB (serl33) (Cell Signaling, Beverly, MA), anti phospho-c- jun (ser63) (Cell Signaling, Beverly, MA), anti-fas (epitope corresponding to amino acids 1-335), anti-p53 (epitope corresponding to amino acids 1-393), anti-NfKB (epitope corresponding to the carboxy terminal of NfkB), anti-TRAF2 (epitope to the terminal carboxy of TRAF2; Santa Cruz, Santa Cruz, CA), anti-ARMS (epitope to the terminal carboxy; provided by Dr. Moses Chao, New York Medical Center, New York). The results are summarised in Table 1.

Table 1 MARKER Expression in AD 11 mice with Restoration of normal values after the respect to non transgenic mice administration of anti-NGF and anti- p75 antibodies Fas increased no NFKB increased no p53 increased no CREB decreased yes C jun increased yes TRAF-2 increased yes ARMS increased yes c-IAP-1 decreased yes C-IAP-2 decreased yes Example 3. Prevention of the decreased expression of c-IAP-1 and c-IAP-2 in anti-NGF mice treated with hybridomas secreting anti-NGF antibodies.

NGF deprivation determines, in transgenic mice AD 11, a decreased expression of c-IAP- 1 and c-IAP-2 in the basal forebrain and in the hippocampus, determined by means of immunohistochemistry (Fig. 5 and Fig. 6). To determine whether the induction of resistance in the cholinergic neurons of the basal forebrain determines and increase in the expression of c-IAP-l and of c-IAP-2, the sections of the brain of the mice treated with the hybridomas were processed with immunohistochemistry.

The sections containing the basal forebrain and the hippocampus were processed to highlight the number of positive cells for c-IAP-1 and c-IAP-2. For this purpose, the

sections were washed twice in TBST buffer. After blocking the endogenous peroxidase in a 3% solution of hydrogen peroxide, non specific bindings were blocked in 10% of bovine foetal serum dissolved in TBST. After half an hour, the sections were incubated with the primary antibody directed against c-IAP-2 and c-IAP-2 (antibodies in anti-ChAT rabbit; 1: 100, Santa Cruz, Santa Cruz, CA). Incubation took place at 4°C for 12 hours.

After washing the sections in TBS buffer, the sections were incubated for 3 hours in the biotylinated secondary antibody directed against rabbit IgG (1: 100, Vector Labtek).

The signal deriving from the reaction between primary and secondary antibody was amplified by means of incubation with avidin conjugated with peroxidase (Vector Labtek). The reaction was then developed by incubation with the diaminobenzidine chromogen (Sigma, St. Louis, MO).

Observation of the sections highlights the fact that precocious administration ofoDll, 27/21, LLG 17 antibodies and of immunoadhesin LLG 65 determines not only resistance to its deprivation in adult age in the cholinergic cells of the basal forebrain (Fig. 7,8), but also a resistance to the decreased expression of c-IAP-1 and c-IAP-2 in the cerebral target areas of cholinergic innervation (Fig. 9,10).

Example 4 Analysis of the antigen-antibody interaction by Surface Plasmon Resonance (SPR) by means of BIACORE The monoclonal antibodies (Mab) aDll and aSV5 were immobilised on a CM5 chip (BIACORE) by means of primary amines according to the protocol of the manufacturer company. The four channels of the chip were split in the following way: - cell 1 (Fcl): white, dextrane surface activated and deactivated - cell 2 (Fc2): Mab aDl 11 at low concentration-3 00 RU - cell 3 (Fc2): Mab aSV5-500 RU - cell 4 (Fc2): Mab aDl 11 at high concentration-1000 RU PBS was used as run buffer, with flow of 30 yL/min.

After each cycle, the chip was regenerated with a brief injection of 10 mM Glycine buffer, pH 1.5. Data analysis was performed using the BIAevaluation 3.1 program.

Proteins used: -mo710clo7lal antibodies aD11 and aSVr : both monoclonal antibodies were obtained and purified according to standard procedures. The aDl l and anti-SV5 immunoglobulins were expressed in the surnatant by means of culture of hybridoma cells

and concentrated by precipitation with 29% ammonium sulphate followed by dialysis at 4°C in PBS using Spectra-Por 12/14K (Spectrum) membranes at 4°C. Both immunoglobulins were purified by means of affinity chromatography using a column of Protein G Sepharose (Pharmacia) and eluted with lOmM HCI ; following dialysis in PBS using Spectra-Por 12/14K (Spectrum) membranes at 4°C, each specimen was concentrated by means of Centricon 50KDa (Amicon) ultrafiltration units. The concentration of the purified protein was determined by Lowry assay (Bio-Rad).

-m-NGF : commercial (company: Alomone) - la-NGF - m proNGF : amplification of the mouse proNGF (m-proNGF) for RT-PCR from lysate of mouse submandibular glands. Cloning in the vector pETM-13 (EMBL protein expression and purification unit) for expression in E. Coli. Expression, refolding and purification according to the protocol described in Rattenholl et al., 2001 b.

- lz poNF - GSTpro : fusion protein of the NGF precursor with GST at the N-terminal. It was obtained by cloning only the precursor of the murine NGF in the vector pGEX-4T-2 for expression in E. Coli. Expressed as a soluble protein (grown in LB medium with Ampicilline antibiotic until an OD600=0. 5; induction of the expression with IPTG 0.5 mM ; induction for 5h at 30°C). Purified by means of affinity chromatography for GST (GSTrap, Amersham Bioscience) and eluted in PBS+ 10 mM glutation, dialysed in PBS with Spectra-Por 8/1 OK (Spectrum) membranes and concentrated with Centricon 10KDa (Amicon) ultrafiltration unit. The concentration of the purified protein was determined by Lowry assay (Bio-Rad).

Example 5 proNGF mutants and fusion constructs with lexA The authors of the present invention have recently devised a method which allows in vivo, in the natural environment, antibodies against antigens of different nature using antibodies expressed and active within the cell (intracellular antibodies).

The use of intracellular antibodies, i. e. of antibodies or portions thereof, expressed and active within a cell, constitutes a valid technology to interfere with the function of a protein, in its physiological intracellular context. In this way, the technology allows to obtain an inhibition of the functionality of the target protein (protein knock-out).

Therefore, a phenotype of interest can be conferred to a cell or to an organism by means

of an appropriate intracellular antibodies. The technology of the intracellular antibodies is based on two advantageous aspects: 1) the virtually unlimited wealth of the repertory of the immune system (real or artificial), which provides a source of molecules able to react at high affinity and specificity against any protein, and 2) the possibility of addressing a protein (and hence an antibody) into different intracellular compartment, by means of appropriate intracellular localisation, autonomous and dominant. Recently, a development of the technology of intracellular antibodies has been introduced, which allows its use as a methodology of choice for functional genomic programs: the so-called IACT (Intracellular Antibody Capture Technology) (Visintin, M. et al. , 2002). This methodology allows to isolate or select, among all the antibodies directed against a given protein, the sub-population that is able to function efficiently in vivo, in the intracellular environment. The technology is described in the international patent application PCT W003014960, incorporated in full herein. The authors of the present invention have further and advantageously developed this technology for the creation of an antibody library expressed directly in the double hybrid format. The entire procedure to obtain intracellular antibodies starting from genes, as inputs, is facilitated, abbreviated, accelerated and made yet more parallel, because the genes of interest are directly inserted in the vector for the IACT selection, short-circuiting the need to express and purify the corresponding proteins. The technology, called SPLINT (Visintin, M. , et al. , 2004) is described in the patent application RM2002A000588.

In order to find the domain of the proNGF which is better suited for in vivo selection of the SPLINT library, a series of deletion mutants have been created. The starting sequence of the proNGF is the following (SEQ 11) No. 1) : 1 agagagcgcc tggagccgga ggggagcgca tcgagagtga ctttggagct ggccttatat 61 ttggatctcc cgggcagctt tttggaaact cctagtgaag atg ctg tgc ctc aag cca gt 121 g aaa tta ggc tcc ctg gag gtg gga cac ggg cag cat ggt gga gtt ttg gcc tgt ggt cg 181 t gca gtc cag ggg gct gga tgg cat get gga ccc aag ctc acc tca gtg tct ggg ccc aa 241 t aaa ggt ttt gcc aag gac gca gct ttc tat act ggc cgc agt gag gtg cat agc gta at 301 q tcc atg tta ttc tac act etc ! ate act gcg ttt ttq atc aac gta cag gca gaa ccg ta 361 c aca gat agc aat gtc cca gaa gga gac tct gtc cct gaa gcc cac tgg act aaa ctt ca 421 g cat tcc ctt gac aca gcc ctc cgc aga gcc cgc agt gcc cct act gca cca ata gct gc 481 c cga gtg aca ggg cag acc cgc aac atc act gtg gac ccc aga ctg ttt aag aaa cgg ag 541 a ctc cac tca ccc cgt gtg ctg ttc agc acc cag cct cca ccc acc tct tca gac act ct 601 g gat cta gac ttc cag gcc cat ggt aca atc cct ttc aac agg act cac cgg agc aag cg 661 c tca tac acc cac cca gtc ttc cac atg ggg gag ttc tca gtg tgt gac agt gtc agt gt 721 g tgg gtt gga gat aag acc aca gec aca gac atc aag ggc aag gag gtg aca gtg ctg gc 781 c gag gtg aac att aac aac agt gta ttc aga cag tac ttt ttt gag acc aag tgc cga gc 841 c tcc aat cct gtt gag agt ggg tgc cgg ggc atc gac tac aaa cac tgg aac tca tac tg

where: underlined: precursor-long double underlined: pre-sequence of the short precursor italics : pro-sequence of the short precursor bold: NGF times: 5'UTR & 3'UTR List of baits used for the selection of SPLINTS 1. proNGF All the proNGF from 348 to 698 (pro+ first 14 aa of NGF) from SEQ ID No. 1: underlined: NGF The pro-NGF protein was fused in the same reading framework to the lexA binding domain. The level of expression of the described fusion protein was evaluated by Western blot, using anti protein-lexA polyclonal antibodies Pro-NGF was cloned in the polylinker downstream of lexA as a product of PCR cut with the BamHI-PstI enzymes. Primers used: pro Bam for 5'TAT AAT GGG ATC CGT GAA CCG TAC ACA GAT AGC AAT G 3' (SEQ ID No. 3) and pro Pst back 5'TGT ATA TGT ACT GCA GGT CAC ACT GAG AAC TCC CCC ATG 3' (SEQ ID No. 4)

in the pMICBDl vector cut the same way.

2. proNGFmut All the proNGF from 348 to 698 (pro+ first 14 aa of NGF) with mutation in the KKRR site in KEAR: proNGF wild type (SEQ ID No. 2) <BR> <BR> <BR> <BR> MLCLKPVKLGSLEVGHGQHGGVLACGRAVQGAGWHAGPKLTSVSGPNKGFAKDAAFYTGR SEVHSVMSMLFY<BR> <BR> <BR> <BR> <BR> <BR> TLITAFLIGVQAEPYTDSNVPEGDSVPEAHWTKLQHSLDTALRRARSAPTAPIAARVTGQ TRNITVDPRLFK KRRLHSPRVLFSTQPPPTSSDTLDLDFQAHGTIPFNRTHRSKRSSTHPVFHMGEFSVCDS VSVWVGDKTTAT <BR> <BR> <BR> <BR> DIKGKEVTVLAEVNINNSVFRQYFFETKCRASNPVESGCRGIDSKHWNSYCTTTHTFVKA LTTDEKQAAWRF IRIDTACVCVLSRKATRRG* proNGF mut (SEQ ID No. 5) gaa ccg ta 361 c aca gat agc aat gtc cca gaa gga gac tct gtc cct gaa gcc cac tgg act aaa ctt ca 421 g cat tec ctt gac aca gcc ctc cgc aga gcc cgc agt gcc cct act gca cca ata gct gc 481 c cga gtg aca ggg cag acc cgc aac atc act gtg gac ccc aga ctg ttt aaq qaa gcg ag 541 a ctc cac tca ccc cgt gtg ctg ttc agc acc cag cct cca ccc acc tct tca gac act ct 601 g gat cta gac ttc cag gac cat ggt aca atc cct ttc aac agg act cac cgg agc aag cg 661 c tca tcc acc cac cca gtc ttc cac atg ggg gag ttc tca gtg double underlined: mutated site underlined: NGF (SEQ ID No. 6) MLCLKPVKLGSLEVGHGQHGGVLACGRAVQGAGWHAGPKLTSVSGPNKGFAKDAAFYTGR SEVHSVMSMLFY <BR> <BR> <BR> <BR> TLITAFLIGVQAEPYTDSNVPEGDSVPEAHWTKLQHSLDTALRRARSAPTAPIAARVTGQ TRNITVDPRLFK EARLHSPRVLFSTQPPPTSSDTLDLDFQAHGTIPFNRTHRSKRSSTHPVFHMGEFSVCDS VSVWVGDKTTAT <BR> <BR> <BR> <BR> DIKGKEVTVLAEVNINNSVFRQYFFETKCRASNPVESGCRGIDSKHWNSYCTTTHTFVKA LTTDEKQAAWRF IRIDTACVCVLSRKATRRG* Pro-NGF was mutated according to a standard mutagenesis (QuikChange Site-Direct Mutagenesis Kit, Stratagene) in the site indicated above using the primers: M-bait for 5'CCC AGA CTG TTT AAG GAA GCG AGA CTC CAC TCA CCC3' (SEQ ID No. 7) and M-bait back 5'GGG TGA GTG GAG TCT CGC TTC CTT AAA CAG TCT GGG 3' (SEQ ID NO. 8).

3. pro Precursor only without NGF. Sequence from 348 to 656; with mutation (SEQ ID No. 5). gaa ccg ta 361 c aca gat agc aat gtc cca gaa gga gac tct gtc cct gaa gcc cac tgg act aaa ctt ca 421 g cat tcc ctt gac aca gcc ctc cgc aga gcc cgc agt gcc cct act gca cca ata gct gc 481 c cga gtg aca ggg cag acc cgc aac atc act gtg gac ccc aga ctg ttt aaa aaa gcg aa 541 a etc cac tca ccc cgt gtg ctg ttc agc acc cag cct cca ccc acc tct tca gac act ct

601 g gat cta gac ttc cag gcc cat ggt aca atc cct ttc aac agg act cac egg agc aag cg 661 c Double underlined: mutation The Pro was cloned in the polylinker downstream of lexA as a product of PCR cut with the BamHI-PstI enzymes. Primers used: pro Bam for 5'TAT AAT GGG ATC CGT GAA CCG TAC ACA GAT AGC AAT G 3' (SEQ ID No. 3) and pro-only-Pst-back 5'TGT ATA TGT ACT GCA GGT CAG CGC TTG CTC CGG TGA G 3' (SEQ ID No. 9) in the pMICBDl vector cut the same way.

4. pro-10 Only precursor with deletion of 10 aa at the C-term. Sequence from 348 to 626 (SEQ ID No. 5). With mutation gaa ccg ta 361 c aca gat agc aat gtc cca gaa gga gac tct gtc cct gaa gcc cac tgg act aaa ctt ca 421 g cat tcc ctt gac aca gcc ctc cgc aga gcc cgc agt gcc cct act gca cca ata gct gc 481 c cga gtg aca ggg cag acc cgc aac atc act gtg gac ccc aga ctg ttt aag gaa gcg ag 541 a ctc cac tca ccc cgt gtg ctg ttc agc acc cag cct cca ccc acc tct tca gac act ct 601 g gat cta gac ttc cag gcc cat ggt aca atc underlined: mutation The Pro-10 was cloned in the polylinker downstream of lexA as a product of PCR cut with the BamHI-PstI enzymes. Primers used: pro Bam for 5'TAT AAT GGG ATC CGT GAA CCG TAC ACA GAT AGC AAT G 3' (SEQ ID No. 3) and C-term-10 5'TGT ATA TGT ACT GCA GGT CAG ATT GTA CCA TGG GCC TGG 3' (SEQ ID No. 10) in the pMICBDl vector cut the same way.

5. pro-20 Only precursor with deletion of 20 aa at the C-tenn. Sequence from 348 to 596. (SEQ ID No. 5). With mutation gaa ccg ta 361 c aca gat ago aat gtc cca gaa gga gac tct gtc cct gaa gcc cac tgg act aaa ctt ca 421 g cat tcc ctt gac aca gcc ctc cgc aga gcc cgc agt gcc cct act gca cca ata gct gc 481 c cga gtg aca ggg cag acc cgc aac atc act gtg gac ccc aga ctg ttt aag gaa gcg ag 541 a ctc cac tca ccc cgt gtg ctg ttc age acc cag cct cca ccc acc tct tca gac act ct 601 g underlined: mutation The Pro-20 was cloned in the polylinker downstream of lexA as a product of PCR cut with the BamHI-PstI enzymes (primers used:

pro Bam for 5'TAT AAT GGG ATC CGT GAA CCG TAC ACA GAT AGC AAT G 3' (SEQ ID No. 3) and C-term-20 5'TGT ATA TGT ACT GCA GGT CAC AGA GTG TCT GAA GAG GTG G 3' (SEQ ID No. 11) in the pMICBDl vector cut the same way.

6. pro-30 Only precursor with deletion of 30 aa at the C-term. Sequence from 348 to 566 (SEQ ID No. 5). With mutation gaa ccg ta 361 c aca gat agc aat gtc cca gaa gga gac tct gtc cct gaa gcc cac tgg act aaa ctt ca 421 g cat tcc ctt gac aca gcc ctc cgc aga gcc cgc agt gcc cct act gca cca ata get gc 481 c cga gtg aca ggg cag acc cgc aac atc act gtg gac ccc aga ctg ttt aag gaa gcg ag 541 a ctc cac tca ccc cgt gtg ctg ttc agc acc underlined: mutation The Pro-30 was cloned in the polylinker downstream of lexA as a product of PCR cut with the BamHI-PstI enzymes (primers used: pro Bam for 5'TAT AAT GGG ATC CGT GAA CCG TAC ACA GAT AGC AAT G 3' (SEQ ID No. 3) e C-term-30 5'TGT ATA TGT ACT GCA GGT CAG GTG CTG AAC AGC ACA CGG 3' (SEQIDNo. 12) inthepMICBDl vector cut the same way.

7. pro-41 Only precursor with deletion of 41 aa at the C-term. Sequence from 348 to 533. (SEQ ID No. 5). With mutation gaa ccg ta 361 c aca gat agc aat gtc cca gaa gga gac tct gtc cct gaa gcc cac tgg act aaa ctt ca 421 g cat tcc ctt gac aca gcc ctc cgc aga gcc cgc agt gcc cct act gca cca ata gct gc 481 c cga gtg aca ggg cag acc cgc aac atc act gtg gac ccc aga ctg ttt aag gaa gcg underlined: mutation The Pro-41 was cloned in the polylinker downstream of lexA as a product of PCR cut with the BamHI-PstI enzymes (primers used: pro Bam for 5 TAT AAT GGG ATC CGT GAA CCG TAC ACA GAT AGC AAT G 3' (SEQ ID No. 3) and C-term-41 5'TGT ATA TGT ACT GCA GGT CAC GCT TCC TTA AAC AGT CTG GG 3' (SEQ ID NO : 13) in the pMICBDl vector cut the same way.

8. pro N-10 Only precursor with deletion of 10 aa at the N-term. Sequence from 378 to 656 (SEQ ID No. 5). With mutation gga gac tct gtc cct gaa gcc cac tgg act aaa ctt ca 421 g cat tcc ctt gac aca gcc ctc cgc aga gcc cgc agt gcc cct act gca cca ata gct gc

underlined: mutation The Pro N-10 was cloned in the polylinker downstream of lexA as a product of PCR cut with the BamHI-PstI enzymes (primers used: N-term-10 5'TAT AAT GGG ATC CGT GGA GAC TCT GTC CCT GAA G 3' (SEQ ID No. 14) and pro-only-Pst-back 5'TGT ATA TGT ACT GCA GGT CAG CGC TTG CTC CGG TGA G 3'(SEQ ID No. 9) in the pMICBDl vector cut the same way.

Pro-NGF and the various deletion mutants were expressed in the yeast strain L40 in which the preliminary tests were conducted to verify protein expression and the transactivating capabilities of the various baits. All baits were found to be expressed in good quantity. The results of the transactivation tests are listed in Table 2.

The proNGF antigen and its deletion mutants are expressed in excellent quantity in the yeast but most of them were extremely strong transactivators for the reporter genes (HIS3 and lacZ) used in our assay (see Fig. 18 and Tables 2 and 3).

Table 2. proNGF and its deletion mutants were expressed in the yeast strain L40 and the transactivating property of the His and lacZ genes for each bait was observed. As is readily apparent, only the pro-30 bait does not exhibit transactivation of the lacZ gene, though the histidine gene remained transactivating. bait-UKW-UKWH X-gal proNGF + + + proNGFmut + + + pro + + + Pro-10 + + + Pro-20 + + +/- Pro-30 + + - Pro-41 +/ proN-10 + + +

Table 3. proNGF and its deletion mutants were expressed in the yeast strain L40 together with the activator domain of the VP16 transcription and the transactivating property of the His and lacZ genes for each bait was observed. As is readily apparent, only the pro-30 bait does not exhibit transactivation of the lacZ gene, though the histidine gene remained transactivating. Bait + VP16-UKWL-WHULK X-gal ProNGF+ VP16 + + + ProNGFmut + VP 16 + + + Pro + VP16 + + + Pro-10 + VP16 + + + Pro-20 + VP16 +/ Pro-30 + VP16 + + Pro-41 + VP 16 + + + proN-10 + VP16

To lower the transactivating property of the various baits constructed, some tests were made using different concentrations of the 3AT histidine inhibitor. Table 4 describes the results obtained: Table 4. Transactivation test of the proNGF mutants in medium containing the 3AT histidine inhibitor. As is readily apparent, 3AT was found to be quite an effective method to eliminate the transactivation of the histidine reporter gene. In light of these results, the pro-10 mutant was selected, grown on a medium which contains 75mM of 3AT as bait for the selection of the SPLINT library. bait-UKWH-UKWH-UKWH-UKWH-UKWH-UKWH + 25mM 3AT +SOmM3AT +75mM3AT +100mM3AT +125mM3AT +150mM3AT proNGF - proNGFmut - pro +/- Pro-10 + + Pro-20 +/- - Pro-30 Pro-41 proN-10 +/-+/-

Example 6 Pro-10 was chosen as bait for the selection of scFvs via SPLINT The SPLINT library was challenged against lexA-proNGF-10 as described in (Visintin et al. , 1999; Visintin and Cattaneo, 2001; Visintin et al. , 2002). Only 9 clones grew in a medium without histidine + 75mM 3AT and became blue after the beta-gal assay. All nine clones were isolated from the yeast following the method described in Visintin, M., et al. , 2001, and the isolated DNA was analysed for BstNI fingerprinting and sequencing as described. The individual clones are further challenged against proNGF-10 and a non relevant antigen (lamin) by means of IACT. Only one of these clones was found to be a true positive.

The deletion mutants of proNGF pro-10, pro-20, pro-30 and pro-41 were secondarily used for an in vivo epitope mapping (IVEM) (the technology is described in the international patent application W00235237A, incorporated in full herein) of the other selected anti- proNGF. The results of this experiment are summarised in Table 5 and in Figure 19.

Table 5: The results obtained show that IVEM has allowed to identify the epitope (Ll- L106) on the proNGF protein recognised by the antibody selected with IACT. Baits + anti-proNGF-10-UKWL-WHULK X-gal + 40mM 3AT +40mM3AT Pro-10 ++ Pro-20 Pro-30 +/- Pro-41 + +/-+ The anti-proNGF scFv was also sequenced. The sequence is highlighted below.

Amino acidic sequences of the light chain (VL) and heavy chain (VH) of the anti- proNGF VL (SEQ ID No. 15) DAEIVLTKSPALMFASPGEKVTMTCSASSSVSYMHWYQRSKAPPPKLWIYDTSKLASGVP GRFSGSGSGNSY SLTISSMEAEDVATYYLFSGEWVPVHVRRGTKLEIKR VH (SEQ ID No. 16) EVMLVESGGGLVKPGGSLKLSCTASGFTFSNYAMSWIRQSPEKRLEWVGEISNGGSNTYY PGSVTGRFTISR DNAEDTLYLEMSRLRSEDTAIYYCVRDGFYAMDYWSQGTSVTVSSASV Nucleotidic sequence of the anti-proNGF scFV (SEQ ID No. 17) 3 CAAAAAAAAAAAAAAAGTGGCCAAGCGGAGCGCGATGCCGAAATTGTTCTCACCAAGTCT 62 63 CCAGCACTCATGTTTGCATCTCCAGGGGAGAAGGTCACCATGACCTGCAGTGCCAGCTCA 122 123 AGTGTAAGTTACATGCACTGGTACCAGAGAAGTAAAGCACCTCCCCCAAAACTCTGGATT 182 183 TATGACACATCAAAACTGGCTTCTGGAGTCCCAGGTCGCTTCAGTGGCAGTGGGTCTGGA 242

243 AACTCTTACTCTCTCACGATCAGCAGCATGGAGGCTGAAGATGTTGCCACTTATTATCTG 302 303 TTTTCAGGGGAGTGGGTACCCGTACACGTTCGGAGGGGGACAAAGTTGGAAATAAAACGT 362 363 TCCGGAGGGTCGACCAGCGGTTCTGGGAAACCAGGTTCCGGTGAAGGCTCGAGCGGTACC 422 423 GAAGTGATGCTGGTGGAGTCTGGGGGAGGCTTAGTGAAGCCTGGAGGGTCCCTGAAACTC 482 483 TCCTGTACAGCCTCTGGATTCACTTTCAGCAACTATGCCATGTCTTGGATTCGCCAGTCT 542 543 CCAGAGAAGAGGCTGGAGTGGGTCGGAGAAATTAGTAACGGCGGTAGTAACACCTACTAT 602 603 CCAGGCAGTGTGACGGGCCGATTCACCATCTCCAGAGACAATGCCGAAGACACCCTGTAC 662 663 CTGGAAATGAGCAGGCTGAGGTCAGAGGACACGGCCATATATTACTGTGTAAGGGACGGT 722 723TTCTATGCTATGGACTACTGGAGTCAAGGAACCTCAGTCACCGTCTCCTCAGCTAGC GTC 782 underlined: VL double underlined: VH Anti-proNGF scFV amino acidic sequence (SEQ ID No. 18) DAEIVLTKSPALMFASPGEKVTMTCSASSSVSYMHWYQRSKAPPPKLWIYDTSKLASGVP GRFSGSGSGNSY SLTISSMEAEDVATYYLFSGEWVPVHVRRGTKLEIKRSGGSTSGSGKPGSGEGSSGTEVM LVESGGGLVKPG GSLKLSCTASGFTFSNYAMSWIRQSPEKRLEWVGEISNGGSNTYYPGSVTGRFTISRDNA EDTLYLEMSRLR SEDTAIYYCVRDGFYAMDYWSQGTSVTVSSASVS In order to characterise the anti-proNGF antibody from the biochemical aspect, the pro- NFG protein and the anti-proNGF scFv were made to be expressed in E. Coli and then purified according to methods described below: anti-proNGF The anti-proNGF selected and validated with IACT was further analysed with biochemical methods to verify its solubility, propensity to aggregation and stability (Worn and Pluckthun, 1999). The scFv was cloned in the vector DAN3 (Sblattero and Bradbury, 2000) in the sites BssH2-NheI and expressed in the non suppresser bacterial strain HB2151. The periplasmatic extract was subsequently purified by means of NiNTA affinity column for proteins with His-tag (see Figure 20) followed by an additional purification on filtration gel column (superdex 75) (see Figure 21).

Analysis of the antibody purified in filtration gel has shown that the analysed scFv is mostly found in the elution peak corresponding to the monometric form of an scFv (Fig.

21).

Example 7 Characterisation by means of ELISA To evaluate the degree of affinity of the purified antibody, an ELISA (Enzyme-Linked Immunosorbent Assay) test was used. On a 96-well microtiter plate (Falcon)

functionalised with the antigen (or with the antibody, in the second experiment) the antibody's specific recognition ability is assayed.

The functionalisation was obtained dissolving the protein in a 100 mM NaHCO3, solution, pH 9.6 according to a concentration of 10yg/mL, and filling each well of the plate with 100 AL of the solution obtained. The plate was then left in incubation for 16 hours at 4°C and subsequently subjected to two washings with PBS.

Before incubation with the antibodies or with the antigens, a"blocking"passage has to be performed, in which the wells were filled with 200 jus of a 3% solution powdered skim milk in PBS (MPBS), and incubated for 1 hour at ambient temperature. After washings with PBS, it was possible to proceed with the antibody or antigen incubation, in a 3% MPBS solution (100 its in each well). The dilutions used are shown schematically in Table 2.

The incubations were conducted for 2 hours at ambient temperature; after each incubation, the well was washed with TPBS (PBS solution with 0. 05% of Tween 20) and PBS for three times.

This was followed by incubation with the primary antibody (1°AB), diluted in a 3% MPBS solution (100 uL in each well). The plate was incubated for 90 minutes at ambient temperature and subsequently washed 3 times with TPBS and PBS.

Subsequently, incubation was conducted with the secondary antibody (2°Ab), of different species according to the primary antibody used, but always conjugated with horseradish peroxidase (HRP) for the subsequent colorimetric recognition, according to the dilutions shown in Table 2. The incubation (100 yL in each well) lasted 1 hour at ambient temperature.

After washing with TPBS and PBS, the colouring development was conducted, which takes place with TMB (3,3', 5, 5'-tetramethylbenzidine). As soon as the solution starts to colour in the plate, development is blocked with H2SO4 0.5 M and the spectrophotometric reading of the plate at 450 nm is conducted.

To achieve the ideal conditions for the characterisation of the anti-proNGF in ELISA, a panel of proteins was prepared: - anti-pro-10 : ScFv obtained as described in the previous paragraph - ScFv control: expressed in laboratory -m-NGF : commercial (company: Alomone) -h-NGF

-m-proNGF : amplification of the mouse proNGF (m-proNGF) for RT-PCR from lysate of mouse submandibular glands. Cloning in the vector pETM-13 (EMBL protein expression and purification unit) for expression in E. Coli. Expression, refolding and purification according to the protocol described in Rattenholl et al., 2001 b -h-proNGF GST : expressed by the vector pGEX-4T-2 (Amersham Bioscience) as a soluble protein (grown in LB medium with Ampicilline antibiotic until an OD600=0. 5; induction of the expression with IPTG 1 mM; induction for 5h at 30°C). Purified by means of affinity chromatography for GST (GSTrap, Amersham Bioscience) and eluted in PBS+ 10 mM glutation, dialysed in PBS with Spectra-Por 8/1 OK (Spectrum) membranes and concentrated with Centricon 10KDa (Amicon) ultrafiltration unit. The concentration of the purified protein was determined by Lowry assay (Bio-Rad).

GSTpro : fusion protein of the NGF precursor with GST at the N-terminal. It was obtained by cloning only the precursor of the murine NGF in the vector pGEX-4T-2 for expression in E. Coli. Expressed as a soluble protein (grown in LB medium with Ampicilline antibiotic until an OD600=0. 5; induction of the expression with IPTG 0.5 mM; induction for 5h at 30°C). Purified by means of affinity chromatography for GST (GSTrap, Amersham Bioscience) and eluted in PBS+ 10 mM glutation, dialysed in PBS with Spectra-Por 8/1OK (Spectrum) membranes and concentrated with Centricon lOKDa (Amicon) ultrafiltration unit. The concentration of the purified protein was determined by Lowry assay (Bio-Rad).

-ProGST : fusion protein of the NGF precursor with GST at the C-terminal. It was obtained by cloning only the precursor of the murine NGF in the vector pET41-b (Novagen) for expression in E. Coli. Expressed as a soluble protein (grown in LB medium with Kanamycin antibiotic until an OD600=0. 5; induction of the expression with IPTG 1 mM; induction for 5h at 30°C). Purified by means of affinity chromatography for GST (GSTrap, Amersham Bioscience) and eluted in PBS+ 10 mM glutation, dialysed in PBS with Spectra-Por 8/1 OK (Spectrum) membranes and concentrated with Centricon 10KDa (Amicon) ultrafiltration unit. The concentration of the purified protein was determined by Lowry assay (Bio-Rad).

-aSV5 : surnatant, obtained in the laboratory (Hanke et al. , 1992) 9E10 : surnatant, obtained in the laboratory (Evan et al. , 1985) - aDll : surnatant, obtained in the laboratory (Cattaneo et al., 1988)

- all secondary antibodies conjugated with peroxidase are commercial (DAKO) The experimental parameters adopted in the various ELISA are summarised in Tabl Figures 22 and 23 instead show the results obtained in the various ELISA assays.

Table 6 Coating incubation 1° Ab Experiment 1 m-NGF αpro-10 (3 µg/mL) 20 & 21 αSV5 α mouse HRP (1: 1000) (surnatant, 1: 10) h-NGF apro-10 (3 pg/mL) 20 & 21 aSV5 a mouse HRP (1: 1000) (surnatant, 1: 10) m-proNGF αpro-10 (3 µg/mL) 20 & 21 αSV5 α mouse HRP (surnatant, 1: 10) (1: 1000) GSTpro: αpro-10 (3 µg/mL) 20 & 21 αSV5 a mouse HRP (1: 1000) (surnatant, 1: 10) ProGST apro-10 (3 pg/mL) 20 & 21 aSV5 a mouse HRP (surnatant, 1: 10) (1: 1000) GST apro-10 (3 pg/mL) 20 & 21 aSV5 a mouse HRP (surnatant, 1: 10) (1: 1000) m-NGF ScFv control (2 µg/mL) 9E10 α mouse HRP (surnatant, 1: 10) (1: 1000) h-NGF ScFv control (2 pg/mL) 9E10 a mouse HRP (surnatant, 1: 10) (1: 1000) m-proNGF ScFv control (2 pg/mL) 9E10 a mouse HRP (1: 1000) (surnatant, 1: 10) GSTpro: ScFv control (2 pg/mL) 9E10 α mouse HRP (1: 1000) (surnatant, 1: 10) ProGST ScFv control (2, ug/mL) 9E10 a mouse HRP (1: 1000) (surnatant, 1: 10) GST ScFv control (2 pg/mL) 9E10 α mouse HRP (1: 1000) (surnatant, 1: 10) Experiment 2 αpro-10 h-NGF (10 µg/mL) αNGF (1:8000) α rbt HRP (1: 2000) apro-10 h-proNGF (10 pg/mL) aNGF (1: 8000) a rbt HRP (1 : 2000) apro-10 m-proNGF (10 Hg/mL) aNGF (1: 8000) a rbt HRP (1: 2000) apro-10 h-NGF (10, ug/mL) aD11 a rat HRP (1: 2000) (surnatant 1: 2) apro-10 h-proNGF (10 pg/mL) aD11 a rat HRP (1: 2000) (surnatant 1: 2) apro-10 m-proNGF (10, ug/mL) aD11 a rat HRP (1: 2000) (surnatant 1: 2) apro-10 GSTpro (10 vg/mL) aGST (1: 8000) a rbt HRP (1: 2000) αpro-10 ProGST (10 µg/mL) αGST (1: 8000) a rbt HRP (1: 2000) αpro-10 GST (10 µg/mL) αGST (1: 8000) a rbt HRP (1: 2000) From the sensograms provided in Example 4, several considerations are possible.

-The alfaDll antibody interacts differently with NGF and with proNGF. In particular, NGF has a greater affinity for the antibody, both because of a very rapid association, and because of a very slow dissociation. The proNGF, whose association is slower and whose dissociation is much more rapid, has a different behaviour.

The behaviour of both antigens is specific, since in the case of the channel with aSV5 the same type of effect does not take place.

The interaction of the proNGF with alfaDll is due solely to the NFG part, since the negative GSTpro (containing the part of the precursor in fusion with the GST) does not respond to the antibodies in any way.

It can therefore be deduced that, when expressed in vivo, the alfaDll antibody is able to alter the balance between pro-survival action of NGF and the pro-apoptotic function of pro-NGF because a determines a practically irreversible sequestration of mature NGF.

Therefore this antibody can be used to set up systems for selecting molecules with a pro- NGF blocking action. The selection may be conducted in vitro and in vivo, through the expression of recombinant antibodies in murine models. The prototype of these models is represented by the anti-NGF AD11 mouse.

An example of these inhibitors is represented by the anti-proNGF antibody. The authors have isolated an antibody in scFv anti-proNGF format which specifically recognises proNGF both in vivo and in vitro. The anti-proNGF is also highly soluble and it has good stability and it could be the first representative of a class of molecules and of antibodies neutralising the toxic activity of the pro-NGF.

Example 8. Prevention of the cholinergic deficit in anti-NGF mice by implanting hybridomas secreting antibodies neutralising pro-NGF Anti-proNGF hybridomas (secreting monoclonal antibodies neutralising the pro-10 mutated form (see Example 5) of proNGF, selected as in Example 6 and myeloma P3X63Ag8 (P3U, cellular control line) were grown Dulbecco's modified Eagle's culture medium supplemented with 10% of bovine foetal serum. Before injection, cells were washed four times in the Hank saline solution (HBSS) and re-suspended in HBSS at the concentration of 2 x 105 cells/llL The P3U cellular control line and the cells secreting the anti-proNGF antibody were injected into the right lateral ventricle of anti-NGF transgenic mice. All injections were performed 8 days after birth. For this purpose, the mice were anaesthetised in icy water.

The cells were injected using a 17 gauge diameter needle connected, through a polyethylene cannula, to a 25 al Hamilton syringe. The injection point was identified by the following reference points: 1.5 mm lateral to median suture and 0.5 mm at Bregma.

One microlitre of the cell suspension was injected into each mouse. After the injection, the mice were treated with cyclosporin A (Roche, Basel, Switzerland, 15 mg/kg) every second day, to prevent rejection of the implant. The injection of hybridomas allows a constant supply for about 15-20 days in the basal forebrain (Molnar et al., 1998). After this period, the levels of the antibody decrease until they become insufficient for any neutralising action (Molnar et al, 1998)..

The mice injected with the hybridomas were allowed to grow until reaching 2 months and 6 months of age. They were then sacrificed by injection of Chloralium hydrate (8 J. L of a 10.5% solution per animal) and the tissues fixed by means of transcardiac perfusion of a cold 4% paraformadehyde solution in phosphate buffer.

The brain was removed and post-fixed in the same fixative. After cryoprotection in 30% saccharose, the brains were sectioned using a refrigerated slide microtome.

The sections containing the basal forebrain were processed to highlight the number of cholinergic cells. For this purpose, the sections were washed twice in Tris/HCl 0.1 M buffer to which were added 0.15 M of NaCl and 0.3% of Triton-X 100 (TBST). After blocking the endogenous peroxidase in a 3% solution of hydrogen peroxide, non specific bindings were blocked in 10% of bovine foetal serum dissolved in (TBST). After half an hour, the sections were placed in incubation with the primary antibody directed against choline acetyltransferase (antibody made in anti-ChAT goat; 1: 500, Chemicon, Temecula). Incubation took place at 4°C for 12 hours.

After washing the sections in Tris/HCl 0.1 M buffer to which were added 0.15 M of NaCl (TBS), the sections were incubated for 3 hours in the biotylinated secondary antibody directed against goat IgG (1: 200, Vector Labtek).

The signal deriving from the reaction between primary and secondary antibody was amplified by means of incubation with a peroxidase (Vector Labtek). The reaction was then developed by incubation with the diaminobenzidine chromogen (Sigma, St. Louis, MO).

After collecting the sections on gelled slides, the number of ChAT-positive cholinergic cells was quantified with neurostereology techniques. The basal forebrain is delimited dorsally by the corpus callosum. The anatomic boundaries that define the basal forebrain

are: dorsally, the corpus callosum; centrally, the ventral surface of the brain; laterally, the medial line that passes through the rostral branch of the front commissure. The rostral boundary is determined by the plane that passes through the knee of the corpus callosum, whilst the caudal boundary is constituted by the plane that passes through the first section that contains the front commissure. All anatomical references were taken from the Franklin and Paxinos (1997) mouse Atlas.

The volume of the basal forebrain was calculated using Cavalieri's method (Michel and Cruz-Orive, 1988; Peterson et al. , 1999). A grid was overlaid on the monitor of the computer on which the images of the basal forebrain were projected using a 2. 5X lens.

The volume of each square of the grid was calculated multiplying the thickness of the sections (T) times the area of each square that formed the grid (a). The total volume (Vref) was obtained counting the squares superposed on the basal forebrain in each section (Pi) and multiplying the sum thereby obtained times the volume of each square of the grid, according to the formula Vref= (SPi) (a) (T).

The estimate of the total number of the cholinergic neurons of the basal forebrain was obtained with the optical fractionating method (West, 1993; Peterson et al. , 1999). The images, acquired at 60X enlargement with a Zeiss transmitted light microscope connected to a CCD television camera, were projected on the monitor of a computer. Numerical density (Nv) was determined using a grid superposed on the monitor screen. The number of neurons for each grid was counted by focusing in tissue at a known distance (20 um), as described by Sterio (1984). The number of cells calculated in this volume provides the average number of cells/volume (Nv). The total number of neurons was calculated multiplying Nv times the total volume of the basal forebrain (calculated according to Cavalieri's procedure). The volume and the total number of neurons was calculated for n = 4 for each experimental group. Statistical analysis was conducted according to T-tail test.

The quantification of the total number of cholinergic neurons in the basal forebrain has demonstrated that the injection of hybridomas secreting the anti-proNGF neutralising antibodies at an early age determines a prevention of the cholinergic deficit which will take place at a late age.

Example 9. Inhibition of the binding of proNGF to sortilin by incubation with anti- proNGF.

The inhibition by the anti-proNGF antibody of the binding of proNGF to sortilin was studied according to the method described in Mahadeo et al. (1994) and Esposito et al.

(2001). The human melanoma cells A875 (American Type Culture collection), transfected to determine the expression of the sortilin receptor (2x10/ml) were incubated (4 °C, 40 min) with Il25-proNGF (2-20 x 10-1° M) in the presence or absence of an excess of anti-proNGF (500 times in excess relative to the concentration of pro-NGF). The ligand bound to the receptor was then separated from the free pro-NGF ligand in the middle by centrifuging in the presence of bovine serum. The radioactivity of Ilzs-proNGF bound to the cell was determined in the pellet. The specificity of the binding of Il25- proNGF to the sortilin receptor was determined subtracting the counts obtained after incubation in the presence of an excess of pro-NGF not marked with the radioactive element. The experiments were conducted in triplicate and they demonstrated that anti- proNGF is able to inhibit the specific binding of pro-NGF to the sortilin receptor (92% inhibition).

Example 10. Effects of knocking-out the selector p75 on the Alzheimer phenotype of anti-NGF mice It will be demonstrated that the p75/sortilin receptor complex is involved in the pathogenesis of the Alzheimer phenotype observed in anti-NGF mice and in AD11-VK mice. As described above, in patent application RM2004A000212, the AD11-VK mice have a neurodegenerative phenotype which can be superposed to that of AD 11 mice. The AD11-VK mice have been crossed with p75NTR°" (-/-) mice, in which the genic exon corresponding to the extracellular domain of the p75NTR receptor was deleted by gene knockout with homologous recombination (Jackson Laboratories, Lee et al. , 1992).

Therefore in p75NTR°" (-/-) mice, signalling through p75NTR is inhibited. From this cross, AD 12 have thus been obtained in which there is the expression of the light chain of the oDll antibody and simultaneously the lack of expression of that portion of the receptor whereon the proNGF, the NGF and the other neurotrophins are bound.

The evaluation of the effect on the Alzheimer phenotype was conducted by means of histological analysis. For this purpose, the AD12 mice, the AD11-Vk mice, the p75NTRexonIII(-/-) mice and non transgenic mice were then were then sacrificed by injection of Chloralium hydrate (8 pL of a 10.5% solution per animal) and the tissues

fixed by means of transcardiac perfusion of a cold 4% paraformadehyde solution in phosphate buffer.

The brain was removed and post-fixed in the same fixative. After cryoprotection in 30% saccharose, the brains were sectioned using a refrigerated slide microtome. The sections containing the hippocampus were processed to highlight the number of cholinergic cells.

For this purpose, the sections were washed twice in Tris/HCl 0.1 M buffer to which were added 0.15 M of NaCl and 0. 3% of Triton-X 100 (TBST). After blocking the endogenous peroxidase in a 3% solution of hydrogen peroxide, non specific bindings were blocked in 10% of bovine foetal serum dissolved in (TBST). After half an hour, the sections were placed in incubation with the primary antibody directed against the ß-amyloid protein (antibody made in anti-A4 goat; 1: 100, Santa Cruz, Santa Cruz). Incubation took place at 4°C for 72 hours. After washing the sections in Tris/HCl 0.1 M buffer to which were added 0.15 M of NaCl (TBS), the sections were incubated for 3 hours in the biotylinated secondary antibody directed against goat IgG (1: 200, Vector Labtek). The signal deriving from the reaction between primary and secondary antibody was amplified by means of incubation with avidin conjugated to alkaline phosphatase (Vector Labtek). The reaction was then developed by incubation with the nitroblue tetrazolium chromogens and the p- bromo-chloro indolyl toluidine salt (Sigma, St. Louis, MO). After collecting the sections on gelled slides, the number of deposits marked with beta-amyloid was quantified with neurostereology techniques.

The results, shown in Figure 24, have shown that the AD11VK mice have a high number of beta-amyloid deposits, whilst AD12 mice have a number of deposits comparable to that of non transgenic mice. This result demonstrated that the amyloid component of the neurodegenerative phenotype in AD11 mice and in AD11-VK mice disappears if signalling through p75NTR is inhibited.

BIBLIOGRAPHY Aigner, TG, et al. (1991) Exp Brain Res 86: 18-26.

Arendt, T, et al. , (1985) Neuroscience 14: 1-14.

Capsoni, S, et al. (2000) Proc Natl Acad Sci USA 97: 6826-6831.

Capsoni, S, et al. (2002a) Mol Cell Neurosci 21: 15-28.

Capsoni, S, et al (2002b) Brain Aging 2: 24-43.

Cattaneo, A et al. (1988) J Neurochem 50: 1003-1010.

Cattaneo, A, et al. (1999) J Neurosci 19: 9687-9697.

Chao, MV, et al. (1986) Science 232: 518-521.

Chu, ZL, et al. (1997) Proc Natl Acad Sci USA 94: 10057-10062.

Crutcher, KA, et al. (1993) J Neurosci 13: 2540-2550.

Debeir, T, et al. (1999) Proc Natl Acad Sci USA 96: 4067-4072.

DeKosky, et al. (1992) Ann Neurol 32: 625-632.

Drachmann, DA, & Leavitt, J (1974) Arch Neurol 30: 113-121.

Esposito, D. et al. J. Biol. Chem. 276,32687-32695 (2001).

Etienne, P, et al. (1985) Neuroscience 19: 1279-1291.

Evan, G. I. , et al. (1985) In: Mol Cell Biol, Vol. 5, p. 3610-6.

Fagan, AM, et al (1997) J Neurosci 17: 7644-7654.

Fahnestock, M, et al (1996) Mol Brain Res 42: 175-178.

Fahnestock M. , et al. 2001, Mol. Cell. Neurosc. , vol. 18, pp. 210-220 Fahnestock M. , et al. 2004, J. Neurochem, vol. 89, pp. 581-592 Fibiger, HC (1991) Trends Neurosci 14: 220-223.

Galfre, G. e Milstein, C. (1981) Methods Enzymol 73: 3-45.

Geula, C (1998) Neurology 51: S18-S29.

Hamers-Casterman, C. et al. (1993) Nature 363,446-448.

Hanke, T. , Szawlowski, P. and Randall, R. E. (1992) J Gen Virol 73,653-60.

Kaplan, DR et al. (1991) Nature 350: 158-160.

Klein, R et al. (1991) Cell 65: 189-197.

Korhonen, L, et al. (2001) Mol Cell Neurosci 17: 364-372.

LeBlanc, AC (2003) Prog Neuropsycol Biol Psych 27: 215-229.

Lee KF, et al. Cell 69,737-749, 1992.

Lee, VM, et al. (2001) Annu Rev Neurosci 24: 1121-1159.

Lee R. , et al. Science, vol. 294, pp. 1945-1948 Lehericy, S, et al. (1993) J Comp Neurol 330: 15-31.

Li, Y, et al. (1995) J Neurosci 15: 2888-2905.

Mahadeo D, et al. J Biol Chem. 269, 6884-6891, 1994.

Marin, DB, et al. (2002) Geriatrics 57: 36-40.

McGeer, PL, & McGeer, EG (2001) Neurobiol Aging 22: 799-809.

Molnar, M, et al. (1998) Eur J Neurosci 10: 3127-3140.

Mrzljak, L, & Goldman-Rakic, PS (1993) Cereb Cortex. 3: 133-147.

Mufson, EJ, et al. (1996) NeuroReport 8: 25-29.

Mufson, EJ, et al. (1997) Exp Neurol 146: 91-103.

Nanduri, J, et al. (1994) J Neurosci Res 37: 433-444.

Nykjaer A. , et al. 2004, Sortilin is essential for proNGF-induced neuronal cell death, Nature, vol. 427, pp. 843-848 Nilsson, et al. (1986) J Neural Transm 67: 275-285.

Nordberg, A, et al. (1992) J Neurosci Res 31 : 103-111.

Rattenholl A. , et al. 2001 a, Eur. J. Biochem. , vol. 268, pp. 3296-3303 Rattenholl A. , et al. 2001 b, J. Mol. Biol. , vol. 305, pp. 523-533 Ruberti, F. , et al. (2000) J Neurosci 20: 2589-2601.

Rylett, RJ, et al. (1983) Brain Res 289: 169-175.

Salehi, A, et al. (2000) Exp Neurol 161: 245-258.

Salvesen, GS, & Duckett, CS (2002) Nature Mol Cell Biol Rev 3: 401-410.

Schubert, P, et al. (2001) Mech Ageing Dev 123: 47-57.

Scott, SA, et al. (1995) J Neurosci 15: 6213-6221.

Selkoe, D (2001) Physiol Rev 81: 741-766.

Tran, J, et al. (1999) BBRC 264: 781-788.

Vantini, G, (1989) Neuron 9: 267-273. C Sblattero, D. and Bradbury, A. (2000) Nat Biotechnol 18, 75-80.

Suter U. , Heymach J. V. Jr and Shooter E. M., 1991, The EMBO Journal, vol. 10 no. 9 pp. 2395-2400 Visintin, M. and Cattaneo, A. (2001) In: R. Kontermann, & Dubel, S. , eds (Ed) Antibody Engineering, Vol. 1, Springer Lab Manual. Springer, Heidelberg, Germany, p. 790.

Visintin, M. , et al. (2002) J Mol Biol 317,73-83.

Visintin, M. , et al. (1999) Proc Natl Acad Sci U S A, Vol. 96, p. 11723-8.

Visintin, M. , et al. (2004). J Immunol Methods, Vol. 290, p. 135 Wenk, GL (1997) Neurobiol Learn Mem 67: 85-95.

Whitehouse, P (1982) Science 215: 1237-1239.

Wilcock, GK (1982) J Neurol Sci 57: 407-417.

Wiese, S, et al. , (1999) Nat Neurosci 2: 978-983.

Worn, A. and Pluckthun, A. (1999) Biochemistry 38, 8739-50.