MXPA99004339A - Purification of neurotrophins - Google Patents

Abstract

Methods are provided for large scale purification of neurotrophins, including mature NGF, suitable for clinical use. The methods provide means to separate neurotrophins from various less desirable misprocessed, misfolded, size, glycosylated, or charge forms. Compositions of neurotrophins, including mature NGF, substantially free of these variants are also provided.

Description

PURIFICATION OF NEUROTROPHINS FIELD OF THE INVENTION This invention relates to an improved method for purifying neurotrophins, mainly those in the NGF family, more particularly nerve growth factor (NGF) and neurotrophin 4/5 (NT-4/5), and neurotrophin 3 (NT 3) of the variants, impurities, and contaminants associated with them, particularly when they are produced by fermentation of bacterial or mammalian cells.

BACKGROUND OF THE INVENTION The production of large quantities of relatively pure, biologically active polypeptides and proteins is economically important for the manufacture of pharmaceutical formulations, enzymes, and other specialty chemicals, for humans and animals. For the production of many proteins, recombinant DNA techniques have become the method of choice due to the large amounts of proteins REF .: 30099 exogenous that can be expressed in mammalian host cells, and in bacteria, and in other host cells. The main structure of a mammalian NGF (mouse NGF) was first elucidated by Angeletti and Bradshaw, Proc. Na tl. Acad. Aci. USA 68: 2417 (1971). The primary structure of its precursor, pre-pro-NGF, has been deduced from the nucleotide sequence of the mouse NGF cDNA (Scott et al, Na t ure 302: 538 (1983); Ullrich et al. Na t ure 303: 821 ( 1983)). The homologous human NGF gene (hNGF) has also been identified (Ullrich, Symp. On Quan, Bi ol., Col. Spring Harbor 48: 435 (1983); US Patent No. 5,288,622, issued on February 22, 1994, which is incorporated by reference herein, its homology to the mouse NGF is approximately 90% and 87%, on the levels of the amino acid and nucleotide sequence, respectively, due to the shortage of naturally occurring human NGF, Since then, it has been prepared from natural sources in sufficient quantities to characterize itself biochemically in fine detail.The additional neurotrophic factors related to NGF have been detected since then, including neurotrophic factor derived from the brain. BDNF) (Leibroc et al., Na t ure, 341: 149-152 (1989)), neurotrophin 3 (NT-3) (Kaisho et al., FEBS Le tt., 266: 187 (1990)), Maisonpierre et al. Sci ence, 247: 1446 (1990); Rosenthal et al. Ne uron 4: 767 (1990)), and neurotrophin 4/5 (NT-4/5) (Berkmeier et al., Neuron, 7: 857-866 (1991)). GDNF, a distant member of the TGF-β superfamily, and neurturin ("NTN") are two potent, recently identified, structurally related, central nervous system and sympathetic sensorial survival factors (Lin et al., Sci en ce 260: 1130-1132 (1993), Henderson et al, Sci en 266: 1062-1064 (1994), Buj-Bello et al., Neuron 15: 821-828 (1995); Kotzbauer et al., Na t ure 384: 467-470 (nineteen ninety six) ) . The production of recombinant protein involves the transfection of host cells with the DNA encoding the protein, and the development of the cells under conditions that favor the expression of the recombinant protein. The prokaryote E. coli has been a favored host because it can be elaborated to produce recombinant proteins with high yields at low cost. There are numerous North American patents on the general bacterial expression of DNA encoding proteins, including US Pat. No. 4,565,785 on a recombinant DNA molecule comprising a bacterial gene for an extracellular or periplasmic carrier protein and the non-bacterial gene; US Patent No. 4,673,641 on the co-production of a foreign polypeptide with an aggregate-forming polypeptide; U.S. Patent No. 4,738,921 on an expression vector with a trp promoter / operator and the trp LE fusion with a polypeptide such as IGF-I; U.S. Patent No. 4,795,706 on the expression of control sequences to be included with a foreign protein; and US Patent No. 4,710,473 on the plasmids of AD? circular, specific such as those that code for IGF-I. Genetically engineered pharmaceutical bioproducts are typically purified from a supernatant containing a variety of various host cell contaminants. NGF, in particular, has been notoriously purified to a varying degree, with varying degrees of effort and success using a number of different methods. See, for example, Longo et al., IBRO Handbook, vol. 12, pp. 3-30 (1989); Patent Nort eameri cana No. 5,082,774, which describes the production of? GF CHO cells; Bruce and Heinrich (Ne urobi o.Aging 10: 89-94 (1989); Schmelzer et al. J. Neurochem. 59: 1675-1683 (1992); Burton et al., J. Neurochem. 59: 1937-1945 (1992) These efforts have been mainly at the laboratory scale, however, the preparative isolation of the recombinant human GF which results in pharmaceutical purity and high yield, essentially free of variants, has eluded the technique. the technique for an efficient protocol for selectively separating neurotrophins, particularly? GF and the? GF family of neurotrophins, their variants and other molecules, and other polypeptides with high pl. The large-scale purification process of neurotrophins should be Applicable to the starting material from various sources, including the fermentation broth, the bacterial or lysed mammalian cells, to provide for the clinical needs.In addition, since the present inventors have discovered In previously unknown, difficult to separate neurotrophin variants, for example NGF variants, the methods presented herein are particularly useful for providing commercially useful amounts of recombinant neurotrophins, including human NGF (rhNGF), rhNT-3 and rhNT-4 / 5 and the desirable genetically engineered mutants thereof, which are substantially free of undesirable variants. These and other objects of the invention will now be apparent to one of ordinary skill in the art.

BRIEF DESCRIPTION OF THE TECHNIQUE In one embodiment of the invention, there is provided a process for purifying a neurotrophin, particularly one in the NGF family, including NGF, NT-3, NT-4/5, and BDNF which share recognition by a highly homologous family of receptors (trks), preferably rhNGF, rNT-3, rhNT-4/5, rhBDNF or the desirable forms, genetically engineered, thereof, by the use of hydrophobic interaction chromatography (HIC). In view of the discovery by the present inventors of certain undesirable neurotrophin variants arising from the recombinant production of a neurotrophin, as reported herein, the use of HIC can separate chemically different or even misfolded forms of a neurotrophin, of intact, desired neurotrophin, correctly folded. The variants that can be removed are those that differ from correctly folded neurotrophin, mature in hydrophobicity, including partially processed precursor sequences, mature glycosylated forms and containing the precursor (when present from the culture of eukaryotic cells), and misfolded and partially folded variants (in general bacterial cell culture when folding steps are used in vi tro). For example, HIC is particularly useful for eliminating the partially processed precursor sequences of NGF, the glycosylated species of NGF and the precursor (when present from the culture of eukaryotic cells), and the misfolded and partially folded variants of the culture of bacterial cells and fold steps in vi tro) from mixtures of mature NGF. NGF has an N-linked glycosylation site in Asn45. In the case of rhNT4 / 5 re-folded, expressed in bacteria, HIC separates NT-4/5 correctly folded from incorrectly folded forms. As a result of the process described herein, neurotrophin is essentially free of these variants. For the purification of the neurotrophin, preferably the functional group of the HIC resin is a phenyl group, while the octyl and butyl groups may be useful. Particularly preferred embodiments include HIC Phenyl Toyopearl resins, Fast Flow Substitution Phenyl Sepharose, TSK-Phenyl 5PW, or the like. In yet another embodiment, a process for the purification of a neurotrophin, particularly one of the NGF family, preferably rhNGF, rNT-3, rhNT-4/5 or the desirable forms, engineered by the same, is provided by the use of preparative cation exchange chromatography, which separates charge-modified variants, such as the oxidized, isoasp and deamidated forms of the mature neurotrophin. Particularly preferred embodiments use High Performance SP-Sepharose, Fractogel EMD S03, or the polyaspartic acid resin, of which PolyCAT A is particularly preferred. More preferably, the high-performance SP-Sepharose or Fractogel EMD S03 resins are used on a large scale. In yet another embodiment of the invention, HIC and cation exchange chromatography are used to prepare a desired neurotrophin composition, for example recombinant mature NGF, preferably human NGF, which is substantially homogeneous, for example, substantially free of the variants of the invention. process and loading, for example misfolded variants and chemical variants, and is also substantially pure with respect to protein content. In one embodiment of this invention, there is provided an improved process for the separation of neurotrophins, particularly those of the NGF family, preferably recombinant NGF, NT-3, NT-4/5, and their desirable forms, engineered by the related undesirable variants, for example, fermentation variants, cleaved by protease, glycosylation variants, misfolded variants, by means of reverse phase liquid chromatography. More preferably, NGF is the homodimeric form 120/120 or 118/118. As a result of the process described herein, neurotrophin is, more preferably, essentially free of variants. In yet another embodiment, a process for purifying neurotrophins, particularly those of the NGF family, of the related variants using elution conditions involving physiological pH is provided. In yet another embodiment a process for the purification of a neurotrophin is provided, which results in considerable improvement in its homogeneity. In still another embodiment, the invention provides a process for separating neurotrophins from the NGF family of variants thereof comprising: a) loading a buffer containing the neurotrophin and its variant at a pH of about 5 to about 8 a column of hydrophobic interaction chromatography; b) washing the column, c) eluting the neurotrophin with a buffer at a pH of about 5 to 8; d) loading the eluent containing neurotrophin onto a cation exchange chromatography column at a pH of about 5 to 8; and e) the elution of the neurotrophin from the column, with a buffer to a saline gradient at a pH of about 5 to 8 preferably pH 6. The neurotrophin is more preferably rhNGF. In one embodiment of the invention, a chromatography step on silica gel is provided which efficiently removes the host cell proteins from the neurotrophin fraction, which is preferably a fraction of NGF. In one embodiment of the invention, a process step is provided in which the 120 amino acid NGF is subjected to trypsin-like protease treatment to selectively remove the terminal RA dipeptide from the C-terminal VRRA, to produce the species 118. Se prefers an immobilized trypsin column. The invention also relates to the neurotrophin composition and to the formulation prepared by the processes of the invention, and to the uses for the composition and the formulation. A neurotrophin composition is provided which is substantially homogeneous, for example, substantially free of process variants and loading variants, for example, misfolded variants and chemical variants, and is also substantially pure with respect to the content of protein. Preferably, mature human NGF, mature human or rat NT-3, and mature NT-4/5 are provided in this form. In a preferred embodiment, NGF is species 120, and more preferably form 118, more preferably as a homodimer, for example, 118/118.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 describes a chromatogram of HP in SP-Sepharose. A mixture containing NGF from a fermentation of 12 kL after HIC chromatography was loaded on a High Resolution resin of SP-Sepharose (column dimension 1.0 x 35 cm, a bed volume of 27.5 ml) of resin 25 omnifit SPSHP. Buffer A was 0.2 M NaCl, 20 mM succinate, pH 6.0 (23 ms). Buffer B was 0.7 M NaCl, 20 mM Succinate, pH 6.0 (63 ms). The column was first equilibrated in buffer A. The combined HIC from 345 ml to 0.24 mg / ml was adjusted to 25 mM = uccinate, pH 6.0 (17 ms) of which 362 ml were loaded giving a resin loading of 3 mg / ml; 82.5 mg of NGF were loaded. The loading speed was 40 cm per hour. The NGF was eluted using a column volume gradient 22 from 30% to 80% buffer B (0.35 to 0.60 M NaCl, 37 to 57 ms). The elution speed was 60 cm per hour. The absorbance units (280 nm) and mS / cm were plotted against the fraction numbers and the elution volume in me. Fractions containing NGF were determined and combined. In this case, fractions 43 through 60 were combined to obtain a sample of 99 ml of NGF at 0.38 mg / ml, giving approximately a recovery of 46%.

Figure 2 depicts a SP-NPR HPLC (HPIEX) cation exchange analysis of Phenyl Sepharose Fast Flow combined and the HP combined SP-Sepharose. The chromatogram is marked for each one.

Figure 3 describes an analysis of C4 RPHPLC of selected fractions (main pool, front edge and trailing edge of the main peak pool) from the HP chromatography step of SP-Sepharose. The three signals are overlapped and thus marked. As can be seen, the main type (containing mature NGF) is separated from several minor peaks containing the variant NGFs.

Figure 4 describes the sequence of human prepro-NGF (SEQ ID NO: 1). Indicated in the figure is the first amino acid of mature NGF (position 1) and the last amino acid of the form 120 NGF (position 120). A preferred amino acid sequence of NGF is that of the mature form 118 (from position 1 to position 118). The variant forms are also shown; mature 120 (position 1 to position 120); mature 117 (position 1 to 117); R120 (position -1 to 120); and the sites for the other bad processing that occurs at the N- and C-terminal ends, including the mature variants 114 (amino acids 1 to 114), 115 (amino acids 1 to 115) and 117 (amino acids 1 to 117). A major misprocessed variant, proteolytically poorly processed, has the N-terminal cleavage between the amino acids R (-39) and S (-38) in the pro-sequence of NGF. The likely Met start is underlined. Other N-terminal variants include truncated forms of NGF, with the most common being cleavage occurring between amino acids H8 and R9 and between R9 and GlO.

Figure 5 depicts the amino acid sequences of human neurotrophins NGF (SEQ ID NO: 2), mouse NGF (SEQ ID NO: 3), BDNFG (SEQ ID NO: 4), NT-3 (SEQ ID NO: 5) ) and NT-4/5 (SEQ ID NO: 6). The boxed regions indicate the homologous cysteine-containing regions involved in the cysteine knot portion (De Young et al., Pro tin Sci. 5 (8): 1554-66 (1996)).

Figure 6 depicts the chromatography pattern of rhNT-4/5 on the Fast Flow resin column of DEAE-Sepharose (DEFF). The chromatogram marked "NT45DE1: 1_UV" is a measurement of UV absorbance at 280 nm. The chromatogram labeled "NT45DE1: l_Condl" is a conductivity measurement of the elution fractions. The fractions containing neurotrophin that were combined are indicated by the horizontal arrow marked "Combined".

Figure 7 describes the chromatography pattern of rhNT-4/5 on the Fast Flow resin column SP-Sepharose. The chromatogram labeled "NT45SFF1: 1_UV" is a measurement of UV absorbance at 280 nm. The chromatogram labeled "NT45SFF1: l_Condl" is a measure of the conductivity of the elution fractions. The fractions containing neurotrophin that were combined are indicated by the horizontal arrow marked "Combined".

Figure 8 depicts a preparative C4-RP-HPLC chromatography pattern, typical of rhNT-4/5 under the conditions described in the text. The absorbance at 280 nm was checked periodically. The fractions containing neurotrophin that were combined are marked with a horizontal arrow marked "Combined".

Figure 9 provides the analytical HPLC chromatography pattern of the NT4 / 5 samples verified during refolding at the indicated times. The conditions of the column are described in the text. "NT-4/5 Std" indicates the elution pattern of an intact NT-4/5, correctly folded, used as a standard. The pattern marked "SSFF (0.5 m)" describes the analysis of? T-4/5 eluted with 0.5 M sodium chloride from the S-Sepharose Rapid Flow column before refolding. As? T-4/5 retracts, the elution pattern resembles that of the standard.

Figure 10 depicts a chromatogram showing the separation of the correctly folded, intact, T-4/5 from misfolded variants on a hydrophobic interaction chromatography column, the Phenyl Toyopearl 650M column. The misfolded variants are less hydrophobic than the correctly-folded? T-4/5, and elute in side-to-side flow, while the misfolded variants (Peak B) that are more hydrophobic elute at a higher organic solvent concentration that necessary to elute the? T-4/5 correctly folded (Peak A).

Figure 11 describes the elution pattern of NT-4/5 and the variants of a cation exchange resin, SP-Sepharose. The absorbance at 280 nm was checked periodically. Peak A contains carbamylated and cut variants, while Peak B contains NT-4/5 correctly folded, intact. The fractions containing? T-4/5 that were combined are indicated by the horizontal arrow marked "Combined".

Figure 12 depicts a preparative poly-CAT A resin chromatography pattern, typical of rh? T-4/5 under the conditions described in the text.

Figure 13 describes a 16% SDS-PAGE analysis (Tris-glycine system, pre-emptied, from? Ovex, Inc., San Diego, California) under reducing conditions to evaluate the purity and homogeneity of the samples taken from the indicated steps of the purification process of rh? T4 / 5 described in the text. The gel was stained with Coomassie-R250 to detect the protein (Andrews, Electrophoresis, Oxford University Press: New Yark, 1986). The band labeled "Load DE" contains a sample of the PEI mixture that was loaded onto the Fast Flow column of DE-Sepharose; the band marked "DE FT" contains a sample of the flow through the Fast Flow column of DE-Sepharose; the "Combined S" band contains a sample from the combined fractions containing NT-4/5 eluted from the S-Sepharose Rapid Flow resin column, before refolding; the "Combined Refolding" band contains a sample of the combined after the refolding was completed; the "Combined C4" band contains a sample of the combined fractions after the RP RP-HPLC preparative; and the "Combined PolyCat A" band contains a sample from the combined NT-4/5 fractions from the PolyCat A HPLC column.

Figure 14 depicts the UV absorbance pattern of the fractions from S-Sepharose Rapid Flow chromatography of a mixture containing sulfonyl rhNT3, produced by bacteria.

Figure 15 depicts Macroprep High S cation exchange chromatography of a mixture containing rh.NT-3 forms folded in vi tro. The resin was purchased from Biorad. The dimensions of the column were 9 x 9 cm. A combined SSFF of 700 μm containing the refolding (after 36 hours of refolding), pH 6.8, was loaded onto the Macroprep column at a flow of approximately 310 ml / minute. The conditions are given in the text.

Figure 16 describes a High-Pitch Sepharose Rapid Flow chromatography(hydrophobic interaction chromatography) of refolded rhNT-3, purified by cation exchange, to eliminate misfolded variants.

Figure 17 describes a chromatography of High Resolution of SP-Sepharose from the combination of HIC-rhNT-3.

DETAILED DESCRIPTION Definitions As used herein, "neurotrophin" refers to a neurotrophin, preferably a neurotrophin of the NGF family, including NGF, NT-3, NT-4/5 and BDNF, of any species, including murine, bovine, ovine, porcine, equine, avian, and preferably human, in a native sequence or in a genetically engineered form, and from any source, whether natural, synthetic or recombinantly produced.For example, "NGF" refers to the nervous growth factor of any species, including murine, bovine, ovine, porcine, equine, avian, and preferably human, in a native sequence or in a variant form engineered by genetic engineering, and from any source, whether natural, synthetic, or recombinantly produced. Neurotrophin is recombinantly produced In a preferred method, neurotrophin is cloned and its DNA expressed, for example, in mammalian cells, or in bacterial cells. Ostrados in the present can be applied to neurotrophins GDNF and neurturin. Preferred for human use is mature NGF, of native, human sequence, more preferably a sequence of 120 amino acids and still more preferably a sequence of 118 amino acids. More preferably, this native sequence NGF is recombinantly produced. The preferred amino acid sequence for human pre-pro-NGF and mature human NGF are provided by U.S. Patent No. 5,288,622, which is specifically incorporated by reference herein. The 120 amino acid form, without the additional post-translational modifications, is a preferred form in the homodimeric form (eg, 120/120). Still more preferred is form 118, with no further post-translational modifications, particularly as a homodimer (eg, 118/118). By "substantially pure" is meant a degree of purity of total neurotrophin, eg, NGF, to the total protein where there is at least 70% neurotrophin, more preferably at least 80%, and even more preferably increasing to at least 90% , 95% or 99%. A particularly preferred purity is at least 95%. By "essentially pure" it is meant that the composition is at least 90% or more pure for the desired neurotrophin. By "substantially free of neurotrophin variant" is meant a composition at which the percent of the desired neurotrophin species to the total neurotrophin (including less desirable neurotrophin species) is at least 70% the desired neurotrophin species, more preferably at least 80%, and at least more preferably increasing to at least 90%, 93%, 95% or 99%. By "essentially free" it is meant that the composition contains at least 90% or more of the desired neurotrophin. A particularly preferred level is at least 95% of the desired neurotrophin, for example, the intact, correctly folded rhNGF 118/118 species. The other undesirable species or forms may be poorly processed forms or chemical variants, for example variants with altered charge, resulting from the fermentation or purification process, or preferably all of the above, as described herein. For example, when NGF is folded back after bacterial synthesis the "species" or "variants" may include the misfolded or partially folded forms. By "badly folded" variant is meant a variant of neurotrophin that differs from neurotrophin by the pairing of its cysteine residues or by the particular cysteine residues which are free or blocked. Misfolded variants may also have the same cysteine pairing as neurotrophin, but have a different three-dimensional conformation resulting from misfolding. By "chemical" variant is meant a variant that differs chemically from neurotrophin, for example by having an altered charge, by carbamylation, deamidation, oxidation, glycosylation, or proteolytic cleavage. The buffers for the aspect of the column of this invention generally have a pH in the range of about 5 to 8. Buffers that will control pH within this range include, for example, citrate, succinate, phosphate, MES, ADA, BIS-TRI S-propane, PIPES, ACES, imidazole, diethyl malonic acid, MOPS buffer, MOPSO, TES, TRIS such as TRIS-HC1, HEPJES, HEPPS, TRICINE, glycinamide, BICINE, glycine lgl icine, and borate buffers . A preferred shock absorber is a MOPSO shock absorber. As used herein "alcohols" and "alcoholic solvents" are understood in the sense of commonly used terminology for alcohol, preferably alcohols with 1 to 10 carbon atoms, more preferably methanol, ethanol, isopropanol, n-propanol, or t-butanol, as well as glycerol, propylene glycol, ethylene glycol, hexylene glycol, polypropylene glycol, and polyethylene glycol, and more preferably ethanol or isopropanol. Such alcohols are solvents that, when added to the aqueous solution, increase the hydrophobicity of the solution by decreasing the polarity of the solution. MOPSO is 3- (N-morpholino) -2-hydroxypropanesulfonic acid. HEPES is N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid. The alcohol reagent is 95 parts by volume (Especially Denatured Alcohol Formula 3A and 5 parts by volume of isopropyl alcohol).

MES is 2- (N-morpholino) ethanesulfonic acid. UF / DF means ultrafiltration / diafiltration. TMAC is tetramethylammonium chloride. TEAC is tetraethylammonium chloride. NGF-120 means the full length of the nerve growth factor 120/120. NGF-118 means the mature, homodimeric NGF molecule of 118 residues. The oxidized NGF means the variant molecule of NGF, Metsulfóxid? 37, which is reported in the present as about 80% as biologically active as the native, mature NGF. Isoasp NGF means the iso bristled variant molecule of NGF, Asp93. Deamidated NGF means NGF which has Asn45 converted to Asp45. RNGF means an NGF molecule with an extra residue of arginine at its N-terminus. CHO stands for Chinese hamster ovary cells. The resins described herein include MACROPREP HIGH S cation exchange resin (BIO-RAD Laboratories, strong cation exchange, SO3 functional group, nominal particle size 50 m, nominal pore size 1000 A); silica gel (not derivatized); Low Rapid Flow Substitution of Phenyl Sepharose (Pharmacia, highly crosslinked 6% agarose, particle size 45-165 microns); SP-Sepharose HP (Pharmacia; highly crosslinked 6% agarose, particle size 34 microns); Phenyl Toyopearl 650 M ÍTosoHaas; particle size 40-90 microns); and Fractogel EMD SO3-650 S (EM Separations, a North American company of E. Merck (Germany), particle size of 25 to 40 m).

Modalities for Carrying Out the Invention Neurotrophins belong to a family of small, basic proteins, which play a crucial role in the development and maintenance of the nervous system. The first identified and probably best understood member of this family is nerve growth factor (NGF). See U.S. Patent No. 5,169,762, issued December 8, 1992. Recently, sequentially related but distinct polypeptides with functions similar to NGF have been identified. For example, the brain-derived neurotrophic factor (BDNF), also referred to as neurotrophin 2 (NT2), was cloned and sequenced by Leibrock et al. (Nature, 341: 49-152).

[1989]). Several groups identified a neurotrophic factor originally called neuronal factor (NF), and now referred to as neurotrophin 3 (NT3).

(E fors et al, Proc. Natl. Acad. Sci. USA, 87: 5454-5458 [1990]; Hohn et al., Nature, 344: 339 [1990]; Maisonpierre et al., Science, 247: 1446 [1990] Rosenthal et al., Neuron, 4: 767 [1990], Jones and Reichardt, Proc. Natl. Acad. Sci. USA, 87: 8060-8064 [1990], Kaisho et al., FEBS Lett., 266: 187 [1990 ]).

Neurotrophin 4/5 (referred to as either NT4 or NT5) has been identified (Hallbook et al., Neuron, 6: 845-858 [1991], Berkmeier et al., Neuron, 7: 857-866 [1991]; Ip et al. , Proc. Natl. Acad. Sci, USA, 89: 3060-3064 [1992]). U.S. Patent No. 5,364,769 issued November 15, 1994, describes human NT-4/5 and the processes for its recombinant expression and is incorporated herein by reference. Also reported are chimeric or pantropic neurotrophins, such as those reported in U.S. Patent No. 5,488,099 issued January 30, 1996, in Urfer et al., EMBO-J-13 (24): 5896-909 (1994), and in International Patent WO 95/33829 published December 14, 1995 (incorporated by reference herein) in which the neurotrophin has been modified to bind to more than one receptor, or contains a receptor binding activity not normally present to a significant degree in the native neurotrophin. Of particular interest are the neurotrophins designated MNTS-1 and D15A NT3. Also of particular interest are neurotrophins that have a NGF amino acid backbone but are modified to bind to different trkA receptors, such as trkB or trkC. Preferred are those in which amino acid substitutions have been made in NGF with an amino acid of a corresponding position in NT-3 which is responsible for trk receptor binding for NT-3. Such NGF mutants have the receptor binding activity similar to NT-3, while retaining the pharmacokinetic and purifying behavior of NGF (Urfer et al., Bi ochemi s try 36 (16): 4775-4781 (1997)) . These NGF mutants may also lack trkA binding activity (Shih et al, J. Biol.Chem. 269 (44): 27679-86 (1994).) Such NGF mutants are particularly preferred neurotrophins for use in The invention described herein The isolation of a recombinant human neurotrophin, for example, rhNGF, involves the separation of the protein from a variety of various contaminants of host cells, each step involving special buffers that make it possible for the Sufficient separation The final or penultimate processing step for a neurotrophin is complicated by the presence of several neurotrophin variants that are co-purified using conventional chromatographic means When a re-folding step is included in the recovery and purification process , variants include misfolded forms of neurotrophin, variants may also include those that differ chemistry. of neurotrophin, such as the carbamylated, deamidated, deamidated or proteolytically excised forms. In the case of NGF, these species consist mainly of dimeric forms, -homodimers, for example, 120/120 or 117/177 when desired 118/188, or heterodimers, eg, 120/118, 117 / 188--, chemically modified variants - glycosylation, mono-oxidized, isoaspartate variants, truncated forms at the N-terminus and at the C-terminus, and dimers thereof. The invention makes possible the large-scale production of neurotrophins, particularly rhNGF, in sufficient quantities for therapeutic uses, such as, for example, the treatment of Al.zheim.ex disease, peripheral neuropathies, including diabetic neuropathy and related to AIDS, and the like. In view of the similarity in sequence and conformation between NGF and other neurotrophins, preferably those in the NGF family, the methods of the present invention can be applied to prepare these neurotrophins substantially free of misprocessed, misfolded or partially folded variants, of glycosylation and / or loading. In the present invention, column resins and conditions are identified, which are favorable for selectively removing neurotrophins from these and other closely related variants. Neurotrophins include NT-3, NT-4/5, NT-6, BDNF, and the engineered forms, including the heterodimeric, chimeric or pantropic forms thereof. Preferably, neurotrophins are human or highly homologous to the human amino acid sequence, preferably more than 80%, more preferably more than 90%, and still more preferably more than 95% homologous to human sequences. A neurotrophin engineered by genetic engineering will retain at least 50% of the trk receptor binding function of the native neurotrophin it resembles, preferably at least 75% and more preferably at least 80%. These genetically engineered forms are those that retain sufficient hydrophobic or high pl character of the native neurotrophin, to preserve similar functioning in the process described herein.

As described below, the process described herein has been successfully applied to rhNGF, rhNT-3 and rhNT-4/5. For example, rhNT-4/5, which was elaborated in E. col i, was isolated in inclusion bodies and reduced and solubilized from the inclusion bodies. The reduced NT-4/5 was partially purified by Fast Flow chromatography on DE Sepharose and by S-Sepharose Rapid Flow chromatography. The S-Sepharose Rapid Flow Combination was re-folded into a buffer containing guanidine for 24 hours. The misfolded forms of NT-4/5 were removed by chromatography as described herein, on a large scale. The misfolded (carbamylated = and trimmed) forms of NT-4/5 were removed by high-resolution cation exchange chromatography using a PolyCat A HPLC resin or HP SP-Sepharose resin, in column format, on a large scale. The purified rhNT-4/5 was ultrafiltered and diafiltered in an acid buffer for the formulation. One embodiment of the invention involves the purification of a neurotrophin from its related variants, usually after the neurotrophin has already been purified from most other impurities, typically in the final or near final step before desalination or diafiltration. before the formulation. The variants related to the mixture may include not only residual variants from a fermentation, but also variants produced if the neurotrophin is degraded in storage or during processing. Neurotrophins suitable for use with the embodiments of the invention can be prepared by any means, but are preferably recombinantly prepared. A nucleic acid molecule encoding the neurotrophins discussed herein is available from several sources, for example, through chemical synthesis using the known DNA sequence or through the use of standard cloning techniques known to those skilled in the art. technique. • The cDNA clones that possess the neurotrophin, for example, the hNGF coding sequence, can be identified by the use of specifically designed oligonucleotide hybridization probes, based on the known sequence of the neurotrophin.

After obtaining a molecule having the neurotrophin coding sequence, the molecule is inserted into a suitable cloning vector for expression in the chosen host cell. The cloning vector is constructed to provide the appropriate regulatory functions required for efficient transcription, translation and processing of the coding sequence. If the neurotrophin is prepared recombinantly, the appropriate host cells for expressing the DNA encoding the neurotrophin are prokaryotic, yeast, or higher eukaryotic cells. Suitable prokaryotes for this purpose include bacteria such as archaebacteria and eubacteria. Preferred bacteria are eubacteria, such as Gram-negative or Gram-positive organisms, e.g., Enterobacteriaceae such as Escherichia, e.g., E. coli, Enterobacter, Erwinia, Klebsiella, Proteus, Salmonella, for example, Salmonella typhimurium, Serratia, for example, Serratia marcescans and Shigella; Bacilli such as B. subtillis and B. licheniformis (for example, B. licheniformis 41P described in DD 266,710 published on April 12, 1989); Pseudomonas such as P. aeruginosa; Streptomyces; Azotobacter; Rhi zobia; I saw treoscill a; and Paracoccus. The right guests of E. col i include E. coli W3110 (ATCC 27,325.}, -B, coli 94 (ATCC 31,446), E. coli B, and E. coli X1776 (ATCC 31,537) These examples are illustrative rather than limiting It is incorporated herein in its entirety PCT publication WO95 / 30686 published on November 16, 1995. The publication is particularly relevant for its description of bacterial synthesis and NGF folding in vi tro The products of that process can be subject to the methods of purification of The present invention Mutant cells of any of the aforementioned bacteria can also be employed., it is necessary to select the appropriate bacteria taking into consideration the replicability of the replicon in the cells of a bacterium. For example, the species of ------. coli, Serra ti a, or Salmonell a may be suitably used as the host when well-known plasmids such as pBR322, pBR325, pACYA177 or pKN410 are used to supply the replicon. Strain W3110 from E. coli is a preferred host or a parent host because this is a common host strain for fermentations that produce recombinant DNA. Preferably, the host cell secretes minimal amounts of proteolytic enzymes. For example, strain W3110 can be modified to effect a genetic mutation in genes encoding proteins endogenous to the host, with examples of such hosts including E. col i W3110 strain 1A2, which has the complete tonA genotype?; E. coli W3110 strain 9E4, which has the complete tonA genotype? ptr3; E. col i W3110 strain 27C7 (ATCC 55,244), which has the complete genotype tonA ptr3 phoA? E15? (a-rgF-lac) 169? degP? ompT kan < r >; E. col i W3110 strain 37D6, which has the complete genotype tonA ptr3 phoA? E15? (argF-lac) 169? degP? ompT? rbs7 ilvG kan r; E. col i W3110 strain 40 B4, which is strain 37D6 with a deletion mutation degP not resistant to kanamycin; and a strain of E, coli having the mutant periplasmic protease described in U.S. Patent No. 4,946,783 issued August 7, 1990. Human NGF has been expressed in E. col i. The isolation and sequence of the gene encoding the β-subunit of hNGF and its expression as a heterologous protein in E. col i, were described in U.S. Patent No. 5,288,622. The teachings herein are also suitable for providing human, mature NGF, produced by mammalian cells. Through the use of recombinant techniques, free human ß-NGF from other mammalian proteins was expressed. The expression of hNGF in E. coli using two genes containing altered amino termini, resulted in the expression of a fused protein, which was described by Iwai et al., Chem. Pharm. Bul l. 34: 4724 (1986). In addition to prokaryotes, eukaryotic microbes such as filamentous fungi or yeasts are suitable as expression hosts for vectors encoding neurotrophin. Saccharomyces cerevi if ae, or common baking yeast, is the most commonly used among lower eukaryotic host microorganisms. However, a number of other genera, species and strains are commonly available and useful herein, such as Schi zosa ccharomyces pombe [Beach and Nurse, Nature, 290: 140 (1981), - European Patent EP-139,383, published on May 2, 1985]; guests Kluyveromyces [U.S. Patent No. 4,943,529; Fleer et al., Bio / Technology, 9: 968-975 (1991)] such as, for example K. lactis [MW98-8C, CBS683, CBS4574; Louvencourt et al., J. Bacteriol., 737 (1983)], K. fragilis (ATCC 12.424), K. bulgaricus (ATCC 16.045), K. wickeramii (ATCC 24,178), K. waltii (ATCC 56, 500.). , K. drosophilarum [ATCC 36.906; Van den Berg et al., Bio / Technology, 8: 135 (1990], K. thermotolerans, and K. marxianus; yarrowia [EP-402, 226]; Pichia pastoris [EP-183,070; Sreekrishna et al., J. Basic Microbiol., 28: 265-278 (1988)]; Candida; Trichoderma reesia [EP-244,234]; Neurospora crassa [Case et al., Proc. Natl. Acad. Sci. USA, 76: 5259 -5263 (1979)], Schwanniomyces such as Schwanniomyces occidentalis [EP-394,538 published October 31, 1990], and filamentous fungi such as Neurospora, Penicillium, Tolypocladium [WO 91/00357 published January 10, 1991], and Aspergillus hosts such as A. nidulans [Ballance et al., Biochem. Biophys., Res. Commun., 112: 284-289 (1983); Tilburn et al., Gene, 26: 205-221 (1983); Yelton et al. Proc. Natl. Acad. Sci. USA, 81: 1470-1474 (1984)] and A. niger [Kelly and Hynes, EMBO J., 4: 475-479 (1985)]. Suitable host cells, suitable for the expression of the DNA encoding neurotrophin, can also be derived from multicellular organisms. Such host cells are capable of complex processing and glycosylation activities. In principle, any culture of higher eukaryotic cells is suitable, be it a vertebrate or invertebrate culture. Examples of invertebrate cells include plant cells and insect cells. Numerous strains of baculoviruses and variants and permissible insect host cells, corresponding from hosts such as Spodoptera frugiperda (caterpillar), Aedes Aegypti (mosquito), Aedes Albopictus (mosquito), Drosophila melanogaster (fruit fly), and Bombyx mori, have also been identified. See, for example, Luckow et al., Bio / Technology, 6: 47-55 (1988); Miller et al., In Genetic Engineering, Setlow, J.K. et al., eds., Vol. 8 (Plenum Publishing, 1986), pp. 277-279; and Maeda et al., Nature, 315: 592-594 (1985). A variety of viral strains for transfection are publicly available, for example, the Ll variant of Autographa californica NPV and the Bm-5 strain of Bombyx mori? PV, and such viruses can be used herein, particularly for the transfection of cells of Spodoptera frugiperda. Human? GF has been produced in insect cells as reported in US Patent No. 5,272,063, issued December 21, 1993. Plant cell cultures of cotton, corn, potato, soy, petunia, tomato, and tobacco , can be used as guests. Typically, plant cells are transfected by incubation with certain strains of the bacterium Agroba cteri um t umefa ci ens, which have been previously manipulated to contain the DNA encoding neurotrophin. During the incubation of the plant cell culture with A. tumefaciens, the DNA encoding the neurotrophin is transferred to the plant cell host such that it is transfected, and, under appropriate conditions, will express the DNA encoding the neurotrophin. In addition, regulatory and signal sequences compatible with plant cells are available, such as the nopalin-synthase promoter and the polyadenylation signal sequences (Depicker et al., J. Mol. Appl. Gen., 1: 561 ( 1982)). In addition, DNA segments isolated from the upstream (5 ') region of the 780 gene of the T-DNA are capable of activating or increasing the transcription levels of the plant-expressible genes in plant tissue containing recombinant DNA (European Patent EP -321,196, published June 21, 1989). Examples of mammalian host cell lines useful are the monkey kidney CVl line transformed by SV40 (COS-7, ATCC CRL 1651); the human embryonic kidney line [293 or 293 cells subcloned for growth in suspension culture, Graham et al., J., Gen Virol., 36: 59 (1977)], baby hamster kidney cells (BHK, ATCC CCL 10) Chinese hamster ovary cells / DHFR [CHO, Urlaub and Chasin, Proc. Na tl, Acad. Sci., USA, 77: 4216 (1980)], mouse sertoli cells [TM4, Mather, Biol. Reprod., 23: 243-251 (1989)], monkey kidney cells (CVl, ATCC CCL 670), African green monkey kidney cells (VERO-76, ATCC CRL.1587), human cervical carcinoma cells ( HELA, ATCC CCL 2), canine kidney cells (MDCK, ATCC CCL 34), buffalo rat liver cells (BRL 3A, ATCC CRL 1442), human lung cells (W138, ATCC CCL 75), human liver cells (Hep G2, HB 8065); mouse mammary tumor (MMT 060562, ATCC CCL51); TRl cells [Mather et al., Annals N.Y. Acad. Sci., 383; 44-68 (1982)]; MRC 5 cells; FS4 cells; and a human hepatoma line (Hep G2). A preferred method is expression in CHO cells. The exon of human NGF containing prepro-NGF, can be used to achieve expression of mature, secreted NGF (including forms 118 and 120) using appropriate promoters and vectors (Patent North American No. 5,288,622). Stable, transfected CHO cell cultures that secrete mature forms of NGF are useful in the invention as discussed in the Examples herein. The host cells are transfected and preferably transformed with the expression or cloning vectors described above and cultured in conventional nutrient media, modified as appropriate for the induction of promoters, the selection of transformants, or the amplification of the genes coding for the sequences desired. Transfection refers to the taking of an expression vector with a host cell, whether or not any coding sequences are in fact expressed. Numerous methods of transfection are known to the person of ordinary skill in the art, for example, CaP04 and electroporation. Successful transfection is generally recognized when any indication of the operation of this vector occurs within the host cell. Transformation means the introduction of DNA into an organism, so that DNA is replicable, either as an extrachromosomal element or as a chromosomal integrant. Depending on the host cell used, the transformation is performed using standard techniques appropriate for such cells. Treatment with calcium using calcium chloride, as described in section 1.82 of Sambrook et al., Molecular Cloning: A Laboratory Manual [? Ew York: Cold Spring Harbor Laboratory Press, 1989], or electroporation is generally used for prokaryotes or other cells that contain substantial cell wall barriers. The infection with Agroba ct eri um t umefaci ens is used for the transformation of certain plant cells, as described by Shaw et al., Gen, 23: 315 (1983) and WO 89/05859 published on June 29, 1989. In addition , the plants can be transformed using ultrasound treatment as described in International Patent WO 91/00358 published January 10, 1991. For mammalian cells without such cell walls, the calcium phosphate precipitation method of Graham and van der Eb. I saw rol ogy, 52: 456-457 (1978) is preferred. The general aspects of mammalian cell host system transformations have been described by Axel in US Patent No. 4,399,216 issued August 16, 1983. Transformations in yeast are typically carried out according to the Van Solingen method and collaborators, J. Bact. , 130: 946 (1977) and Hsiao et al., Proc. Na tl. Acad. Sci. (USA), 76: 3829 (1979). However, other methods for the introduction of DNA into cells can also be used, such as by nuclear microinjection, electroporation, fusion of bacterial protoplasts with intact cells, or polycations, for example, polybrene, polyornithine, etc. For various mammalian cell transformation techniques, see Keown et al., Me thods in Enzymol ogy (1990) Vol. 185, pp. 527-537, and Mansour et al., Na ture, 336: 348-352 (1988). Preferably, the gene for hNGF is inserted into the vector, to have available a methionine initiation codon, preferably one of the two methionine initiation codons identified by Ullrich et al., Na t ure, 303: 821-825 (1983) ). The hNGF gene (Ullrich et al., Col d Spring Harbor Symposium on Quan t. Bi ol. XLVIII, pp. 435 (1983); US Patent No. 5,288,622) has two closely adjacent ethionines that are likely to be used as the start codons of translation (position 1 refers to the N-terminal serine residue of mature hNGF). In contrast, the cDNA of the mouse submaxillary glands for? GF, the most extensively studied of nerve growth factors, has a methionine at position -187 in addition to those at positions -121 and -119. In a preferred embodiment for expression in mammalian cells, the prepro-? GF sequence is present.

If prokaryotic cells are used to produce neurotrophin, they are cultured in suitable media in which the promoter can be constitutive or artificially induced as generally described for example in Sambrook et al., Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory Press, New York 1989). Any necessary supplements in addition to the carbon, nitrogen and inorganic phosphate sources may also be included, at appropriate concentrations introduced alone or as a mixture with another supplement or medium such as a complex nitrogen source. If mammalian host cells are used to produce neurotrophin, they can be cultured in a variety of media. Commercially available media such as Ham's FIO (Sigma), Minimum Essential Medium ([MEM], Sigma), RPMI-1640 (Sigma), and Dulbecco's Modified Eagle Medium ([DMEM], Sigma) are suitable for culturing host cells. In addition, any of the media described in Ham and Wallace, Meth., Can be used as a culture medium for the host cells. Enz., 58:44 (1979); Barnes and Sato, Anal.

Biochem., 102: 255 (1980); US Patent Nos. 4,767,704; 4,657,866; 4,927,762; 5,122,469; or 4,560,655; International Patents WO 90/03430; WO 87/00195; or U.S. Patent No. 30,985, the descriptions of which are incorporated by reference herein. Any of these means can be supplemented as necessary with hormones and / or with other growth factors (such as insulin, transferrin, or epidermal growth factor), salts (such as sodium chloride, calcium, magnesium, and phosphate), buffers (such as HEPS), nucleosides (such as adenosine and thymidine), antibiotics (such as the drug Gentamicin MR), trace elements (defined as inorganic compounds usually present at final concentrations in the micromolar range), and glucose or a source of equivalent energy. Any other necessary supplements may be included at appropriate concentrations that would be known to those of skill in the art. The culture conditions, such as temperature, pH, and the like, are those previously used with the host cell selected for expression, and will be apparent to one of ordinary skill in the art. In general, the principles, protocols, and practical techniques for maximizing the productivity of mammalian cell cultures in vi tro, can be found in Mammalian Cell Biotechnology: A Practical Approach, M. Butler, ed. (IRL Press at Oxford University Press, Oxford, 1991). The above process can be used whether the neurotrophin is produced intracellularly, produced in the periplasmic space, or directly secreted into the medium. Typically, the culture fluid is harvested after an appropriate period and from standard identification assays, for example, immunoassays such as ELISA and Western blot analysis, or biological assays, such as differentiation of PC12 cells (Greene, LA , Trends Ne uros ci. 7:91 (1986)). Tests to determine the type and degree of variants described herein are known in the art or are provided or cited in the Examples (see, for example, Sch elzer et al., "Ne uroch em, 59 (5). : 1675-83 (1992) and Burton et al, J. Neurochem 59 (5): 1937-45 (1992), which are incorporated by reference herein The neurotrophin composition prepared from the cells is preferably subject to at least one purification step prior to HIC Examples of suitable purification steps include those described herein, including affinity chromatography, other techniques for protein purification such as chromatography on silica, chromatography on heparin -Sepahrose, chromatography on an anionic or cationic exchange resin (such as a column of polyaspartic acid), chromatofocusing, and preparative SDS-PAGE, depending on the neurotrophin to be recovered and the initial culture used. In a modality where the neurotrophin is directly secreted into the medium, the medium is separated from the cellular waste by centrifugation, and the broth or clarified fermentation medium is then used for the purification on silica gel. For chromatography on silica, typically the broth is passed through non-derivatized silica particles such that the neurotrophin polypeptide adheres to the silica particles; the silica particles are washed to remove contaminants; and the polypeptide is eluted from the silica particles with a buffer comprising an alcoholic or polar aprotic solvent and an alkaline earth metal, alkali metal, or inorganic ammonium salt. In a preferred embodiment of the invention, the Macroprep High S Cation Exchange Chromatography is used to separate the neurotrophin from its variants, as well as to decrease the bulk contaminants. This resin can be used as an early step in the purification of a neurotrophin from the culture of mammalian cells, preferably for the fractionation of the cell culture medium, harvested. In a further embodiment, the Macroprep High S Cation Exchange Chromatography is used more preferably immediately after a step of protein re-folding. The very high flow property of this cation exchange column allows a large volume of the diluted refolded protein or neurotrophin of the harvested cell culture medium to be easily concentrated, before the subsequent chromatography steps, such as HIC or SP-Sepharose. , by adjusting the conditions, so that the neurotrophins are linked to the column. In addition, the cation exchange nature of the resin allows the removal of unbound bulk proteins, and some poorly processed variants and chemically modified variants (eg, altered charge, variant oxidized with MET37). More importantly when misfolded variants or chemically modified variants (eg poorly processed variants, glycosylation variant) (as of the production of the mammalian cell culture)) that differ in hydrophobicity from the native neurotrophin are present, the Macroprep resin allows at least a partial removal of these hydrophobic variants, substantially enriching the native neurotrophin, as determined herein . While not intended to be limiting, it is believed that the backbone of the resin support has a hydrophobic content that promotes nonspecific interactions between the neurotrophins and the resin, which has been advantageous as shown herein. Typically, for a pH elution buffer of about 5 to 8, more preferably 6 to 8, a TMAC concentration of 0 to 3 M is useful. Sodium acetate, when present to increase the ionic strength of the buffer of elution, allows the use of lower concentration of TMAC. Chloride is a preferred substitute for the acetate ion. In an example of the modality where the neurotrophin is produced in the periplasmic space, the culture medium or lysate is centrifuged to remove the particulate cellular waste. The soluble protein and membrane fractions can then be prepared if necessary. The neurotrophin can then be purified from the fraction of the soluble protein and the membrane fraction of the culture lysate, depending on whether the neurotrophin is bound to the membrane, whether it is soluble, or is present in an aggregated form. The neurotrophin after this is solubilized and then subsequently re-folded using an appropriate buffer. The details for this method of isolation from the periplasm to produce the re-folded protein are described below.

The non-native, insoluble neurotrophin is isolated from the prokaryotic host cells in a suitable isolation buffer by any suitable technique, for example, one involving the exposure of the cells to a buffer of adequate ionic strength to solubilize most of the host proteins, but in which the aggregated neurotrophin is substantially insoluble, and disintegrating the cells to release the inclusion bodies and make them available for recovery for example by centrifugation. This technique is well known, and is described for example in U.S. Patent No. 4,511,503. Briefly, the cells are suspended in the buffer (typically at pH 5 to 9, preferably about 6 to 8, using an ionic strength of about 0.01 to 2 M, preferably 0.1 to 0.2 M). Any suitable salt, including sodium chloride, is useful to maintain a sufficient ionic strength value. The cells, while suspended in this buffer, are then disintegrated by lysis using commonly employed techniques such as, for example, mechanical methods, for example, a Manton-Gaulin microfluidizer, a French press, or a sonic oscillator, or by chemical or enzymatic methods. Examples of chemical or enzymatic methods of cell disintegration include the formation of spheroplasts, which involves the use of lysozyme to lyse the bacterial wall (Neu et al, Bi ochem Bi ophys Res. Comm., 17: 215 (1964)), and osmotic shock, which involves the treatment of viable cells with a solution of high tonicity and with a wash with cold water of low tonicity to release the polypeptides (? Eu and collaborators, J. Bi o?. Chem., 240: 3685-3692 (1965)). A third method, described in U.S. Patent No. 4,680,262, involves contacting the transformed bacterial cells with an effective amount of a lower alkanol having 2 to 4 carbon atoms, for a time and at a temperature sufficient to kill and lyse the cells. After the cells are disintegrated, the suspension is typically centrifuged to turn the inclusion bodies into a pellet or button. In a modality, this step is carried out at approximately 500 to 15,000 x g, preferably approximately 12,000 x g, in a standard centrifuge for a sufficient time which depends on the volume and design of the centrifuge, usually from about 10 minutes to 0.5 hours. The resulting pellet or button contains substantially all of the insoluble polypeptide fraction, but if the cell disintegration process is not complete, it may also contain intact cells or broken cell fragments. The termination of cell disintegration can be evaluated by resuspending the pellet or button in a small amount of the same buffer and examining the suspension with a phase contrast microscope. The presence of fragments of broken cells or whole cells indicates that additional disintegration is necessary to remove the fragments or cells and the associated non-retractable polypeptides. After such disintegration, if required, the suspension is again centrifuged and the pellet or button is recovered, resuspended and analyzed. The process is repeated until the visual examination reveals the absence of broken cell fragments in the concentrated material in the form of a button or until the additional treatment can not reduce the size of the resulting pellet. In an alternative embodiment, the neurotrophin is isolated from the periplasmic space by solubilization in a suitable buffer. This procedure can be the solubilization in si t u that involves the direct addition of the reagents to the fermentation vessel after the neurotrophin has been recombinantly produced, thereby avoiding extra steps of harvest, homogenization, and centrifugation to obtain the neurotrophin. The remaining particulates can be removed by centrifugation or filtration, or combinations thereof. If the neurotrophin is being deployed, the degree of unfolding is adequately determined by non-native neurotrophin chromatography, including RP-HPLC. The increase in the peak area for the non-native material indicates how much non-native neurotrophin is present. Once obtained from the solubilized inclusion bodies or in a subsequent purification step, the neurotrophin is properly re-folded into an active conformation as described below. If the neurotrophin is no longer in soluble form before it is re-folded, it can be solubilized by incubation in chaotropic agent containing alkaline buffer and reducing agent in amounts necessary to substantially solubilize the neurotrophin. This incubation takes place under conditions of neurotrophin concentration, incubation time, and incubation temperature that will allow the solubilization of the neurotrophin to occur in the alkaline buffer. The measurement of the degree of solubilization of the neurotrophin in the buffer is suitably carried out by determination of the turbidity, by means of analysis of the fractionation of the neurotrophin between the supernatant and the button or pellet after centrifugation on reduced SDS gels, by means of a test. of protein (for example, the Bio-Rad protein assay kit), or by HPLC. The pH range of the alkaline buffer for solubilization is typically at least about 7.5, with the preferred range being about 8-11. Examples of suitable buffers that will provide a pH within this latter range include glycine, CAPSO (3- [cyclohexylamino-2-hydroxy-1-propanesulfonic acid), AMP (2-amino-2-methyl-1-propanol), CAPS (3- [cyclohexylamino] -l-propanesulfonic acid), CHES (2- [N-cyclohexylamino] ethanesulfonic acid), and TRIS-HC1 (Tris [hydroxymethyl] aminomethane hydrochloride). The preferred buffer herein is glycine or CAPSO, preferably at a concentration of about 20 M, at a pH of about 8.5 to 11., preferably from about 10 to 11. The concentration of the neurotrophin in the buffered solution for solubilization should be such that the neurotrophin will be substantially solubilized and partially or completely reduced and denatured. Alternatively, the neurotrophin may be initially insoluble. The exact amount to be employed will depend, for example, on the concentrations and types of other ingredients in the buffered solution, particularly the type and amount of the reducing agent, the type and amount of the chaotropic agent, and the pH of the buffer.

For example, the concentration of the neurotrophin can be increased at least three times if the concentration of the reducing agent, e.g., DTT, is concurrently increased, to maintain a DTT: neurotrophin ratio of about 3: 1 to 10: 1. It is desirable to produce a more concentrated solubilized protein solution before refolding by dilution. Thus, the preferred concentration of the neurotrophin is at least about 30 mg / ml, with a more preferred range of 30 to 50 mg per ml. For example, the neurotrophin can be solubilized at a concentration of about 30-50 mg / ml in 5M to 7M urea, 10mM DTT and diluted, for example, to about 1 mg / ml for folding. After the neurotrophin is solubilized, it is placed or diluted in a refolding buffer containing alcoholic or aprotic solvent at 5-40% (v / v), a chaotropic agent, and an alkaline earth metal, alkaline earth metal salt, or of ammonium. The buffer can be any buffer for the first buffered solution, with CAPSO, glycine and CAPS which are preferred at pH 8.5-11, particularly at a concentration of approximately 20 mM, and more preferably CAPSO and glycine. The neurotrophin can be diluted with the refolding buffer, preferably at least five times, more preferably at least about ten times. Alternatively, the neurotrophin can be dialyzed against the refolding buffer. The refolding can be carried out at about 2 ° C-45 ° C, more preferably at about 2 ° C-8 ° C for at least about one hour. The solution also optionally contains a reducing agent and an osmolyte. The reducing agent is suitably selected from those described above for the solubilization step in the given concentration range. Its concentration will depend especially on the concentrations of the alkali metal salt, alkaline earth metal or ammonium salt, neurotrophin and solvent. Preferably, the concentration of the reducing agent is from about 0.5 to 8 mM, more preferably from about 0.5 to 5 mM, even more preferably from about 0.5 to 2 mM. Preferred reducing agents are DTT and cysteine.

The oxygen in the refolding solution may optionally be depleted by the addition of an inert gas, for example helium or argon, to displace the oxygen. The optional osmolyte is preferably sucrose (in a concentration of about 0.25 to 1 M) or glycerol (in a concentration of about 1 to 4 M). More preferably, the concentration of sucrose is about .mu.M and the glycerol concentration is about 4M. The initial concentration of the neurotrophin in the folding buffer is such that the ratio of the correctly folded to the misfolded, recovered form will be maximized, as determined by HPLC, RIA, or bioassay. The preferred concentration of neurotrophin (which results in the maximum yield of correctly folded conformer) is in the range of from about 0.1 to 15 mg / ml, more preferably from about 0.1 to 6 mg / ml, and still more preferably from about 0.2 to 5 mg / ml. The degree of refolding that occurs after this incubation is adequately determined by the RIA title of neurotrophin or by HPLC analysis with the increasing RIA titer, or the peak size of correctly folded neurotrophin, which correlates directly with the increasing amounts of the correctly folded, biologically active neurotrophin conformer present in the buffer. Incubation is carried out to maximize the performance of the correctly folded neurotrophin conformer and the proportion of the correctly folded neurotrophin conformer to the misfolded, recovered neurotrophin conformer, as determined by RIA or HPLC., and to minimize the performance of associated, multimeric neurotrophin as determined by mass balance. Alternatively, the species can be determined by means of the methods provided below and in the examples. Guanidine is a preferred denaturing agent for refolding. After the neurotrophin is withdrawn, the following procedures as shown herein, individually or in combination, are exemplary of the purification procedures suitable for obtaining greater purity and homogeneity: fractionation on cation exchange columns; hydrophobic interaction chromatography (HIC); and chromatography on silica. Whether a refolding step is part of the process or not, a preferred step for the separation of neurotrophin from its variants is separation on the hydrophobic interaction chromatography resin. During the fermentation, purification or refolding of the protein in vi tro, some protein may be chemically modified, poorly processed, or may not retract into its native three-dimensional structure, but rather to other structures that differ with respect to its stability, solubility, immunogenicity, or bioactivity. These variants can be eliminated during recovery to avoid undesirable side effects such as antigenicity or loss of potency. If the variant is insoluble it can be easily removed by solid / liquid separation techniques such as centrifugation and filtration. However, if the variant is soluble, higher resolution absorption techniques such as chromatography will be required to eliminate it. When procaryotic cells are produced or when they fold back, the neurotrophins form a misfolded, stable, soluble variant. The misfolded neurotrophin has an altered disulfide pairing pattern, and three-dimensional structure altered in relation to the native neurotrophin and lacks native pharmacological activity. When produced "in eukaryotic cell culture, e.g., mammalian cell culture, the variant forms are typically poorly processed forms.They also typically lack of native pharmacological activity and must be eliminated." HIC has been found to be suitable for separating these variants from the native neurotrophin.HIC involves the sequential adsorption and desorption of the protein from the solid matrices, mediated through the non-covalent hydrophobic bond.In general, the sample molecules in a high-salt buffer, are loaded onto the HIC column, the salt in the buffer interacts with the water molecules to reduce the solvation of the molecules in the solution, thereby exposing the hydrophobic regions in the sample molecules that are consequently adsorbed by the column HIC: The more hydrophobic the molecule, the less salt is needed to promote the bond. Therefore, a decreasing salt gradient is used to elute the samples from the column. As the ionic strength decreases, the exposure of the hydrophilic regions of the molecules increases, and the molecules elute from the column in order of increasing hydrophobicity. Elution of the sample has also been achieved by the addition of mild organic modifiers or detergents to the elution buffer. HIC is reviewed in Protein Purification, 2nd ed., Springer-Verlag, New York, pages 176-179 (1988). The strength of the association between a protein and a matrix depends on various factors, including the size and hydrophobicity of the immobilized functional group, the polarity and surface tension of the surrounding solvent, and the hydrophobicity of the protein. The binding capacity of HIC matrices tends to be lower due to the need for the immobilized hydrophobic ligand to be widely spaced. In addition, the capacity of a medium for a given protein varies inversely with the level of hydrophobic impurities in the sample preparation. In order to resolve a desired protein of the variants and other impurities, while simultaneously maximizing capacity, it is necessary to identify a suitable HIC solid phase medium, as well as the mobile phases suitable for loading, washing and elution. As determined herein, the most appropriate means for the separation of correctly folded and misfolded neurotrophins or the poorly processed forms of the correctly processed, intact forms were those that have immobilized phenyl functional groups. Phenyl-based HIC media from different vendors showed different efficiency for the resolution of these forms of neurotrophin. Better results were achieved with the Phenyl Toyopearl medium from TosoHaas and Phenyl Sepharose Fast Flow Low Sub (low substitution). TSK Phenyl 5PW was also adequate. Other functional groups immobilized by HIC may function to separate these forms. The examples were the octyl groups, such as those on Pharmacy Octyl Sepharose CL4B media, and propyl groups, such as those on Baker's High Propyl media. Less preferred are resins with alkoxy, butyl, and isoamyl functional groups.

HIC was useful for the separation of neurotrophins from their variants in mammalian cell culture. For example, as determined herein, the culture of CHO cells expressing rhNGF contained proteolytically incorrectly processed variants, such as those in which a partial precursor sequence is present, eg, NGF precursor, NGF hybrid precursor, and NGF cleaved precursors. Also found in mammalian cell culture media are the glycosylated and glycosylated NGF forms of the incorrectly processed proteolytic variants. Undesirable glycosylated forms, which in the case of NGF can be observed as a species of higher molecular weight (greater than 2,000 kD), could generate an unwanted antigenic response in a patient, and contribute to the poor quality or activity of the patient. product. HIC effectively separated the hydrophobic variants, primarily the non-proteolytically poorly processed N-ter variants, including the glycosylated forms, from rhNGF. As shown in the examples, the NGF of cut precursor sequence and containing the precursor sequence, and the glycosylated forms of NGF and NGF containing the precursor sequence, eluted at the leading edge of the NGF peak during phenyl-HIC. In this way, a rhNGF composition that was substantially free of these species, and which was particularly suitable for a subsequent step such as high-resolution cation exchange chromatography, could be obtained. HIC is applicable to other neurotrophins, as well as to NGF, notwithstanding the source. For example, HIC is useful for separating the NGF monomers from the dimers, either homo- or heterodimers depending on the monomeric forms present, as well as for distinguishing the dimeric forms that also differ in hydrophobicity, which are obtained after the refolding in vi or when they are produced and secreted from mammalian cells. A preferred source of neurotrophin mixtures for use with HIC is the culture of mammalian cells, more preferably the culture of CHO cells. The culture is preferably subject to at least one prior purification step as discussed herein. HIC is particularly effective in separating glycosylated, poorly processed variants from the native recombinant neurotrophin. In the case of rhNGF, the glycosylated and prepro forms of NGF are less hydrophobic than native NGF, thereby eluting before native NGF. The misfolded forms of neurotrophins (when they are produced by bacteria) are also more hydrophobic, eluting earlier than the native neurotrophin. The most preferred HIC resin for the separation of neurotrophin forms was that having immobilized phenyl functional groups. Phenyl-based HIC media from different suppliers showed different efficiency for the resolution of these forms of NGF. Of the phenyl-HIC resins, Phenyl Toyopearl medium by TosoHaas is more preferred, and Phenyl Sepharose Fast Flow Low Sub (low substitution) and TSK Phenyl 5PW are preferred. Preferred functional groups of HIC include the alkoxy, butyl, and isoamyl portions. Using HIC, a variety of mobile phase conditions can be used to differentially wash and elute neurotrophin forms. These mobile phases may contain several different chemical species that influence the association between a neurotrophin and the stationary phase in different ways. Neurotrophin correctly folded and badly folded, for example, NT-4/5, can be resolved on a column of HIC by decreasing saline gradients or gradual decrease, for example of the salt in mobile phase, for example ammonium sulfate , the concentration of sodium chloride, the concentration of acetate. The salts can influence the binding of a neurotrophin to the resin by modulating the surface tension of the mobile phase. Other agents that affected surface tension were sodium citrate and tetramethylammonium chloride, as discussed in the Examples. The variants can also be resolved by column chromatography by eluting the bound protein with increasing gradients or by gradually increasing the concentration of the relatively polar organic solvents. Examples of suitable solvents include ethanol, acetonitrile, and propanol. The strength of the association between the neurotrophin forms and the HIC resin also depended on the pH of the mobile phase, with the neutral conditions being preferred. The relative hydrophobicity of the correctly folded neurotrophin and the misfolded neurotrophin also depended on the pH of the solution.

Separation of native neurotrophin variants could also be obtained by varying various properties of the mobile phase simultaneously during gradual or gradient elution. For example, a mobile phase that simultaneously varies the salt concentration and the concentration of apolar solvent during the elution, would provide better resolution than when only the salt was varied. For HIC, the salts discussed herein may be used, including ammonium sulfate, citrate, acetate, and potassium chloride. Depending on the salt used, the salt concentration is typically 0.5 to 3 M, more preferably 0.5 to 2.5 M, to achieve the binding of the neurotrophin to the resin. For example, a link buffer of 0.8 to 1.5 M salt is preferred for NGF, with higher salt concentrations leading to the precipitation of NGF on the resin, resulting in less recovery. For NT-3 a binding buffer at pH 7 with a salt concentration of 1.0 to 2.5 M is preferred, with 1.25 to 1.75 M of NaCl which is more preferred, and 1.5 M which is most preferred. For NT-4/5 a binding buffer at pH 7 with a salt of 1 to 3 M is preferred, with 2 to 2.75 M which is more preferred, and 2.5 M sodium chloride which is most preferred. In the case of NT-4/5, when 2.5 M NaCl was preferred for loading, 2M NaCl was preferred for elution with the organic solvent present (e.g., 10% alcohol, pH 7). Preferably, a decrease in the salt concentration is used to elute and separate a neurotrophin and its variants. In order to achieve elution, the salt concentration in the elution buffer is typically lower than that in the charge buffer, but this may be the same concentration when it is compensated with the organic solvent. In addition, the use of the organic solvent has another advantage, as has been found in the present, that the addition of an organic solvent improves the elution pattern resulting in narrower peak profiles. In addition to ethanol, other organic solvents discussed herein may be used, including propanol, isopropanol, and lower alkylene glycols, such as propylene glycol, ethylene glycol, and hexylene glycol. The organic solvent at 5-25% (v / v), more preferably 5 to 20% (v / v), will typically elute a correctly folded neurotrophin. Elution with organic solvent can be either gradient or gradual. The pH range is preferably close to neutral to slightly acidic, from pH 5 to 8, more preferably pH 6 to 8, pH 6.5 to 7.5, and most preferably pH to 7. Any of the buffers discussed herein, including MOPSO, MOPS, HEPES, phosphate, citrate, ammonium, acetate, can be used as long as these buffer to the desired pH. In view of the discovery by the present inventors of certain undesirable neurotrophin variants arising from the recombinant production of a neurotrophin, as reported herein, the use of high resolution cation exchange chromatography, preferably in the preparative mode, allows the separation of the modified variants in charge, such as the carbamylated, oxidized, isoasp, deamidated forms, and certain cut forms (for example truncated forms in the C-terminus of NGF) from the native neurotrophin. For example, the shapes cut at the end (for example 2 to 4 deletions of N-terminal amino acids) that result in charge alteration, which may occur during bacterial fermentation as in the case of NT-4/5 and NT -3, can now be eliminated. In the case of neurotrophins produced in mammalian cell culture, C-terminal truncation can occur in the highly loaded terminal region. For example, NGF form 118 can be poorly processed or cleaved at its C-terminus to forms 117, 114 and 115. These can be separated from native NGF form 118 by high-resolution cation exchange chromatography. Particularly preferred modalities use High Resolution SP-Sepharose, Fractogel EMD S03, or polyaspartic acid resin, of which PolyCAT A is a particularly preferred. More preferably, on a large scale, the resins of SP-Sepharose of High Resolution and Fractogel EMD S03 are used. The compositions obtained by the processes described herein will be substantially pure, more usual and preferably essentially pure neurotrophin, and will be substantially free of neurotrophin variants, more preferably essentially free of neurotrophin variants. For example, a typical SP-Sepharose cocktail after purification of NGF from the CHO cell culture contains about 92% of 118, 4.6% of 120, and 1% of deamidated NGF, 1% of oxidized NGF and 1 % of NGF isoasp. Routinely, the amount of each species is in the range of about 85 to 93% for 118, 0 to 5% for 120 (depending largely on the degree of endogenous and / or exogenous proteolysis being used), 0 to 5% for 117, 0 to 3% for the deamidated forms, 0 to 2% for the isoasp forms, and 0 to 2% for the oxidized forms. The purity of NGF (all species) is routinely greater than 99.5%. After the neurotrophin is eluted from the column, it is suitably formulated in a composition with a carrier, preferably a pharmaceutical composition with a physiologically acceptable carrier. The neurotrophin compositions are preferably sterile. The neurotrophin compositions of the invention also find use in vi tro, for example, to promote the development and survival of neurons in culture. • The chemical and physical stability of recombinant human nerve growth factor (NGF) in aqueous solution was investigated between 5 and 37 ° C, in the pH range of 4.2 to 5.8. The chemical stability of NGF increased with the increase in pH. In the succinate buffer at pH 5.8, the physical stability of NGF decreased due to protein aggregation. Based on stability data at 5 ° C and accelerated degradation studies at 37 ° C, it was found that the optimal formulation was the acetate buffer at pH 5.5 (see International Document WO 97/17087 which is incorporated by reference at the moment) . Reverse phase HPLC was the primary stability indication method, which shows conversion of Asn-93 to iso-Asp, which is the main degradation pathway at 5 ° C. The quantification of NGF degradation by cation exchange chromatography was complicated by the rearrangement of monomeric variants of NGF in various mixed dimers over time (dimeric exchange). The treatment of the samples and the controls with dilute acid quickly balanced the monomeric distribution in the dimers, allowing the degradation of NGF to be quantified in the absence of dimeric exchange. Benzyl alcohol and phenol were evaluated for compatibility and stability with rhNGF in two liquid formulations for multipurpose purposes. These two formulations consist of 0.1 mg / ml protein in 20 mM sodium acetate at pH 5.5 and 136 mM sodium chloride with and without 0.01% pluronic acid (F68) as a surfactant. The final concentrations of benzyl alcohol and phenol in each of these two formulations were 0.9 and 0.25%, respectively. Based on the 12-month stability data, rhNGF is more stable with benzyl alcohol than with phenol in these formulations. The rh? GF formulation preserved in benzyl alcohol with the presence of surfactants is as stable as the formulation without the added surfactant, indicating that the addition of F68 to the multiple dose formulation of rh? GF is not required for stability purposes. Therefore, a formulation consisting of 0.1 mg / ml protein in 20 mM acetate, 136 mM sodium chloride, 0.9% benzyl alcohol, pH 5.5, is recommended for rh? GF used for multiple dosing in clinical trials in Phase III. This multiple dose formulation of rh? GF passed the conservation efficacy test of the USP (North American Pharmacopoeia) and the EP (European Pharmacopoeia) after 6 months at 5 ° C, and is as stable as the current liquid formulation at 2 mg / ml. However, the formulation should avoid exposure to intense light due to the presence of benzyl alcohol as a preservative, which is sensitive to light. In general, the compositions may contain other components in amounts preferably that do not detract from the preparation of the stable, liquid or lyophilizable forms and in amounts suitable for effective and safe pharmaceutical administration. The neurotrophin is formulated with a pharmaceutically acceptable carrier, for example, one that is non-toxic to patients and at the doses and concentrations employed, and is compatible with other ingredients of the formulation. For example, the formulation preferably does not include oxidizing agents and other compounds known to be harmful to the polypeptides. This formulation step is achieved through desalination or diafiltration using standard technology. In general, the formulations are prepared by contacting the neurotrophin uniformly and intimately with liquid carriers or finely divided solid carriers, or both. Then, if necessary, the product is formed into the desired formulation. Preferably, the carrier is a parenteral carrier, more preferably a solution that is isotonic with the patient's blood. Examples of such carriers include water, saline, Ringer's solution and dextrose solution. Non-aqueous vehicles such as fixed oils and ethyl oleate are also useful herein, as well as liposomes. The carrier suitably contains minor amounts of additives such as substances that improve isotonicity and chemical stability. Such materials are non-toxic to containers at the doses and concentrations employed, and include buffers such as phosphate, citrate, succinate, acetic acid, and other organic acids or their salts; antioxidants such as ascorbic acid; low molecular weight polypeptides (less than about ten residues), eg, polyarginine or tripeptides; proteins such as serum albumin, gelatin, or immunoglobulins; hydrophilic polymers such as polyvinylpyrrolidone; amino acids, such as glycine, glutamic acid, aspartic acid, or arginine; monosaccharides, disaccharides, and other carbohydrates including cellulose or its derivatives, trehalose, glucose, mannose, or dextrins; chelating agents such as EDTA; sugar alcohols such as mannitol or sorbitol; counterions such as sodium; and / or nonionic surfactants such as polysorbates, poloxamers or PEG. The final preparation can be liquid or solid lyophilized. The neurotrophin to be used for therapeutic administration must be sterile. Sterility is easily achieved by filtration through sterile filtration membranes (eg, 0.2 micron membranes). Therapeutic neurotrophin compositions are generally placed in a container having a sterile access port, for example, a bag for intravenous solution or a bottle having a plug pierceable by a needle for hypodermic injection. The above formulations are also suitable for in vi tro uses. The neurotrophin will ordinarily be stored in unit dose or multiple dose containers, for example, sealed ampoules or flasks, as an aqueous solution, or as a lyophilized formulation for reconstitution. As an example of a lyophilized formulation, 10 ml vials are filled with 5 ml of sterilized aqueous neurotrophin solution by filtration, at 1% (w / v) and the resulting mixture is lyophilized. The infusion solution is prepared by reconstituting the lyophilized neurotrophin using water injection, bacteriostatic. The therapeutically effective dose of a neurotrophin formulation is administered to a patient. By "therapeutically effective dose" is meant herein a dose that produces the effects for which it is administered. The exact dose will depend on the disorder that is to be treated, and will be ascertainable by one of skill in the art using known techniques. In general, the neurotrophin formulations of the present invention are administered at about 0.01 μg / kg to about 100 mg / kg per day. Preferably, from 0.1 to 0.3 μ / kg. Furthermore, as is known in the art, adjustments for age as well as for body weight, general health, sex, diet, time of administration, drug interaction and severity of the disease may be necessary, and will be evaluable with routine experimentation by those skilled in the art. Typically, the physician will administer neurotrophin formulations of the invention until a dose is reached that repairs, maintains and optimally restores neuronal function. The progress of this therapy is easily verified by conventional tests. The neurotrophin is optionally combined with or administered in conjunction with other neurotrophic factors including NGF, NT-4/5, NT-3 and / or BDNF, and is used with other conventional therapies for nerve disorders. In the case of NGF, preferably a composition comprises a pharmaceutically effective amount of nerve growth factor and a buffer containing pharmaceutically acceptable acetate. The composition may have a pH of 5 to 6. The buffer is preferably sodium acetate. The acetate concentration is preferably 0.1 to 200 mM. The composition preferably has a NGF concentration of 0.07 to 20 mg / ml. And, the composition optionally further contains a pharmaceutically acceptable preservative, such as benzyl alcohol, phenol, m-cresol, methyl paraben, or propyl paraben. Preferably, the preservative is benzyl alcohol. The concentration of benzyl alcohol is preferably from 0.1 to 2.0%. The composition may optionally contain a pharmaceutically acceptable surfactant. Y, the composition may optionally, but preferably, contain a physiologically acceptable concentration of sodium chloride. A more preferred composition contains the nerve growth factor at a concentration of at least 0.1 mg / ml and an acetate ion concentration of 10 M to 50 mM. Still more preferably, the composition contains the nerve growth factor at a concentration of 0.1 to about 2.0 mg / ml and the acetate ion at a concentration of 10 mM to 50 M. A more preferred composition contains NGF at a concentration of 0.1 mg / ml. ml, a concentration of sodium acetate of 20 mM, pH 5.5, a concentration of sodium chloride of 136 mM, and benzyl alcohol at 0.90% (v / v). Another embodiment contains a NGF concentration of 2.0 mg / ml, a concentration of sodium acetate of 10 M, pH 5.5, and a sodium chloride concentration of 142 mM. It is preferable to formulate NGF with 0.1 mg / ml, 20 mM sodium acetate, 136 mM sodium chloride, 0.9% benzyl alcohol (v / v), at a pH of 5.5. As discussed herein, homodimer 118/118 is a preferred form of NGF. NGF is purified at pH 6 to 8 to maintain the normal dimeric form. However, the percentage of cut shapes (proteolytically) represents monomeric forms, which become apparent and can be determined by reverse phase HPLC. The acidic conditions of the analytical HPLC dissociate the dimers. The existence of different dimeric forms of NGF - 120/120, 120/118, 118/118, etc. - has been published (Schmelzer et al., J. Ne uroch em 59: 1675-1683 (1992), which is specifically incorporated in the present in its entirety, mainly by its analytical tests and bioassays that were used in the concurrent studies, as well as by its general teachings). That publication reported that the in vi tro activities were the same for each dimeric form. However, in contrast, the present studies demonstrate for the first time that the 120/120 dimer is less active, approximately 80 to 90% as active, as the 118/118 species, using a radio-receptor-based test. In one form of the test, rat PC-12 cell membranes are isolated and used for competitive binding between the NGF standard and the various test species.

The RRA has receivers of P75 and trkA. It was also found in the present that the species 117/117 is as active as the species 188/118. In addition, the use of a PC-12-based assay in the present confirmed the finding of receptor-based assay, which shows that the 120/120 form is approximately as active as the 118/118 form. Also incorporated in its entirety specifically by reference is Burton et al., J. Neurochem. 59: 1937-1945 (1992) mainly for his analytical essay and bioassay, which were carried out in the current studies, as well as for his general teachings. It is believed that the 118/118 form is more bioavailable in human patients than the 120/120 form. The increase in bioavailability is at least 4 to 5 times. This difference is significant, surprising and unexpected in view of the technique. The following examples are offered by way of illustration and not by way of limitation. The description of all citations in the specification are expressly incorporated by reference herein.

EXAMPLES Example I Purification of the NGF homodimer 118/118 This example illustrates the purification of NGF and the explanation for each step. As in each of the Examples, one of ordinary skill in the art can easily determine and adjust the column dimensions and flow rates to compensate for initial culture volumes and protein concentrations, as is well known in the art. .

Harvested Cell Culture Fluid Recombinant CHO cells were transfected with an expression vector containing the DNA sequence encoding the 120 amino acid human NGF. To promote secretion and processing, the prepro sequence of NGF was also present. After the culture of the recombinant CHO cells, the cell culture medium was harvested. Cropped Cell Culture Fluid (HCCF) contained NGF species 120, 118, and 117. Approximately 40 to 70% of the. NGF was typically a 118/118 homodimer with the remainder as heterodimers 120/118, 120/120, and a small amount of 118/117. As shown herein, these species can be separated by HP column of SP-Sepharose. Cropped Cell Culture Fluid was concentrated to approximately one twentieth using the 10 kD cutting membranes (either cellulose, composite or polysulfone, which were used interchangeably). To the concentrate was added 0.1 volume of 1.0 M Tris, pH 8.2. The diluted material was microfiltered using a 0.22 micron filter and transferred to a holding tank at 37 ° C for 2 to 18 hours. The conversion of form 120/120 to form 118/118 is catalyzed by an endogenous protease during the maintenance period.

Chromatography on Silica Gel The microfiltrate was adjusted to NaCl IM and applied to a column of silica gel equilibrated in 1M NaCl, 25 mM MOPSO, pH 7. The column was washed with NaCl IM, MOPSO 25 M, pH 7. The suitable pH range is from about 6 to 8, with a preferred pH of 7. The column was then washed with 25 mM MOPSO, pH 7. A low conductivity wash removes the proteins from the host cell. The bound NGF was eluted with 50 mM MOPSO, 0-5 M TMAC, 20% anhydrous reagent grade alcohol (especially denatured alcohol to 94-96% of formula 3A (5 volumes of methanol and 100 volumes of 200 proof ethanol) and 4-6% isopropanol). Other alcohols such as 20% propanol, 20% isopropanol and 20% methanol can be used. As used herein, "alcohols" and "alcohol solvents" are understood in the sense of commonly used terminology for alcohol, preferably alcohols with 1 to 10 carbon atoms, more preferably methanol, ethanol, isopropanol, n-propanol, or t-butanol, and still more preferably ethanol or isopropanol. Such alcohols are solvents that, when added to the aqueous solution, increase the hydrophobicity of the solution by decreasing the polarity of the solution. Ethanol is the most preferred. The lower limit of alcohol is any percentage that elutes, and the upper limit is adjusted by the need to avoid denaturation of the protein. The solvent is preferably from 5% to 25%, more preferably from 5% to 20%, even more preferably from -5 to 15%. TMAC is tetramethylammonium chloride, which is present to elute NGF. TMAC can be in the range of 0.1 to 1 M. With the range of 0.3 to 0.7 M which is most preferred. The amount of TMAC used to elute NGF is a function of pH and alcohol concentration. The lower the pH, the lower amounts of alcohol and TMAC are required. The pH may be between about 4 to 8. In this example the preferred pH was 7, which allows very minimal adjustment of the combined fractions before loading onto the next column. The upper limit of pH is determined by the pH necessary to load the next column, and the lower limit by that useful to elute NGF efficiently.

Rapid Flow Chromatography in S-Sepharose The eluent containing NGF was combined, diluted to a conductivity of less than 15.5 ms / cm with purified water, and the pH was adjusted to 7.0. The material was retained no more than 8 hours since several proteases were still present; however, endogenous protease activity was not observed converting the NGF from 120 amino acids to the 118 form. The material was applied to a S-Sepharose Rapid Flow chromatography column (a S-SEPHAROSE ™ cation exchange resin). Fast Flow agarose (Pharmacia)) equilibrated in 25 mM MOPSO, pH 7. The column was washed with 25 mM MOPSO, pH 7. A suitable pH range is from about 6 to 8, with pH 7 being preferred. The column was then washed with 0.16 M NaCl, pH 7. The bound NGF was eluted with 0.5 M NaCl, pH 7. The molarity of the elution salt can be in the range of 0.3 to 1.0 M, more preferably 0.4 to 0.6 M. The lower limit is adjusted by the utility to elute all the NGF, and the upper limit is adjusted by the need to avoid the elimination of contaminants and cause hydrophobic interactions on the column which could interfere with the elution of NGF. Other salts may be used, with potassium chloride being a preferred alternative. Elution with 0.5 M NaCl, pH 7, is preferred in order to obtain a combined with a small volume. At higher salt concentrations, for example above IM, the tightly bound contaminants can elute.

Phenyl Toyopearl 650M Chromatography The SSFF column fractions containing the NGF were combined, adjusted to 1M NaCl, and applied to a Phenyl Toyopearl 650M column. The column was washed with 25 mM MOPSO, pH 7. A suitable pH is in the range of about pH 5 to 8. The bound NGF was eluted with a linear gradient of 10 CV (column volume) beginning with buffer A in gradient (25 mM MOPSO, pH 7, 1.0 M NaCl) and ending with buffer B in gradient (20% alcohol in 80% MOPSO 25 mM, pH 7). Fractions - containing NGF were analyzed by polyacrylamide gel electrophoresis of SDS-PAGE to determine which fractions harbored the precursor NGF species. Fractions containing NGF were selected and combined to eliminate incorrectly processed variants, such as those in which a partial precursor sequence is present, eg, NGF precursor, NGF hybrid precursor, and cut NGF precursor sequences, to obtain a composition of NGF substantially free of any? GF precursor sequences. The phenyl column also removed the small amount of the glycosylated? GF sequences and glycosylated NGF precursor. The sequences of precursor NGF and cut precursor, together with the glycosylated forms of NGF and NGF precursor eluted at the leading edge of the NGF peak. Thus, this column easily separated NGF from the various NGF species, to obtain a composition of NGF substantially free of these species. In this step the hydrophobic variants of variant NGF, mainly the poorly processed variants, including the proteolytic and glycosylated variants, were separated using HIC. The most suitable means for the separation of the forms of NGF were those having the immobilized phenyl functional groups. Phenyl-based HIC media from different suppliers showed different efficiency for the resolution of these forms of NGF. The best results were achieved with the Phenyl Toyopearl medium by TosoHaas. The HIC Phenyl Sepharose Fast Flow Low Sub (low substitution) and TSK Phenyl 5PW resins worked well. Other functional groups of HIC were less suitable, and less effective, under these conditions, including the alkoxy, butyl, and isoamyl moieties.

The Phenyl Toyopearl combination contained 75% of 118, 10% of 120, 7% of 117, 1.8% of deamidated NGF, 1.4% of oxidized NGF, and 2.0% of NGF isoasp, with the remaining 2.8% being other species of NGF not identified. Optionally, the combined fractions were treated with acid to achieve viral inactivation at a pH of less than 3.95 for a minimum of 15 minutes.

HP Chromatography of SP-Sepharose The combined HIC was diluted with 0.5 to 1 volume of water and the diluted cocktail was adjusted to pH 6. The combined was loaded onto an HP column of SP-Sepharose equilibrated with 0.2 M sodium chloride , 20 mM succinate, pH 6, containing 5% reactive alcohol (as in the silica gel column passage). The column was washed with 0.2 M sodium chloride, 20 mM succinate, pH 6 containing 5% reactive alcohol (alcohol of Formula SDA-3A, alcohol is optionally present). Alcohol helps reduce non-specific (mainly hydrophobic) interactions of NGF with the backbone of the resin. A suitable alcohol range is from about 0 to 10%. The pH of the charge is about 5 to 8, which is chosen to achieve and maintain the maximum stability of NGF and the separation of the NGF variants. The column was washed with two column volumes of equilibrium buffer. The bound NGF was eluted and separated from the variants by a linear gradient of 22 column volumes, by mixing 11 column volumes of buffer A in gradient (0.25 M sodium chloride, 0.02 M succinate, pH 6, containing 5 volumes). % alcohol) and 11 column volumes of buffer B in gradient (0.5 M sodium chloride, pH 6). Alcohol is optional. NGF 118/118 typically eluted at a 0.35 M to 0.40 M sodium chloride concentration. The column fractions were analyzed for NGF content and NGF variants. The fractions are preferably analyzed by C4 reverse phase high-resolution liquid chromatography (C4 RP-HPLC) as described by Schmelzer et al. (1992), supra, and Burton et al. (1992), supra. The fractions were selected and combined to obtain a composition of NGF that was substantially free of the modified variants of NGF, for example, the charged species such as the oxidized, deamidated and isoasp NGF species. The previous HIC column can not effectively remove other variant forms of NGF such as oxidized NGF and isoasp. The HIC column does not effectively eliminate misfolded and glycosylated proteins, which bind stronger to the HIC resin than properly folded NGF, since these tend to be more hydrophobic. Accordingly, the cation exchange resin, eg, HP from SP-Sepharose, was used to remove altered charge variants not eliminated by the HIC resin. The SP-Sepharose pool typically contained approximately 92% of 118, 4.6% of 120 1% of deamidated NGF, 1% of oxidized NGF, and 1% of NGF isoasp. Routinely, the amount of each species is in the range of approximately 85 to 93% for 118, 0 to 5% for 120, 0 to 5% for 117, 0 to 3% for deamidated forms, 0 to 2% for forms isoasp, and 0 to 2% for the oxidized forms. The purity of NGF (all species) is routinely greater than 99.5%.

Formulation The HP combined SP-Sepharose was prepared for the formulation by ultrafiltration / diafiltration in a formulation buffer. An acid buffer, preferably acetate at pH 5, is preferably used as discussed above. The NGF 118/118 composition is substantially free of NGF variants and is substantially pure NGF. The formulated material is suitable for the treatment of neuronal disorders, particularly peripheral neuropathy associated with diabetes and peripheral sensory neuropathy associated with AIDS.

Example II Purification of the NGF 120/120 Homodimer on a Large Scale Cropped Cell Culture Fluid HCCF was obtained from a culture of CHO cells of 12,000 liters in general as described in Example I. The distribution of the NGF species in the HCCF was approximately 40-65% 120/120 homodimer with the heterodimer 120/118 with the rest that is homodimer 118/118. The medium was typically processed quickly to minimize the proteolytic conversion of. 120 to 118.

Macroprep High S Cation Exchange Chromatography The HCCF was loaded onto a Macroprep High S Cation Exchange Chromatography column, washed with 1.5 M sodium acetate, 50 M HEPES pH 7. The bound NGF was eluted with 1.5 M sodium chloride, TMAC 0.25 M, 0.2% thiodiglycol, pH 7. The Macroprep column can be run at pH 5-8 with the adjustment of the acetate concentration. Chloride is a preferred substitute for the acetate ion. The NGF eluted due to the TMAC gradient. TMAC is a salt that has ionic and hydrophobic character, which is a useful property, since the support of the spine of some resins contains hydrophobic material that promotes nonspecific interactions between NGF and the resin. Typically, for a pH elution buffer of about 6 to 8, a TMAC concentration of 0-3 M is useful. Fractions containing NGF were combined.

Chromatography on Silica Gel The combined was directly applied to a chromatography column on silica gel. The silica provides a mixed-mode chromatography resin, which has ionic, polarity, and hydrophobic interactions that play an important role in the binding characteristics of the protein. The column was equilibrated in sodium chloride, IMM, 25 mM MOPSO, pH 7. The column was washed with 1 M sodium chloride, 25 mM MOPSO, pH 7 (preferably from about pH 5.0 to 8.5, more preferably pH 6 to 8, and still more preferably pH 7). The bound NGF was eluted with 25 mM succinate, pH 3.9, 50 mM TEAC (tetraethylammonium chloride). TEAC is a more powerful eluent than TMAC. The pH is preferably in the range of pH 3.5 to 8. However, a pH above 7.5 for prolonged periods of time should be avoided in order to prevent or reduce the formation of the deamidated species of NGF. In general, the lower the pH of the buffer, the lower the concentration of the mixed property salt, eg, TMAC or TEAC, required to elute NGF from the silica column. Shock absorbers that have a good buffer capacity close to pH 4 to 5 are suitable for use. The presence of the salt in the elution buffer is optional, such that the column can be washed with an unsalted MOPSO buffer prior to the application of the elution buffer.

Rapid Flow Chromatography in Phenyl-Sepharose Fractions containing NGF were identified and combined. The combined was adjusted to 0.7 M acetate, pH 7, 25 mM MOPSO. The adjusted mixture was loaded onto a column of Phenyl Sepharose Fast Flow Chromatography equilibrated with buffer A in gradient (0.7 M acetate, 25 mM MOPSO, pH 7). The column was washed with a gradient of 90% buffer A (0.7 M acetate, 25 M MOPSO, pH 7) at a 10% gradient of buffer B (25 mM MOPSO, pH 7.20% propylene glycol). Other glycols can be substituted, such as hexylene glycol. Typically the wash was about 2 to 3 CV or until a stable optical density of baseline is achieved. The wash removed some of the proteins from the host cell. The bound NGF was eluted with a linear gradient of 10 CV from a mixture of 90% buffer in gradient A and 10% buffer in gradient B to a mixture of 10% buffer in gradient A and 90% buffer in gradient B. Sodium chloride or sodium sulfate can replace acetate in HIC buffers. The pH is preferably from about 5 to 8, more preferably from about 5.5 to 7.5, with acceptable pH 6 to 8, and more preferably from about 7. The column separated any remaining precursor sequences, partial precursor sequences, or glycosylated forms, present as a homodimer or a heterodimer of a mature NGF monomer and an NGF monomer that still has part of the precursor sequence present. The precursor and glycosylated forms of NGF are present at the leading edge of the elution peak, such that fractions containing NGF were combined to substantially exclude these species. The HIC blend contained approximately 72% monomer 120, 17% monomer 118, 2.8% monomer 117, 3.6% monomer R120, 0.8% isoasp forms, 1.3% oxidized forms, and 1% forms deamidated, as separated and detected on an analytical HPLC system.

HP Chromatography of SP-Sepharose Fractions containing NGF from the HIC step were combined. On a large scale this was achieved by directing the effluent from the column at the appropriate time to a combination tank (holding tank). The pH of the mixture was adjusted to pH 6, and applied to an SP-Sepharose HP chromatography column. The column was washed with 20 mM succinate, 0.2 M sodium chloride, pH 6 (buffer gradient A). The bound NGF was eluted with a gradient of 22 CV starting with a mixture of 70% buffer in gradient A and 30% buffer in gradient B to a final mixture of 80% buffer in gradient B (0.7 M NaCl / pH 6 ) and 20% buffer in gradient A. The pH is preferably from pH 5 to 8, more preferably pH from 5.7 to 6.5, and still more preferably pH 6. A representative chromatogram is shown in Figure 1. The combination of HP SP -Sepharose routinely contained approximately 95% of the 120 form, 3% of the R120 form, 0.65% of the isoasp form, 0.6% of the oxidized form, 0.6% of the deamidated form. Other unidentified forms of NGF were approximately 0.6%, and comprised dioxidized NGF (Met 37 and Met 92) with a deamidated Asn45 present. An HPIEX analysis comparing a representative HIC pool (loaded on the SP-Sepharose resin) and the pool obtained after SP-Sepharose chromatography is shown in Figure 2. Each of the three main cut-out forms of NGF , 120, 118 and 117, may have variants, but variants, such as the oxidized and isoasp forms, of the predominant cut form during a purification (120 in this example) may mask the analysis of the variants from the less predominant forms (118 and 117 in this example). An HPLC analysis of the fractions from a representative run is shown in Figure 3. HP from SP-Sepharose effectively eliminated variants present in the HIC pool. The R120 form has an additional arginine residue at the N-terminus of NGF; usually the N-terminal amino acid sequence of rhNGF in SSSHP, but R120 has an N-terminal sequence of RSSSHP. In this way the R120 form is more basic than the mature NGF and separated by SO-SHP. It also has lower bioactivity, probably related to the fact that the N-terminal NGF is necessary for the receptor binding (trkA). The oxidized NGF form is a mono-oxidized form having the methionine in the oxidized position 37, producing a more acidic form eluting over the leading edge of the NGF peak. The isoasp form contains a modification of the aspartic acid in amino acid 93. The isoasp form is slightly more basic and thus binds slightly stronger to the HP resin of SP-Sepharose. The NGF species containing isoasp 93 eluted at the trailing edge of the elution peak. Deamidation occurs in asparagine residues, typically in asparagine at position 45. NGF containing deamidated Asn, which produces an Asp at position 45, is slightly more acidic, appearing at the leading edge of the elution peak. Fractogel EMD S03 is a preferred alternative resin to the HP resin of SP-Sepharose for the separation of the charged variants of the NGF species. When this less preferred resin is used, higher concentrations of sodium chloride are needed to elute NGF.

Formulation The bulk material was formulated by UF / DF in the buffer of the formulation as in the previous Example. In the final bulk product, the form 120 was routinely in the range of about 92 to 97%, the form R120 of about 1 to 4%, the isoasp form of about 0.2 to 1.5%, the oxidized form of about 0.2 to 2. %, and the deamidated form of approximately 0.2 to 2%. Forms 117 and 118 were routinely less than about 2%. The final bulk product was routinely at least 99.5% pure NGF (including all species).

Example III Isolation from 118/118 In a preferred embodiment to obtain a substantially pure composition of NGF 118/118 that is substantially free of the NGF variants, the method of Example II was followed with the following modifications. An immobilized trypsin column is used between the Macroprep High S column and the silica column. The Macroprep cocktail is directly loaded onto the immobilized trypsin column, after adjusting the pH to between about pH 5 to 8.5, more typically 6.5 to 7.5, if necessary. The pool was passed through the column, during which time most of the NGF was converted to form 118. Protease digestion converts form 120 to form 118 by cleavage of the C-terminal VRRA to VR. To achieve limited and selective cleavage, a trypsin or trypsin-like protease is used, preferably trypsin, more preferably readily available porcine trypsin, or alternatively bovine trypsin or a recombinant trypsin. Any proteolytic method that provides substantially limited and selective cleavage can be used, but an immobilized trypsin column is preferred in order to minimize contamination of the NGF preparation. The column is run at a pH that leads to protease activity, preferably pH 5.5 to 8.5, more preferably 6.0 to 8.0, and still more preferably 6.5 to 7.5. In this example, the glycosylated NGF was removed by HIC as discussed herein. After an HP step of SP-Sepharose as discussed herein in Example II, but preferably using 22 volumes of salt gradient column from 0.3M to 0.55M, a preferred composition of NGF was obtained for clinical use. A composition with more than 70% of the monomer 118, less than 10% of the monomer 120, and less than 15% of the monomer 117 can be obtained, as determined by RP-HPLC assays. Typically, compositions that are greater than or equal to 90% rhNGF 118/118, more usually greater than or equal to 93% rhNGF 118/118 with less than or equal to about 7% of the deamidated isoAsp and oxidized variants, They are obtained. A means to achieve the highest purity is to avoid selection of the fractions with significant amounts of the variants, such as can be found at the leading or trailing edges of the main neurotrophin peak, for example the peak of rhNGF 118/118.

Example IV. Partial Purification and Refolding of rhNT-4/5 From Bacterial Inclusion Bodies In this example, starting with 10 or 60 liters of fermentation, rhNT-4/5 was purified. The host used to produce the recombinant human NT-4/5 in the fermentation described in this example was a strain of E. col i designated 27C7 / pmNT5DT; although NT-4/5 produced from other strains and organisms is suitable for the purification process described herein. The expression plasmid used in this example contained the mature NT-4/5 coding sequence under the transcriptional and translational control sequences required for the expression of the? T-4/5 gene in E. coli In the plasmid expressing? T-4/5, the transcriptional sequences used for expression in E. col i were provided by the alkaline phosphatase promoter sequence. The lambda site to the transcriptional terminator was located adjacent to the stop codon of? T-4/5. Secretion of the protein from the cytoplasm was directed by the STII signal sequence. Most rh? T-4/5 was found in the cellular periplasmic space as shrinking bodies. The plasmid conferred tetracycline resistance after the transformed host. The fermentation process was carried out at 35 ° C-39 ° C and pH 7.0-7.8. The fermentation was allowed to proceed for 25-40 hours, at which time the culture was cooled before harvesting. The culture was inactivated by thermal treatment using a continuous flow apparatus at 60 ° C or using thermal inactivation in tank at that temperature for 5-15 minutes. The culture inactivated by heat was centrifuged using an AX Alf-laval centrifuge or equivalent. The cells of E. coli were recovered in the button or pella. The cells of E. col i, which express the recombinant human NT-4/5 in inclusion bodies, were lysed by standard means to prepare a paste containing NT-4/5 in the inclusion bodies. Protease inhibitors were not included in the buffer. To isolate inclusion bodies from cellular waste, NT-5 paste from E. coli was resuspended in Tris 0.02 M, pH 8, 5 mM EDTA (10 ml of buffer / gram of paste) using a rotary mechanical dispersion device, for example a Turrax. The cell suspension was passed through a microfluidizer three times at 421.84 kg / cm2 (6,000 psi). The resulting homogenate was centrifuged in a Sorvall RC-3B centrifuge at 5000 rpm for approximately 45 minutes. The supernatant was discarded and the button was resuspended in 20 mM Tris, pH 8, 5 mM EDTA (Extraction buffer) using a turrax apparatus for 2 to 3 minutes at medium speed. The homogenate was centrifuged as described above. The button was resuspended in extraction buffer and centrifuged as described above. The resulting button (s) (referred to as the NT-4/5 inclusion bodies or retractable bodies) was stored at -70 ° C. NT-4/5 was isolated from inclusion bodies as follows. Buttons of the inclusion bodies were suspended in 20 mM Tris, pH 8, 6 M urea, 25 M DTT (10 ml buffer / gram inclusion body) using a turrax apparatus at medium speed for approximately 10 minutes. The suspension was stirred for 40 minutes at 2-8 ° C and centrifuged in a Sorvall RC3B apparatus at 5000 rpm for 45 minutes. PEI (polyethylenimine) was added to 0.1% of the supernatant, which was stirred at 2-8 ° C for 30 minutes. PEI precipitates nucleic acid and other acid-charged molecules. The mixture was centrifuged in a Sorvall RC3B at 5000 rpm for approximately 45 minutes. The supernatant of PEI was loaded on a Fast Flow column of DEFF Sepharose (10 cm x 14 cm; DEFF is a diethylaminoethyl resin) equilibrated in 0.02 M Tris, 6 M urea, 10 mM DTT, pH 8. One equivalent of 1 kg of the solubilized shrinkable bodies was loaded onto the DEFF column. Since the reduced and denatured NT-4/5 does not bind to the DEFF resin, the combined side-to-side flow containing NT-4/5 and 6 M urea was collected (Figure 6), and the pH of the combined was decreased to 5.0 with acetic acid. The combined flow of DEFF with adjusted pH was loaded on a S-Sepharose Rapid Flow column (S refers to the functional group S03 on the resin) equilibrated in 20 mM acetate, pH 5, containing 6 M urea, under whose conditions the NT-4/5 is bonded to the resin. After loading, the S-Sepharose Rapid Flow column was washed with various volumes of equilibrium buffer column. Bound NT-4/5 was eluted with 0.5 M sodium chloride, 20 mM sodium acetate, 6 M urea, pH 5 (Figure 7). The combined SSFF of 0.5 M sodium chloride was dialyzed overnight against 20 mM Tris, 0.14 M sodium chloride, pH 8, conditions that allow NT-4/5 to replicate although incorrectly. The misfolded rhNT-4/5 molecules were added to form a precipitate. The rhNT-4/5 badly folded, aggregate was processed to obtain the correctly folded NT-4/5. The unfolded, aggregated rhNT-4/5 was collected by centrifugation as a button or pellet. The button was resuspended in 0.2 M Tris, pH 8, 4 M urea, 5 mM DTT and stirred at 2-8 ° C for approximately 1 to 2 hours or until the button was dissolved. The final protein concentration was adjusted to approximately 10 mg / ml protein based on the extinction coefficient 1.8 at 280 nm. Oxidized glutathione was added to the solution of the solubilized button, to a final concentration of 20 mM, followed by gentle agitation for 15 to 30 minutes at 2-8 ° C. The oxidized glutathione reacts with the sulfhydryl groups of NT-4/5 to produce the disulfide mixed with NT-4/5-S-glutathione. The disulfide mixed with NT-4/5-SG was diluted to a final concentration of 0.1 to 0.5 mg / ml protein in 100 M Tris, 20 M glycine, 15% PEG 300, Guanidine-HCl IM, pH 8.3. To initiate adequate refolding of NT-4/5, cysteine was added to the refolding mixture, at a concentration of 2 to 4 mM, followed by aeration (by bubbling) the solution with nitrogen or helium for 5 to 60 minutes before sealing the container to exclude oxygen. The refolding of NT-4/5 was allowed to proceed for 18 to 24 hours at 2-8 ° C.

Alternatively, rhNT-4/5 was refolded using sulfitolysis as follows. Buttons of the inclusion bodies (110 g) were suspended in 1.1 liter of 20 mM Tris, 7 M urea, 10 mM glycine, 100 mM sodium sulfite, 10 M sodium tetrathionate, and solubilized using a turrax apparatus by 10 minutes at medium speed. The mixture (1260 ml) was then stirred at 2-8 ° C for 45 minutes. The PEI was added to a final concentration of 0.1% PEI. The mixture was stirred for an additional 30 minutes at 4 ° C and centrifuged for 45 minutes at 5500 rpm in an RC3B centrifuge. The supernatant was loaded on a DEFF column (4.4 cm x 25 cm) equilibrated in 20 mM Tris, 6 m urea, pH 8. The complete flow of DEFF was adjusted to pH 5 with acetic acid and loaded onto a Fast Flow column. of S-Sepharose (4.4 cm x 25 cm) equilibrated with 20 mM acetate, 6 M urea, pH 5. The NT-4/5 was eluted with 25 mM MOPSO, 0.5 M sodium chloride, pH 7. The combined Sodium chloride SSFF 0. 5 M containing the sulphonylated rhNT-4/5, was diluted to approximately 0.1 mg / ml protein and adjusted to 1 M guanidine hydrochloride, 100 mM Tris, 20 mM glycine, 15% PEG 300, pH 8.3. The refolding of NT-4/5 was initiated by the addition of cysteine 2 to 4 mM. The refolding reaction was essentially complete within 24 hours. Aeration with an inert gas, for example helium or nitrogen, to replace the oxygen in the solution, may optionally be carried out.

Example V. Isolation of rhNT-4/5 Correctly Folding of Conformational Variants (Wrongly folded) The refolding mixture of rh.NT-4/5 of Example IV was dialyzed against a solution of pH 4 to 5 overnight to eliminate guanidine and other reagents. To clarify the solution, the solution was centrifuged either for 45 minutes at 5000 rpm or passed through a 0.2 μm filter. The clarified supernatant, containing 0.5 to 5 grams of protein, was adjusted to pH 3 to 5 by the addition of glacial acetic acid and either loaded onto a CR RP-HPLC column or stored frozen at -20 ° C until he was ready for purification. In this example, the acidified and rinsed solution was loaded onto a CR R-HPLC column (3 cm x 50 cm), to which resin was bound in the folded rhNT-4/5. The properly folded NT-4/5 was eluted using a gradient of acetonitrile in a 0.05% trifluoroacetic acid (TFA) solvent system: a gradient of 26 to 40% acetonitrile (over a period of 95 minutes) in trifluoroacetic acid 0.05% at a flow rate of 25 ml / minute. The fractions were collected at intervals of 1 to 1.5 minutes (Figure 8). Fractions were analyzed for correctly folded NT-4/5 by comparing the elution time on a Vydac CR HPLC analytical column (0.21 x 15 cm) to that of a correctly folded NT-4/5 standard (Figure 9). ). The NT-4/5 intact, correctly folded, standard, eluted typically at 19 minutes at a flow rate of 2.5 ml per minute, with a buffer system of 0.5% TFA / acetonitrile. The fractions containing the correctly folded rhNT-4/5 were combined and the pH was adjusted to pH 5 to 7. This combination of the correctly folded rhNT-4/5 also contained the carbamyllated and trimmed forms at the N-terminus of NT-4 /5. The reverse phase preparative liquid chromatography resin is preferably a medium having a particle diameter of about 10 to 40 microns, a pore size of about 200 to 400 Angstroms, and an alkyl group of 4, of 8 or of 18 atoms of carbon. More preferably, the resin has a particle diameter of about 15 to 40 microns and a pore size of about 300 Angstroms, and is a C4 silica medium.

Example VI. An Alternative Isolation of rhNT-4/5 Correctly Folding the Conformational Variants (Wrongly folded) The rhNT-4/5 refolded mixture of Example IV was concentrated to approximately 1 tenth using a Millipore-Pellicon ultrafiltration system with a cellulose membrane of 1.86 m2 (20 square feet) (or polysulfone or equivalent) with a molecular weight cutoff. of 10 kD. The concentrated mixture was dialyzed either overnight against 50 liters of 50 mM acetate, pH 5.5, 50 mM sodium chloride or diafiltered in 50 M acetate, 50 mM sodium chloride, pH 5.5, before filtering through a 0.2 micron membrane. The filtered refolding mixture was adjusted to 2.5 M sodium chloride, 20 mM MOPSO, pH 7 and loaded onto a HIC column, a Toyopearl 650 M phenyl column (10 cm x 19 cm), previously equilibrated in 2.5 M NaCl. , 20 mM MOPSO, pH 7. The column was then washed with equilibration buffer. Some misfolded forms of the rhNT-4/5 molecule eluted in the full-flow fractions, while other misfolded forms were eluted at high concentrations of organic solvents such as 20-40% reactive alcohol. The rhNT-4/5 correctly folded was eluted from the phenyl column using 2 M sodium chloride, 10% reagent alcohol, pH 7 (Figure 10). Other phenyl resins such as phenyl-Sepharose can be used in place of the Toyopearl backbone. The salts discussed herein, including ammonium sulfate, citrate, acetate, and potassium chloride may also be used. Depending on the salt used, the salt concentration is typically 1 M to 3 M, with 2.5 M sodium chloride which is preferred for the charge and 2 M sodium chloride which is preferred for elution when the organic solvent is present. Preferably, a decrease in the salt concentration is used to elute and separate a neurotrophin and its variants. In order to achieve elution, the salt concentration in the elution buffer is typically less than that in the charge buffer, but this may be the same concentration when compensated with the organic solvent. In addition, the use of organic solvent has another advantage, as has been found in the present, that the addition of an organic solvent improves the elution pattern resulting in narrower peak profiles. In addition to ethanol, other organic solvents discussed herein may be used, including propanol, isopropanol, and lower alkylene glycols such as propylene glycol, ethylene glycol, and hexylene glycol. The organic solvent at 5-25% (v / v), more preferably 5 to 20% (v / v), even more preferably 5 to 15% (v / v) will typically elute a properly folded neurotrophin. Elution with organic solvent can be either gradient or gradual. The pH range is preferably close to neutral to slightly acidic, from pH 5 to 8, more preferably pH 5.5 to 7.5, and still more preferably pH 7. Any of the buffers discussed herein, including MOPSO, MOPS, HEPES, phosphate , citrate, ammonium, acetate, can be used as long as they buffer to the desired pH.

Example VII. Purification of rhNT-4/5 Correctly Folding the Chemical Variants The separation of the intact, correctly folded rhNT-4/5 from its chemical variants, including the carbamylated forms and cut at the N-terminus of rh.NT-4/5, was achieved by high-resolution cation exchange chromatography using the resin of SP-Sepharose HP or PolyCat A HPLC resin. When the C4 RP-HPLC column was used to remove misfolded variants, the combined C4 HPLC was adjusted to pH 5 to 7 and loaded onto an HP column of SP-Sepharose 7 cm x 19 cm, equilibrated in 20 M succinate, pH 6, 5% reactive alcohol, 0.2 M sodium chloride. The resin with bound NT-4/5 was washed with equilibrium buffer. The bound rhNT-4/5 was eluted and separated from the trimmed N-terminal and carbamylated forms using 22 column volumes (CV) in gradient from 0.2 M sodium chloride to 0.4 M sodium chloride at pH 6 (e.g. the saline gradient in the equilibrium buffer) (Figure 11). The fractions containing NT-4/5 were combined and formulated in 0.05 M acetate, pH 4 to 5. The intact rhNT-4/5 was identified and distinguished from the variants in the fractions preferably by analytical RP-HPLC, or by SDS-PAGE, as discussed herein, in comparison to the standard. Alternatively, the variant forms of NT-4/5 were removed by high-resolution HPLC cation exchange chromatography on a polyaspartic acid column (PolyCat A, PolyLC, Columbia Md) (9.4 x 200 cm) (Figure 12). The combined HPLC C4 was adjusted to pH 5 to 6 and then loaded onto a PolyCat A column. The chromatography conditions were: buffer A was 20 M phosphate, 5% acetonitrile, pH 6; buffer B was 20 mM phosphate, 5% acetonitrile, 0.8 M potassium chloride, pH 6. The rhNT-4/5 was eluted using a gradient of 25 to 60% of Buffer B in 65 minutes (Figure 12), Fractions were collected at 1 minute intervals and analyzed by C4 analytical HPLC as described above. When HIC was used to eliminate misfolded variants (Example VI above), the correctly folded NT-4/5 combination was dialyzed overnight in 20 M succinate, 0.1 M sodium chloride, 5% reagent alcohol, pH 6 or ultrafiltered / diafiltered in the 20 M succinate buffer. The combined UF / DF or dialysate was then loaded onto an HP column of SP-Sepharose or PolyCat A as described above.

Formulation Fractions containing the intact NT-4/5, correctly folded (from HP step of SP-Sepharose or HPLC PolyCat A.) Were combined and concentrated to 1-5 mg / ml in 20 mM acetate, pH 4 to 5 of formulation buffer. Alternatively, NT-4/5 was formulated using ultrafiltration / diafiltration. The final bulk solution was analyzed by amino acid analysis, N-terminal sequence analysis, mass spectrometry, SDS-PAGE (Figure 13) and biological assays. A kinase receptor activation assay (KIRA), which detects NT-4/5 activation of the autophosphorylation of its receptor tyrosine kinase (trkB) located on a cell membrane, was used to characterize the rh. T-4/5 purified. CHO cells expressing trkB with a gD tag were used. International Patent WO 95/14930, published June 1, 1995, describes the KIRA test and is incorporated by reference herein. RhNT-4/5 had an EC50 of 12.6 ng / ml in this assay. Typically the EC50 of intact NT-4/5, correctly folded, purified as described herein, is 5 to 30, more preferably 10 to 20. The purity of rhNT-4/5 with respect to proteins other than NT- 4/5 was typically 90 to 99%. The homogeneity of NT-4/5 with respect to the carbamylated variants and cut at the N-terminus was from 90 to 99%. More typically, and preferably, the purity and homogeneity are 99% or greater.

Example VIII. Initial Purification, Refolding and Final Purification of rhNT-3 of Bacterial Inclusion Bodies To isolate inclusion bodies from cellular debris, NT-3 paste from E. coli (1 kg) was resuspended in 10 liters of 100 mM sodium acetate, pH 5, using a rotary mechanical dispersion device, for example a turrax apparatus. The cell suspension was passed through a microfluidizer three times at 421.84 kg / cm2 (6000 psi). The resulting homogenate was centrifuged in a Sorvall RC-3B centrifuge at 5000 rpm for 30 minutes. The NT-3 was isolated from the inclusion bodies as follows. The buttons or pellets of the inclusion bodies were suspended in 100 mM Tris, 100 mM sodium chloride, 5 mM EDTA, 100 mM sodium sulfite, 10 mM sodium tetrathionate, 7.5 M urea, pH 8.3 (10 ml / gram of the inclusion body) using a turrax apparatus at medium speed for approximately 10 minutes. The suspension was stirred for approximately one hour at 2-8 ° C. PEI was added (polyethyleneimine) at about 0.15% (final concentration) and stirred at 2-8 ° C for 30 minutes. The mixture was centrifuged in a Sorval RC3B apparatus at 5000 rpm for approximately 30 minutes. The supernatant was filtered with a Gelman Preflow cartridge. The filtered supernatant was diluted with 3 volumes of equilibrium buffer of S-Sepharose Fast Flow (50 mM sodium acetate, urea 5 M, pH 5). The diluted filtered supernatant (conductivity less than 7 mS) is loaded onto a S-Sepharose FF column equilibrated with 50 M sodium acetate, 5 M urea, pH 5.0. The column was first washed with 50 mM sodium acetate, 5 M urea, pH 5, followed by 50 mM MOPS, 5 M urea, 10 mM glycine, pH 7.0. The sulfitolized NT-3 was eluted from the column using a gradient of 10 column volumes from 0 to 0.6 M NaCl to 50 mM, in 50 mM MOPS, 5 M urea, 10 mM glycine, pH 7. The partially purified NT-3 was refolded by diluting the combined S-Sepharose FF to approximately 0.1 mg / ml protein in the refolding buffer containing 0.1 M Tris, 2 M urea, 0.1 M sodium chloride, 15% PEG 300, 10 mM glycine , 25 mM ethanolamine, pH 9.1. Refolding was initiated by the addition of cysteine at approximately 5 M and agitated for 2 to 5 days at 2-8 ° C. Optionally, the refolding buffer can be purged with helium or argon to reduce the concentration of oxygen in the refolding solution. The pH of the refolded cocktail was adjusted to pH 7, filtered and loaded onto a Macroprep High S cation exchange chromatography column equilibrated in 50 mM HEPES, pH 7. After loading the refolding combination with pH adjusted to the Macroprep column, the column was first washed with 50 mM MOPS, pH 7 followed by 50 M MOPS, 0.1 M TMAC, 0.3 M NaCl, pH 7. 'NT-3 was eluted with 50 mM MOPS, 0.25 M TMAC, 1.5 M NaCl, pH 7 The Macroprep blend was then further purified on the Phenyl-Sepharose High-Flow Fast Substitution column. The phenyl column was equilibrated in 50 mM HEPES, 1.5 M NaCl, pH 7 and the combined macroprep was directly loaded onto the phenyl column. The column was washed with equilibrium buffer and then the correctly refolded NT-3 was eluted using a gradient of 15 column volumes going from 50 mM HEPES; 1.5 M sodium chloride, pH 7 to 50 mM HEPES, 10% reagent alcohol, pH 7. The fractions were analyzed either by C4 HPLC or by SDS-PAGE and the fractions containing the correctly folded NT-3 were combined. The phenyl pool was diluted to less than 25 mS (typically about 2 volumes with water) and loaded onto an HP column of SP-Sepharose previously equilibrated in 25 mM MOPSO, pH 7. The column was first washed with equilibrium buffer and NT-3 was eluted from the column using a gradient of 20 column volumes ranging from 0.35 M TMAC to 0.65 M TMAC in 25 mM MOPSO, pH 7. Fractions containing rhNT-3 (as judged by the C4 HPLC) were combined. The HP-combination of SP-Sepharose was concentrated to approximately 1 mg / ml in a membrane of 10,000 molecular weight and then diafiltered with 100 volumes of 10 mM acetate, 140 mM NaCl, pH 5.0.

LIST OF SEQUENCES GENERAL NFORMATION: (i) APPLICANT: Genentech, Inc. (ii) TITLE OF THE INVENTION: Purification of Neurotrophins (iii) SEQUENCE NUMBER: 6 (iv) ADDRESS FOR CORRESPONDENCE: (A) RECIPIENT: Genentech, Inc. (B) STREET: 1 DNA Way (C) CITY: South San Francisco (D) STATE: California (E) ) COUNTRY: USA (F) ZIP: 94080 (V) COMPUTER LEGIBLE FORM: (A) TYPE OF MEDIA: 3.5 inch flexible disk, 1.44 Mb (B) COMPUTER: IBM compatible PC (C) OPERATING SYSTEM: PC-DOS / MS-DOS (D) SOFTWARE: WinPatin (Genentech) i) DATA OF THE CURRENT APPLICATION (A) APPLICATION NUMBER: (B) DATE OF PRESENTATION: (C) CLASSIFICATION: (vii) DATA OF THE APPLICATION P-REVIA: (A) NUMBER OF APPLICATION: 60/030838 (B) DATE OF PRESENTATION: 15 -NOV- 1996 (vii) DATA FROM THE PRIOR APPLICATION: (A) APPLICATION NUMBER: 60/047855 (B) DATE OF SUBMISSION: MAY 29, 1997 (viii) INFORMATION D. THE LAWYER / AGENT: (A) NAME: Torchia, PhD., Timothy E. (B) REGISTRATION NUMBER, '36,700 (C) REFERENCE NUMBER / CASE: P1063R2PCT ix) INFORMATION FOR COMMON TION: (A) TELEPHONE: 650 / 225-8674 (B) FAX: 650 / 952-9881 (2) INFORMATION FOR SEQ ID NO: 1: (i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 241 amino acids (B) TYPE: Amino Acid (D) TOPOLOGY: Linear ix) FROM SCRI PC OF THE SECUENC IA: SEQ I D NO: 1 Met Ser Met Leu Phe Tyr Thr Leu lie Thr Ala Phe Leu lie Gly 1 5 10 15 lie Gln Ala Glu Pro His Ser Glu Ser Asn Val Pro Ala Gly His 20 25 30 Thr lie Pro Gln Val Hxs Trp Thr Lys Leu Gln His Ser Leu Asp 35 - 40 45 Thr Ala Leu Arg Arg Ala Arg Be Ala Pro Ala Ala Ala Ala lie Ala 50 55 60 Wing Arg Val Wing Gly Gln Thr Arg Asn lie Tlir Val Asp Pro Arg 65 70 75 Leu Phe Lys Lys Arg Arg Leu Arg Ser Pro Arg Val Leu Phe Ser 80 85 90 Thr Gln Pro Pro Arg Glu Wing Wing Asp Thr Gln Asp Leu Asp Phe 95 100 105 Glu Val Gly Gly Ala Ala Pro Phe Asn Arg Thr His Arg Ser Lys 110 115 120 Arg Ser Ser Ser His Pro lie Phe His Arg Gly Glu Phe Ser Val 125 130 135 Cys Asp Ser Val Ser Val Trp Val Gly Aep Lys Thr Thr Wing Thr 140 145 150 Asp He Lys Gly Lys Glu Val Val Met Val Leu Gly Vallu Asn He 155 160 165 Asn Asn Ser Val Phe Lys Gln Tyr Phe Phe Glu Thr Lys Cys Arg 170 175 180 Asp Pro Asn Pro Val Asp Ser Gly Cys Arg Gly He Asp Ser Lys 185 190 195 His Trp Asn Ser Tyr Cys Thr Thr Thr His Thr Phe Val Lys Wing 200 205 210 Leu Thr Met Asp Gly Lys Gln Wing Wing Trp Arg Phe He Arg He 215 220 225 Asp Thr Ala Cys Val Cys Val Leu Ser Arg Lys Ala Val Arg Arg 230 235 240 Wing 241 (2) INFORMATION FOR SEQ ID NO: 2: (i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 120 amino acids (B) TYPE: Amino Acid (D) TOPOLOGY: Linear (ix) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 2 Being Ser His Pro He Phe His Arg Gly Glu Phe Ser Val Cys 1 5 10 15 Asp Ser Val Ser Val Trp Val Gly Asp Lys Thr Thr Wing Thr Asp 20 25 - 30 He Lys Gly Lys Glu Val Met Val Leu Gly Val Glu Asn He Asn 35 40 45 Asn Ser Val Phe Arg Gln Tyr Phe Phe Glu Thr Lys Cys Arg Asp 50 55-60 Pro Asn Pro Val Asp Ser Gly Cys Arg Gly He Asp Ser Lys His 65 70 75 Trp Asn Ser Tyr Cys Thr Thr Thr His Thr Phe Val Lys Ala Leu 80 85 90 Thr Met Asp Gly Lys Gln Wing Wing Trp Arg Phe He Arg He Asp 95 100 105 Thr Wing Cys Val Cys Val Leu Ser Arg Lys Wing Val Arg Arg Wing 110 115 120 (2) INFORMATION FOR SEQ ID NO: 3: (i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 120 amino acids (B) TYPE: Amino Acid (D) TOPOLOGY: Linear (ix) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 3 Being Ser Thr His Pro Val Phe His Met Gly Glu Phe Ser Val Cys 1 5 10 15 Asp Ser Val Ser Val Trp Val Gly Asp Lys Thr Thr Wing Thr Asp 20 25 30 He Lys Gly Lys Glu Val Thr Val Leu Wing Glu Val Asn He Asn 35 40 45 Asn Ser Val Phe Arg Gln Tyr Phe Phe Glu Thr Lys Cys Arg Wing 50 55 60 Ser Asn Pro Val Glu Ser Gly Cys Arg Gly He Asp Ser Lys His 65 70 75 Trp Asn Ser Tyr Cys Thr Thr Thr His Thr Phe Val Lys Ala Leu 80"85 90 Thr Thr Asp Glu Lys Gln Wing Wing Trp Arg Phe He Arg He Aep 95 100 105 Thr Wing Cys Val Cys Val Leu Ser Arg -Lys Wing Thr Arg Arg Gly 110 115 120 (2) INFORMATION FOR SEQ ID NO: 4: (i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 118 amino acids (B) TYPE: Amino Acid (D) TOPOLOGY: Linear ix) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: His Ser Asp Pro Wing Arg Arg Gly Glu Leu Ser Val Cys Asp Ser 1 5 10 15 He Ser Glu Trp Val Thr Ala Wing Asp Lys Lys Thr Wing Val Asp 20 25 30 Met Ser Gly Gly Thr Val Thr Val Leu Glu Lys Val Pro Val Ser 35 40 45 Lys Gly Gln Leu Lys Gln Tyr Phe Tyr Glu Thr Lys Cys Asn Pro 50 55 60 Met Gly Tyr Thr Lys Glu Gly Cys Arg Gly He Asp Lys Arg His 65 70 75 Trp Asn Ser Gln Cys Arg Thr Thr Gln Ser Tyr Val Arg Ala Leu 80 85 90 Thr Met Asp Ser Lys Lys Arg He Gly Trp Arg Phe He Arg He 95 100 105 Asp Thr Ser Cys Val Thr Leu Thr He Lys Arg Gly Arg 110 115 118 (2) INFORMATION FOR SEQ ID NO: 5: (i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 119 amino acids (B) TYPE: Amino Acid (D) TOPOLOGY: Linear (ix) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 5 Tyr Ala Glu His Lys Ser His Arg Gly Glu Tyr Ser Val Cys Asp 1 5 10 - 15 Ser Glu Ser Leu Trp Val Thr Asp Lys Ser Ser Ala He Asp He 20 25 30 Arg Gly His Gln Val Thr Val Leu Gly Glu He Lys Thr Gly Asn 35 40 45 Ser Pro Val Lye Gln Tyr Phe Tyr slu Thr Arg Cyß Lys Glu Wing 50 55 60 Arg Pro Val Lys Asn Gly Cys Arg Gly He Asp Asp Lys His Trp 65 70 75 Asn Ser Gln Cys Lys Thr Ser Gln Thr Tyr Val Arg Ala Leu Thr 80 85 90 Ser Glu Asn Asn Lys -Leu Val Gly. Trp Arg Trp He Arg He Asp _ 95"100 105 Thr Ser Cys Val Ser Ala Leu Ser Arg Lys He Gly Arg Thr 110 115 119 (2) INFORMATION FOR SEQ ID NO: 6: (i) CHARACTERISTICS OF THE SEQUENCE: (A) LENGTH: 130 amino acids (B) TYPE: Amino Acid (D) TOPOLOGY: Linear [ix) DESCRIPTION OF THE SEQUENCE: SEQ ID NO: 6: Gly Val Ser Glu Thr Wing Pro Wing Being Arg Arg Gly Glu Leu Wing 1 5 10 15 Val Cys Asp Wing Val Ser Gly Trp Val Thr Asp Arg Arg Thr Wing 20 25 30 Val Asp Leu Arg Gly Arg Glu Val Glu Val Leu Gly Glu Val Pro 35 40 45 Wing Wing Gly Gly Pro Pro Leu Arg Gln Tyr Phe Phe Glu Thr Arg 50 55 60 Cys Lys Wing Asp Asn Wing Glu Glu Gly Sly Pro Gly Wing Gly Gly 65 70 75 Gly Gly Cys Arg Gly Val Asp Arg Arg His Trp Val Ser Glu Cys 80 85 90 Lys Wing Lys Gln Ser Tyr Val Arg Wing Leu Thr Wing His Wing Gln 95 100 105 Gly Arg Val Gly Trp Arg Trp lie Arg He Asp Thr Wing Cys Val not 115 120 Cys Thr Leu Leu Ser Arg Thr Gly Arg Ala 125 130 It is noted that in relation to this date, the best method known to the applicant to carry out the aforementioned invention, is that which is clear from the present description of the invention.

Having described the invention as above, property is claimed as contained in the following:

Claims (29)

1. A process for isolating a recombinant human neurotrophin, or the genetically engineered form thereof, from a mixture containing other proteins and a variant of this neurotrophin which may include a misfolded variant, an incorrectly proteolytically processed variant, or a variant of glycosylation of this neurotrophin, characterized the process because it comprises the separation of the neurotrophin from the other proteins and the neurotrophin variant in the mixture, through the use of a hydrophobic interaction chromatography resin where the separation comprises the charge of the mixing on a hydrophobic interaction chromatography resin and eluting the neurotrophin from the resin with an elution buffer under conditions in which the neurotrophin is separated from the variant.

2. The process according to claim 1, characterized in that the resin comprises a phenyl functional group.

3. The process according to claim 1, characterized in that the mixture loaded on the hydrophobic interaction chromatography resin has a pH of 5 to 8.

4. The process according to claim 1, characterized in that the mixture charged on the resin has a salt concentration of 0.5 M to 3 M.

5. The process according to claim 4, characterized in that the mixture charged on the resin has a salt concentration of 0.5 M to 2.5 M.

6. The process according to claim 5, characterized in that the mixture charged on the resin has a salt concentration of 0.7 M acetate or 1.0 M NaCl at 2.5 M.

7. The process according to claim 1, characterized in that the elution buffer comprises organic solvent.

8. The process according to claim 7, characterized in that the organic solvent is 5% to 20% by volume.

9. The process according to claim 1, characterized in that the pH of the elution buffer is pH 5 to pH 8.

10. The process according to claim 1, characterized in that the elution comprises a decreasing salt gradient.

11. The process according to claim 1, characterized in that the neurotrophin is in the NGF superfamily.

12. The process according to claim 10, characterized in that the neurotrophin is NGF, NT-4/5 or NT-3.

13. The process according to claim 1, characterized in that the neurotrophin is prepared from bacterial culture and refolded in vi tro before using the hydrophobic interaction resin.

14. The process according to claim 1, characterized in that the neurotrophin is isolated from the culture of mammalian cells.

15. The process according to claim 1, characterized in that it further comprises the separation of the neurotrophin from its chemical variants using high resolution cation exchange chromatography resin.

16. The process according to claim 15, characterized in that the separation step by high-resolution cation exchange chromatography comprises charging a mixture comprising the neurotrophin and a chemical variant of that neurotrophin onto a high-cation exchange chromatography resin. resolution, and eluting the neurotrophin from the resin under conditions in which the neurotrophin is separated from the chemical variant, and where the neurotrophin has a high pl.

17. The process according to claim 16, characterized in that the high-resolution cation exchange resin is an SP-Sepharose HP resin, polyaspartic acid, polysulfoethyl cation exchange, or Fractogel EMD S03.

18. The process according to claim 16, characterized in that the eluate containing the neurotrophin from the high resolution cation exchange resin is desalted or subjected to diafiltration and then formulated with a carrier.

19. The process according to claim 1, characterized in that it further comprises the separation of neurotrophin from the other proteins using silica gel resin.

20. A process for isolating a neurotrophin from a mixture of proteins, characterized in that it comprises the separation of the neurotrophin from the others, proteins using a silica gel resin in the absence of an alcoholic or polar aprotic solvent.

21. A process for separating a neurotrophin from a chemical variant of that neurotrophin, characterized in that it comprises the separation of neurotrophin from its chemical variant using high resolution cation exchange chromatography.

22. The process according to claim 21, characterized in that the neurotrophin and its variant are essentially pure before separation.

23. The process according to claim 21, characterized in that the resin is an SP-Sepharose HP resin, polyaspartic acid resin, polysulfoethyl cation exchange resin, or Fractogel EMD S03 resin.

24. The process according to claim 21, characterized in that it also optionally comprises the step of separating the neurotrophin from a misfolded variant of that neurotrophin, using the preparative reverse phase liquid chromatography resin.

25. The process according to claim 24, characterized in that the resin contains a functional group C4.

26. A neurotrophin composition, characterized in that it is prepared by the process according to any of the preceding claims.

27. A composition, characterized in that it comprises a carrier and a neurotrophin that is essentially pure and homogeneous because it is essentially free of its variants.

28. The composition according to claim 27, characterized in that it is sterile.

29. A process for the purification of a neurotrophin, characterized in that it comprises: a) the charge of a mixture comprising the neurotrophin and its variant on a hydrophobic interaction chromatography resin at pH 5 to 8; b) washing the resin with a buffer at pH 5 to 8; c) the elution of the neurotrophin with a buffer at pH 5 to 8, which contains an alcoholic solvent or a polar aprotic at a concentration of about 5 to 25% (v / v); d) loading the eluate containing neurotrophin onto a high resolution cation exchange chromatography resin, at pH 5 to 6; and e) the elution of the neurotrophin from the resin of high-resolution cation exchange chromatography with a buffer at pH 5 to 6 containing a cation at a concentration of 0.2 to 0.5 M.