This invention relates to neuroregeneration.

Background to the Invention

Various mammalian diseases are known which involve progressive chronic degeneration of central nervous system (CNS) and/or peripheral nervous system (PNS) neurons. Examples of such diseases that have been closely studied in humans include Parkinson's Disease, Alzheimer's Disease and various motor neurone diseases.

In Alzheimer's Disease, lesions occur in cerebral cortex neurons, in particular cholinergic neurons projecting from the substantia innominata to the cortex. Treatment regimes are aimed at enhancing the activity of the remaining neurons.

Motor neurone disease, in which there is a chronic degeneration of CNS and PNS motor neurons resulting in a loss of muscle strength and control, can be treated in similar ways.

Acute conditions involving brain injury include the following:

1 ) acute mechanical trauma with axotomy (for example,

penetrating brain injury and spinal cord transection);

2) cerebrovascular accident due to haemorrhage, thrombosis or embolism;

3 ) brain and spinal cord damage due to bacterial or viral infection, and tumour or abscess expansion.

A disadvantage common to the known treatments of the diseases and conditions outlined above is that the treatment is aimed at enhancing the activity of remaining neurons rather than in replacing lost neurons or halting the degeneration that causes the disease. As a result, higher doses of relevant drugs are required as the disease progresses. Further, it is likely that in time the disease goes beyond the point at which treatment can be effective.

A further disadvantage is that there are side effects associated with the drugs used for treating the diseases.

The present invention aims to mitigate at least some of these disadvantages.

Davies et al, 1990 describes the isolation from the posterior half of chick embryo somites of a glycoprotein fraction that causes growth cone collapse of chick sensory neurons. The glycoproteins bind to the lectin peanut agglutinin (PNA) and were shown by SDS-PAGE under reducing conditions to have two major components with apparent molecular weights of 48K and 55K.

The present invention is based on the extraction from grey matter of one or more glycoproteins that have been

implicated in neurodegeneration.

Summary of the Invention

In one aspect the present invention provides glycoprotein obtainable from grey matter, and capable of inducing growth cone collapse.

It is not at present clear whether there are one or more such glycoproteins with growth cone collapse properties.

The glycoprotein is further characterised by dye binding properties, and in particular is capable of binding to the commercially available dye Mimetic Green (obtained from Affinity Chromatography Limited) but does not bind to the dye Mimetic Blue 2.

The glycoprotein is retained by a membrane of 100,000 molecular weight cut off, e.g. an Amicon Diaflow (Trade Mark) YM type membrane with 100,000 molecular weight cut off used in an Amicon Centriprep Concentrator or an Amicon Centricon Concentrator (Centriprep and Centricon are Trade Marks) .

Glycoprotein in accordance with the invention has been extracted from grey matter from a wide range of sources, including mammalian sources such as human, sheep, rat, pig, avian sources such as chicken, and fish sources such as carp. Cross-reactivity of material from different sources has been tested e.g. with chicken-derived glycoprotein being tested for the effect on rat nerve cell growth cones, and vice versa, and similarly for chicken and human material, and in all cases growth cone collapse is induced. For practical reasons, chicken nerve cells are generally used to test growth cone collapse activity

of human-derived glycoprotein.

The glycoprotein of the invention is conveniently extracted from grey matter by chromatography on immobilsed dyes, e.g. Mimetic Green 1 as mentioned above.

The glycoprotein has yet to be fully characterised. However, the cleanest preparations of inhibitory glycoprotein, isolated from avian grey matter as a single band on a non-denaturing gel, show 2 bands to be present at approximate molecular weights of 75K and 62K when examined by SDS-PAGE under reducing conditions . By comparison with non-active material, we have evidence that the lower of these 2 bands is derived from the biologically-active material.

The ability of the glycoprotein to induce growth cone collapse, e.g. as determined by the growth cone collapse assay described in Cook et al, 1991, suggests it may be implicated in the failure of CNS regeneration as well as neurodegeneration, so that it can be used as the basis of attempts to stimulate nerve regeneration by interfering (up or down regulating) with the functioning of the glycoprotein. This view is substantiated by experimental evidence obtained by the present inventors.

They have investigated the quantitative and regional expression of the growth cone collapse-inducing activity in the normal human brain and in that of patients with Alzheimer's disease (AD). Growth cone collapse-inducing activity is substantially diminished in grey matter from AD brain: detergent (CHAPS) extracts of cerebral cortex from 4 normal brains (4-30 hours post-mortem delay; tissue obtained collectively from the frontal, parietal, temporal and occipital lobes ) were incorporated into liposomes and

added to cultures of chick embryo sensory neurons extending on a laminin substratum. One hour later, a mean of 70% of the growth cones had collapsed, changing from a typical spread morphology (with filopodia and lamellipodia) to a needle-like morphology. However, extracts of AD-der.ived cerebral cortex (3 separate brain samples, 3-22 hours post-mortem delay, also of mixed regional origin; diagnosis of AD confirmed by silver- staining for plaques and tangles), with the same overall protein content, gave a mean figure of only 25% growth cone collapse, representing a highly significant reduction over control values.

Further work suggests that the diminution varies between different cortical regions, correlating with the local variation in disease activity as assessed by the density of amyloid plaques. Thus, in 2 AD brains examined in this way, regions of cerebral cortex with little disease activity show normal levels of collapse-inducing activity, while those with increased disease activity show reduced collapse activity.

These observations are consistent with the idea that the pathogenesis of the disease may involve a failure of normal control of nerve sprouting (Geddes and Cotman, 1991). Nerve sprouting has been noted in a number of studies of AD (for example: Scheibel and Tomiyasu, 1978; Probst et al., 1983; Ihara, 1988; McKee et al., 1989; Geddes and Cotman, 1991; Masliah et al., 1991), and could be secondary to partial deaffe antation of neurons (resulting from a primary process of cell death). An alternative, however, is that it might result from a primary reduction (from whatever cause) in the levels of expression of the molecule suppressing nerve growth, for

example on astrocytes surrounding synaptic terminals, leading to aberrant nerve growth and alteration of normal synaptic connectivity. Associated with such a change, neurons could lose normal trophic support, leading secondarily to cell death. Dementia could then result from sprouting and/or cell death. In sum, while the loss of growth cone collapse-inducing activity in AD brain may be secondary to cell death, it is equally plausible that it lies upstream of cell death in the chain of events that leads to full AD pathology.

Finally, it should be noted that the implication of the latter hypothesis is that in the possible treatment of AD, in contrast to the restoration of axon regeneration following acute CNS lesions, it may be appropriate to increase rather than inhibit the levels of nerve growth- inhibitory activity. This could be achieved, for example, by the local administration of appropriate growth factors and cytokines (Johnson et al., 1991).

One approach to stimulate nerve regeneration involves use of antibodies (monoclonal or polyclonal) to the glycoprotein(s) : by binding to the glycoprotein(s) , • biochemical activity may be affected.

The invention thus provides a method of inhibiting neurodegeneration or promoting neuroregeneration comprising administering to a patient a substance that interferes with the functioning of glycoprotein of the invention.

According to a further aspect of the invention there is provided an antibody, preferably a monoclonal antibody, to glycoprotein of the invention.

The invention also provides an antibody preferably a monoclonal antibody to the glycoprotein of the invention, which antibody is for use in a method of treatment of the human or animal body. Preferably, the antibody is for use in treating neurodegenerative diseases, or in promoting neuroregeneration.

In another aspect the invention provides a method of inhibiting neurodegeneration or promoting neuroregeneration, comprising administering to a patient antibodies to glycoprotein of the invention.

The invention finds applicablility in the treatment of conditions and diseases such as acute brain injury, Alzheimer's Disease or motor neurone disease.

Two logical developments from the invention concern the purification of the receptor(s) for the biologically- active glycoprotein, and the cloning of the genes for both the glycoprotein and its receptor(s). The identification of the receptor(s) would provide another entry point for molecular manipulation of the system for the clinical purposes described herein, for example using suitable blocking antibodies or chemical agonists and antagonists. Cloning the genes for both the glycoprotein and its receptor(s), via partial sequence determination of the purified molecules, would enable the production of large quantities of the molecules by insertion of appropriate cDNA in suitable expression vectors. In turn, this would provide a more efficient means of production of antigen for the generation of further antibodies. Local inactivation of gene function in brain lesions, by direct surgical application of anti-sense material or targeted gene disruption, would be a further way of preventing

expression of the biological activity in lesion areas. Further, antibodies would provide an additional means of purifying the growth cone collapse-inducing molecules and related receptors.

An additional use for antibodies and cDNA probes would be for diagnostic purposes, where extracts of mammalian brain biopsy could be subjected to immunoassays (such as ELISA, enzyme-linked immunoadsorbent assay) and/or RNA assay (such as Northern blot analysis or RNAase protection).

There now follows, by way of illustration, a description of embodiments of the invention including a method of obtaining the biologically active material and a method of obtaining antibodies thereto, reference being made to the accompanying figures in which:

Figure 1 is a photograph of an acrylamide gel, illustrating growth cone collapse activity isolated from chicken grey matter by affinity chromatography;

Figure 2 shows histograms representing the level of growth cone collapse activity obtained from various extracts of grey matter;

Figure 3 shows histograms representing the level of growth cone collapse activity of various products of dye chromatography; and

Figure 4 shows histograms representing the level of growth cone collapse activity obtained from extracts grey matter of various different species.

Purification Protocol

Freshly dissected grey matter, e.g. mammalian (human, sheep, rat, mouse etc) or avian (chicken) is homogenized in 2% (w/v) CHAPS (Sigma) (CHAPS is a zwitterionic detergent, 3-[ (3-cholamidopropyl) dimethyl ammonio]-l- propanesulphonate) in phosphate-buffered saline, pH 7.4, at 4°C, centifuged at 100,000g for 1 hour at 4°C, and the supernatant fluid harvested as source material. This is dialyzed against excess 0.2% CHAPS, 25mM phosphate, pH 6.65; any precipitate that fcrms is centrifuged at 100,000g for 1 hour at 4°C. The resulting supernatant fluid is adsorbed onto Mimetic Green (an affinity adsorbent marketed by Affinity Chromatography Ltd). Biological activity in this latter material is monitored by a growth cone collapse assay adapted from Raper and Kapfhammer (1990) (see Coc et al 1991). The column is washed extensively with 0.2% CHAPS, 25mM phosphate, pH 6.65, the eluates also being monitored by optical density at 280nm. V7e find that all the collapse activity is adsorbed when equal volumes of gel and supernatant fluid are interacted.

Once the optical density has returned to baseline the column is elu:ed with IM KCl, 50mM phosphate, pH 8.0, 0.2% CHAPS, whereupon a second peak of protein containing collapse activity is recovered.

The active material present in the second peak can be worked up by one of two protocols: a) use of further dye-affinity chromatography, or b) by hydrophobic interaction chromatography.

In protocol (a) material from the second peak is dialyzed against excess 0.2% CHAPS, 25mM phosphate, pH 6.6 It is

then passed successively through immobilized Brown-10 (a procion dye obtained in immobilized form from Sigma Chemical Co; the batches used by us normally have more that 5mgm dye per ml of gel) followed by Mimetic Blue-2 marketed by Affinity Chromatography Ltd. This part of the precedure is designed to remove contaminating proteins, although in practice using partially purified material from Mimetic Green results in considerable loss of biological activity. Finally, the filtrates from these columns are adsorbed onto immobilized Green-19 (marketed in immobilized form by Sigma Chemical Co.), washed extensively with 0.2% CHAPS 25mM phosphate, pH 6.65, followed by elution with 0.5M KCl, 50 mM phosphate, pH 8.0, 0.2% CHAPS followed by 0.7M KCl, 50mM phosphate, pH 8.0, 0.2% CHAPS, followed by IM KCl, 50mM phosphate, pH 8.0, 0.2% CHAPS. The bulk of the biological activity is recovered in the latter elution. After extensive dialysis against 0.2% CHAPS 25mM phosphate (pH 7.0), the material is fractioned on a non-denaturing slab gel (10% acrylamide; Ha es and Rickwood, 1981).

Molecules separated by this process can be detected by the use of silver-staining reagent for proteins. By way of example, results with avian material are illustrated in Figure 1. This Figure illustrates how biologically-active material isolated from Mimetic Green 19 may be fractionated on a non-denaturing gel (10% acrylamide) and detected by silver-staining reagent. The 2 tracks on the left on the figure are standard marker proteins, ovalbumin and BSA; the track on the right is material which has been purified from Green 19 affinity column chromatography. The biologically-active material runs principally as a single band, which in the case of chicken grey matter runs in the same position as BSA monomer as shown, and from

which it may be electroeluted and shown to possess growth cone collapse-inducing activity. This material shows 2 components of M 62,000 and 75,000 as determind by SDS- PAGE. The sample used in Figure 1 corresponds to approximtely lug protein.

An alternative procedure is to adsorb the active material from Mimetic Green directly to a suitable hydrophobic matrix. From a survey of propyl, butyl, hexyl, octyl and phenyl ligands we have evidence that activity is best adsorbed to the propyl and butyl matrices. Biological activity may be recovered by eluting these matrices with 0.2% CHAPS, 25mM phosphate, pH 7.0. Such material, when subjected to non-denaturing gel electrophoresis (see above), results in a similar co-migration of the active material with the BSA marker.

This protocol has been applied to chicken and rat brain (forebrain, midbrain and cerebellum). Our laboratory has also detected biological activity in 2% CHAPS - extracts of human brain (post-mortem specimens). Activity from this latter source is also adsorbed reversibly by Mimetic Green, and an example of such an experiment is shown ^' in Figure 2.

The histograms of Figure 2 illustrate the fractionation of collapse-inducing activity from extracts of human grey matter on a 5ml column of Mimetic Green. Each histogram represents the level of collapse activity present in the column fractions, which correspond to 0.2 column volumes. The histograms marked RG through to Di are various controls. RG represents the level of collapse found in

CHAPS extracts of rat grey matter, and HG the same found in extracts of human grey matter. PL and PBS are control

values, showing that liposomes (PL) used as a vehicle for the collapse-inducing molecules have themselves no effect on growth cones, and that saline alone (PBS) is likewise without effect. The dye affinity purification protocol requires that the extracts are dialysed against excess 0.2% CHAPS, 25mM phosphate, pH 6.65, before addition to the Mimetic Green. The histogram Di shows the extracts of human grey matter (HG) so treated retain biological activity. Histograms Fl, F5 and F8 show the loading of dialysed material onto Mimetic Green. It will be seen that after 8x0.2 column volumes the Mimetic Green column is fully saturated. The column may then be thoroughly washed and, as seen by comparison of l with 22, any excess activity can be washed off. Note that by washing fraction 22, collapse activity has reached control levels (cf. PL and PBS). The adsorbed biological activity is then elutable using IM KCl, 50mM phosphate (pH 8.0), 0.2% CHAPS, as shown by histograms El to E4.

In addition we have shown the Mimetic Blue 2 may also be used to remove contaminating proteins from biologically- active material eluted from Mimetic Green (see Figure 3). In this experiment a 1ml Mimetic blue 2 column was used and the fractions represented by individual histograms of Figure 3 correspond to 0.4 column volumes. For this experiment the 1st KCl elution fraction from the experiment illustrated in Figure 2 (El) has been brought back to 0.2% CHAPS, 25 mM phosphate (pH 6.65) by dialysis. The histogram shows that most of the collapse-inducing activity passes straight through the column (MBF) and in the 1st washing (Wl) . Subsequent elution with IM KCl, 50 mM phosphate (pH 8.0), 0.2% CHAPS (El, E2 and E3) shows that little if any activity is adsorbed to Mimetic Blue. For comparison, rat grey matter extract (RG) is given as a

positive control, along with the 2nd elution fraction from the Mimetic Green column (MGE2). Negative controls for plain liposomes (PL) and PBS are also shown.

See also the Table. The table illustrates that Mimetic Green can be used in a positive mode to adsorb human activity, and Mimetic Blue 2 in a negative mode to remove non-active protein. It should be noted that, in contrast to avian material, any haemoglobin present in the extracts of human grey matter is not adsorbed to Mimetic Green, possibly obviating the need to use Brown 10.

The biological activity obtained from chicken grey matter at least may be adsorbed by interaction with immobilized peanut agglutinin (PNA), a property associated with the collapse-inducing molecules isolated from chicken somites (Davies et al. , 1990). This is prima facie evidence that the molecules encountered here are glycoproteins carrying O-linked carbohydrate groups (Cook and Keynes, 1992. Relations between mesodermal and neural segmentation. - in Formation and Differentiation of Early Embryonic Mesoderm; eds. R Bellairs & J W Lash, Plenum Press.). In addition, biological activity is lost when the material from somites is treated with a mixture of neuraminidase and O- glycanase, which strengthens the view that glycoproteins bearing O-linked residues are involved. We have also treated the somite material with N-glycanase (under conditions where any endogenous protease activity is carefully inhibited) without any loss of biological activity. This experiment does not, however, rule out the presence of N-linked carbohydrate in the molecule.

Extracts from a variety of species have been prepared under standard conditions (ie. lgm wet weight grey matter

to 2.5 ml homogenisation medium) and tested for their ability to induce collapse of chicken sensory axon growth cones extending on a laminin substratum. The results are shown in Figure 4. Controls for plain liposomes (PL) and phosphate buffered saline (PBS) are shown, along with extracts from carp (CB), chicken (CG), rat (RG), sheep (SG) and human (HG) grey matter. In addition similar extracts of pig grey matter also induce growth cone collapse, and bind to Mimetic Green.

It is to be noted that mammalian activity causes avian (chicken) growth cones to collapse, and the latter source of growth cones is used as the regular test material. Avian activity also causes mammalian (rat) growth cones to collapse. These observations strongly suggest that a common molecular mechanism is shared by avian and mammalian species.

Antibody Preparation

Glycoproteins purified as described above on preparative gels are either electroeluted or directly excised by razor blade and act as antigen. The antigen, in sterile phosphate-buffered saline (0.145M NaCl, lOmM phosphate, pH 7.4) is made into an emulsion with Freund's complete adjuvant using a Berlin adaptor. Emulsion is admistered at several sites across the back of the rabbit by subcutaneous injection, as well as by intramuscular injection into the thigh. Booster injections are given every 10 days using the above procedure, with the major exception that Freund's incomplete adjuvant is used. As an alternative vehicle to Freund's adjuvant we use Hunter's TiterMax, available from Stratech Scientific Limited.

The animals are bled via the marginal ear vein, the blood allowed to clot at room temperature, and the clot allowed to contract overnight at 4°C. Serum is harvested by low speed centrifugation. Serum is used either directly, or an IgG fraction is isolated by immoblized protein A (alternatively Spectragel is used. This is a low molecular weight non-protein, non-carbohydrate synthetic affinity ligand, marketed by Spectrum Medical Industries Inc., Houston, Texas, U.S.A. and has the advantage that the IgG preparation is not contaminated with protein A).

The presence of antibodies is detected either by Western blotting against brain proteins separated by SDS-PAGE or by immunohistochemistry of brain sections. In addi ion, immobilized immune IgG is used to test whether collapse- inducing activity can be removed from extracts of mammalian grey matter.

In practice, monoclonal antibodies would be produced in conventional manner.

In use antibodies to the glycoprotein antigens can be administered directly to the site of the neuronal lesion or into the cerebrospinal fluid via a cerebral ventricle. This could be accomplished by surgical infusion following craniotomy (the blood brain barrier may preclude systemic administration of immunoglobulins) . The induction of transient leakiness of the blood brain barrier, for example by infusion of a hyperosmotic solution, with simultaneous systemic administration of antibodies, may provide an alternative therapeutic strategy. Conditions most readily suited to such procedures comprise: 1) acute mechanical trauma with axotomy (for example, penetrating

brain injury and spinal cord transection) ; 2) cerebrovascular accident due to haemorrhage, thrombosis or embolism; 3) as an adjunct therapy to the treatment of brain and spinal cord damage due to bacterial or viral infection, and tumour or abscess expansion.

The antibodies can also be used in treatment of diseases involving chronic neurodegeneration such as Alzheimer's Disease and motor neurone disease.

The glycoproteins in purified form can also be used in treating humans, preferably in treating diseases caused by excessive neuronal growth.

TABLE

Purification of growth cone collapse-inducing activity from CHAPS extracts of human grey matter by dye affinity chromatography.

Fraction Growth cone collapse Protein (%) (mg/ml)

Extract of 100 12.4 human grey (HG)

1st eluate 85 2.2 from Mimetic Green

(El)

Filtrate from 30 0.2 Mimetic Blue 2 (MBF)

References

G.M.W. Cook, J.A. Davies & R.J. Keynes (1991). Growth cone collapse: a simple assay for monitoring cell-cell repulsion. - in Cell Signalling: Experimental Strategies, pp. 359-366; eds. E. Reid, G.M.W. Cook & J.P. Luzio, Royal Society of Chemistry.

G.M.W. Cook & R.J. Keynes (1992). Relations between mesodermal and neural segmentation. - in Formation and Differentiation of Early Embryonic Mesoderm; eds. R. Bellairs & J.W. Lash, Plenum Press.

J.A. Davies, G.M.W. Cook, CD. Stern & R.J. Keynes (1990). Isolation from chick somites of a glycoprotein fraction that causes collapse of dorsal root ganglion growth cones. Neuron 4, 11-20.

J.W. Geddes and C.W. Cotman, (1991). Plasticity in Alzheimer's disease: too much or not enough? Neurobiol.Ageing 12, 330-333.

B.D. Hames and D. Rickwood, (1981). Gel electrophoresis of proteins: a practical approach. IRL Press Ltd., Oxford and Washington DC.

Y. Ihara, (1988). Massive somatodendritic sprouting of cortical neurons in Alzheimer's disease. Brain Res. 459, 138-144.

A.R. Johnson, J.W. Fawcett, R.J. Keynes and G.M.W. Cook, (1991). Induction of growth cone collapsing activity in cultured astrocytes by FGF and interleukin 1. Society for

Neuroscience Abstracts 17, 598.9.

E. Masliah, M. Mallory, L. Hansen, M. Alford, T. Albright, R. DeTeresa, R. Terry, J. Baudier and T. Saitoh, (1991). Patterns of aberrant sprouting in Alzheimer's disease. Neuron 6, 729-739.

A.C. McKee, N.W. Kowall, and K.S. Kosik, (1989). Microtubular reorganisation and dendritic growth response in Alzheimer's disease. Ann. Neurol. 26, 652-659.

A. Probst, V. Basler, B. Bron and J. Ulrich, (1983). Neuritic plaques in senile dementia of Alzheimer type: a Golgi analysis in the hippocampal region. Brain Res. 268, 249-254.

J.A. Raper and JP Kapfhammer, (1990). The enrichment of a neuronal growth cone collapsing activity from embryonic chick brain. Neuron 4, 21-29.

A.B. Scheibel and U. Tomiyasu, (1978). Dendritic sprounting in Alzheimer's presenile dementia. Exp. Neurol. 60, 1-8.