Pharmaceutical composition comprising erythrocytes encapsulating a PLP-dependent enzyme and, a non-phosphate PLP precursor

Abstract

The invention relates to a pharmaceutical composition containing a PLP-dependent enzyme and optionally its cofactor, pyridoxal phosphate (PLP), and/or a phosphate or non-phosphate precursor of PLP, its use as a drug, its production method and a therapeutic treatment method related to it. The pharmaceutical composition comprises erythrocytes and a pharmaceutically acceptable vehicle, the erythrocytes encapsulating the PLP-dependent enzyme. The PLP-dependent enzyme may be methioninase, tyrosine phenol-lyase, tyrosine aminotransferase or cystathionine beta-synthase.

Claims

1. One or more erythrocyte(s) encapsulating a plurality of pyridoxal phosphate (PLP)-dependent enzyme molecules, the erythrocytes comprising a sufficient amount of pyridoxine kinase (PN-kinase) and pyridoxine phosphate oxidase (PNP oxidase) to produce a sufficient amount of PLP from PLP precursor present in a subject's bloodstream to maintain a sufficient portion of the erythrocyte-encapsulated PLP-dependent enzyme molecules in their holoenzyme forms to preserve enzymatic activity beyond 24 hours after injection or infusion into the subject in need of said enzymatic activity.

2. The erythrocyte(s) of claim 1, wherein the PLP-dependent activity persists in the subject for at least 1-15 days after injection or infusion.

3. The erythrocyte(s) of claim 1, wherein the PLP-dependent activity persists in the subject for at least 15 days after injection or infusion as measured by a depletion percentage of more than 20-50% of the PLP-dependent enzyme's substrate in the plasma of the patient or subject.

4. A pharmaceutical composition comprising the erythrocytes of claim 1 and a pharmaceutically acceptable vehicle.

5. The composition of claim 4, further comprising a non-phosphate precursor of PLP encapsulated in the erythrocytes and/or present outside the erythrocytes.

6. The composition of claim 5, wherein the non-phosphate precursor is selected from pyridoxal, pyridoxine, pyridoxamine, and combinations thereof.

7. The composition of claim 4, further comprising a phosphate PLP precursor, encapsulated in the erythrocytes.

8. The composition of claim 7, wherein the phosphate precursor is selected from the group consisting of pyridoxine phosphate (PNP), pyridoxamine phosphate (PMP), and combinations thereof.

9. The composition of claim 4, comprising from about 0.01 to about 30 mg of PLP-dependent enzyme per ml of erythrocytes.

10. The composition of claim 7, comprising from about 0.01 to about 30 mg of PLP-dependent enzyme per ml of erythrocytes.

11. The composition of claim 4, comprising from about 0.05 to 600 μmol of encapsulated PLP and/or pyridoxine phosphate (PNP) and/or pyridoxamine phosphate (PMP), per liter (L) of erythrocytes.

12. The composition of claim 11, comprising from about 5 to about 50 μmol of encapsulated PLP and/or PNP and/or PMP, per liter (L) of erythrocytes.

13. The composition of claim 4, further comprising pyridoxine kinase (PN-kinase), pyridoxine phosphate oxidase (PNP-oxidase), and an agent inhibiting pyridoxal phosphate phosphatase (PLP-phosphatase).

14. The composition of claim 4, wherein the PLP-dependent enzyme comprises a methioninase, a tyrosine phenol-lyase, a tyrosine aminotransferase or a cystathionine beta-synthase.

15. The composition of claim 7, wherein the PLP-dependent enzyme comprises a methioninase, a tyrosine phenol-lyase, a tyrosine aminotransferase or a cystathionine beta-synthase.

16. The composition of claim 4, comprising from about 0.05 to about 10 mg of PLP-dependent enzyme per ml of erythrocytes.

17. The composition of claim 7, comprising from about 0.05 to about 10 mg of PLP-dependent enzyme per ml of erythrocytes.

18. The composition of claim 4, comprising from about 0.5 to about 100 μmol of encapsulated PLP and/or pyridoxine phosphate (PNP) and/or pyridoxamine phosphate (PMP), per liter (L) of erythrocytes.

19. The composition of claim 4, comprising from about 5 to about 50 μmol of encapsulated PLP and/or pyridoxine phosphate (PNP) and/or pyridoxamine phosphate (PMP), per liter (L) of erythrocytes.

Description

(1) The purified methioninase may then be concentrated and diafiltered. Conservation may be made through freeze-drying and storage at about −80° C. The invention will now be described in more detail by means of embodiments taken as non-limiting examples and with reference to the drawing wherein:

(2) FIG. 1: Description of the method for purifying MGL, according to the method described in EP 0 978 560 and according to the improved method described in the present application. The modifications brought to the method described in patent EP 0 978 560 B1 relate to the steps located after the precipitation step with ammonium sulfate.

(3) FIGS. 2 and 3. Comparison of the intra-erythrocyte concentrations (FIG. 2) and extracellular concentrations (FIG. 3) of PLP after incubation of a RC-MGL-PLP suspension with pyridoxine (PN) at different concentrations. The RC-MGL-PLP suspension, incubated for 3 h and 24 h at room temperature in the absence of pyridoxine (0 mM) has a basal PLP level of about 3.9 μM. The incubation of the suspensions with 2 mM and 4 mM pyridoxine gives the possibility of increasing the intra-erythrocyte PLP concentration to 8 μM after 3 h of incubation (pale grey bars) and gives the possibility of attaining considerably higher levels (11 μM and 14 μM respectively) after 24 h of incubation (dark grey bars).

(4) FIG. 4. Pharmacokinetics of red corpuscles (RCs) loaded with a complex MGL-PLP. The RC-MGL-PLP2 product is obtained by lysis-resealing of a suspension containing 3 mg/ml of MGL and ˜30 μM of PLP. The RC-MGL-PLP3 product is obtained by lysis-resealing of a suspension containing 3 mg/ml of MGL and ˜125 μM of PLP. The RC-MGL-PLP4 product is obtained by lysis-resealing of a suspension containing 5 mg/ml of MGL and 33 μM of PLP. The RC-MGL-PLP5 product is obtained by lysis-resealing of a suspension containing 6 mg/ml of MGL and 100 μM of PLP. Fluorescent labeling of the products (CFSE) allows traceability of the RCs in vivo. The products injected intravenously to the mice CD1 (8 ml/kg for the products RC-MGL-PLP2, RC-MGL-PLP3 and RC-MGL-PLP5 and 10 ml/kg for the product RC-MGL-PLP4) have excellent stability with a survival rate of the injected RCs greater than 75% at 120 h, i.e. 5 days after their administration. For the RC-MGL-PLP4 product, the survival rate is reduced to less than 75% after ˜10 days.

(5) FIG. 5. Pharmacodynamics of the free MGL enzyme. The MGL enzyme was diluted by means of a potassium phosphate solution supplemented with 10 μM of P5P in order to obtain two injectable products (MGL-L1 and MGL-L2). These products were made so as to obtain 1) the same enzyme concentration as the product RC-MGL-PLP2, i.e. 0.45 mg/ml of MGL and 2) a concentration twice greater than the product RC-MGL-PLP2, i.e. 0.90 mg/ml of MGL. Both products are administered intravenously (IV) to CD1 mice (8 ml/kg) with supplementation IV of pyridoxine 6 h after the injection. The plasma L-methionine level is measured by HPLC-MS-MS. The L-Met level in non-treated CD1 was evaluated to be 68 μM. These products MGL-L1 and MGL-L2 both lead to rapid depletion within 15 min after their administration but not long lasting over time.

(6) FIG. 6. Pharmacodynamics of the RC-MGL-PLPs over short times. The product RC-MGL-PLP2 is obtained by lysis-resealing of a suspension containing 3 mg/ml of MGL and ˜30 μM of PLP. The product RC-MGL-PLP3 is obtained by lysis-resealing of a suspension containing 3 mg/ml of MGL and ˜125 μM of PLP. Both products are administered intravenously (IV) to CD1 mice (8 ml/kg) with IV supplementation of pyridoxine 6 h after administration for the mice receiving RC-MGL-PLP2. The plasma L-methionine level is measured by HPLC-MS-MS. The L-Met level in untreated CD1s was evaluated to be 82 μM. Both products RC-MGL-PLP2 and RC-MGL-PLP3 lead to rapid depletion 15 min after their administration reducing the L-Met level to 15.0±3.6 μM and 22.7±1.5 μM respectively and then maintaining more moderate depletion at 35 μM but stable between 48 h and 120 h.

(7) FIG. 7. Pharmacodynamics of the RC-MGL-PLPs over long periods. The product RC-MGL-PLP4 (0.5 mg/ml) is obtained by lysis-resealing of a suspension containing 5 mg/ml of MGL and 33 μM of PLP. The product is administered intravenously (IV) to CD1 mice (10 ml/kg) with IV supplementation of pyridoxine 6 h after administration for the mice receiving RC-MGL-PLP4. The plasma L-methionine level is measured by HPLC-MS-MS. The L-Met level in untreated CD1s was evaluated to be 68 μM. The product RC-MGL-PLP4 leads to rapid depletion 15 min after administration reducing the L-Met level to ˜10 μM and then maintaining more moderate depletion at ˜25 μM but stable between 24 h and 48 h so as to then gradually return to the control values 12 days after injection.

(8) FIG. 8. Residual activity of circulating PLP enzymes. The enzyme MGL was diluted by means of a potassium phosphate solution supplemented with 10 μM of P5P in order to obtain the injectable product MGL-L2. The product RC-MGL-PLP2 is obtained by lysis-resealing of a suspension containing 3 mg/ml of MGL and ˜30 μM of PLP. The product RC-MGL-PLP3 is obtained by lysis-resealing of a suspension containing 3 mg/ml of MGL and ˜125 μM of PLP. The product RC-MGL-PLP4 is obtained by lysis-resealing of a suspension containing 5 mg/ml of MGL and 33 μM of PLP. The product RC-MGL-PLP5 is obtained by lysis-resealing of a suspension containing 6 mg/ml of MGL and 100 μM of PLP. The products are injected intravenously to CD1 mice (8 ml/kg for the products MGL-L2, RC-MGL-PLP2, RC-MGL-PLP3 and RC-MGL-PLP5 and 10 ml/kg for the product RC-MGL-PLP4). The residual activity of the injected MGL enzyme (total RCs) is determined by a measurement of the NH.sub.3 produced by MGL according to the method described in Example 4.

EXAMPLE 1. METHOD FOR OBTAINING AND CHARACTERIZING METHIONINE GAMMA LYASE (MGL)

(9) Production of the strain and isolation of a hyper-producing clone: the natural sequence of MGL of Pseudomonas putida (GenBank: D88554.1) was optimized by modifying rare codons (in order to adapt the sequence stemming from P. putida to the production strain Escherichia coli). Other changes have been made to improve the context of translation initiation. Finally, silent mutations were performed to remove three elements that are part of a putative bacterial promoter in the coding sequence (box-35, box-10 and a binding site of a transcription factor in position 56). The production strain E. coli HMS174 (DE3) was transformed with the expression vector pGTPc502_MGL (promoter T7) containing the optimized sequence and a producing clone was selected. The producing clone is pre-cultivated in a GY medium+0.5% glucose+kanamycin for 6-8 h (pre-culture 1) and 16 h (pre-culture 2) at 37° C.

(10) Fermentation: the production is then achieved in a fermenter with GY medium, with stirring, controlled pressure and pH from the pre-culture 2 at an optical density of 0.02. The growth phase (at 37° C.) takes place until an optical density of 10 is obtained and the expression induction is achieved at 28° C. by adding 1 mM IPTG into the culture medium. the cell sediment is harvested 20 h after induction in two phases: the cell broth is concentrated 5-10 times after passing over a 500 kDa hollow fiber and then cell pellet is recovered by centrifugation at 15900×g and then stored at −20° C.

(11) Purification: The cell pellet is thawed and suspended in lysis buffer (7v/w). Lysis is performed at 10° C. in three steps by high pressure homogenization (one step at 1000 bars, and then two steps at 600 bars). The cell lysate then undergoes clarification at 10° C. by adding 0.2% PEI and centrifugation at 15900×g. The soluble fraction is then sterilized by 0.2 μm before precipitation with ammonium sulfate (60% saturation) at 6° C., over 20 h. Two crystallization steps are carried out on the re-solubilized sediment using solubilization buffer, the first crystallization step is realized by addition of PEG-6000 at 10% (final concentration) and ammonium sulfate at 10% saturation, and the second crystallization is then performed by addition of PEG-6000 at 12% final concentration and 0.2M NaCl (final concentration) at 30° C. The pellets containing the MGL protein are harvested at each stage after centrifugation at 15900×g. The pellet containing the MGL protein is re-suspended in a solubilization buffer and passed over a 0.45 μm filter before being subject to two anion exchange chromatographies (DEAE sepharose FF). The purified protein is then subject to a polishing step and passed over a Q membrane chromatography capsule for removing the different contaminants (endotoxins, HCP host cell protein, residual DNA). Finally, the purified MGL protein is concentrated at 40 mg/ml and diafiltered in formulation buffer using a 10 kDa cut-off tangential flow filtration cassette. Substance is then aliquoted at ˜50 mg of protein per vial, eventually freeze-dried under controlled pressure and temperature, and stored at −80° C.

(12) Characterization: The specific activity of the enzyme is determined by measuring the produced NH.sub.3 as described in example 4. The purity is determined by SDS-PAGE. The PLP level after being taken up in water was evaluated according to the method described in example 5. The osmolarity is measured with an osmometer (Micro-Osmometer Loser Type 15).

(13) The following table summarizes the main characteristics of one produced batch of MGL:

(14) TABLE-US-00001 MGL of P. putida Formulation Freeze-dried (amount per tube: 49.2 mg). Characteristics after being taken up in 625 μL of water: 78.7 mg/ml, ~622 μM of PLP, 50 mM of Na phosphate, pH 7.2, Osmolarity 300 mOsmol/kg. Specific activity 13.2 IU/mg Purity >98%

(15) Discussion of the production method. The method for purifying the MGL described in Example 1 is established on the basis of the method detailed in patent EP 0 978 560 B1 and of the associated publication (Takakura et al., Appl Microbiol Biotechnol 2006). This selection is explained by the simplicity and the robustness of the crystallization step which is described as being particularly practical and easily adaptable to large scale productions according to the authors. This step is based on the use of PEG6000 and of ammonium sulfate after heating the MGL solution obtained after the lysis/clarification and removal of impurities by adding PEG6000/ammonium sulfate steps. The other notable point of this step is the possibility of rapidly obtaining a high purity level during the step for removing the impurities by achieving centrifugation following the treatment of the MGL solution with PEG6000. The impurities are again found in the centrifugation pellet, the MGL being in majority found in solution in the supernatant. Because of this purity, the passing of the MGL solution in a single chromatography step over an anion exchanger column (DEAE), associated with a purification step by gel filtration on a sephacryl S200 HR column, gives the possibility of obtaining a purified protein.

(16) Upon setting into place the patented method for small scale tests, it appeared that the obtained results were not able to be reproduced. According to patent EP 0 978 560 B1, at the end of the step for removing the impurities (treatment with PEG6000/ammonium sulfate and centrifugation), the MGL enzyme is in majority found in the soluble fraction, centrifugation causing removal of the impurities in the pellet. During small scale tests conducted according to the described method in EP 0 978 560 B1, the MGL protein is again in majority found (˜80%) in the centrifugation pellet. The table below lists the percentage of MGL evaluated by densitometry on SDS-PAGE gel in soluble fractions.

(17) TABLE-US-00002 MGL percentage in Purification the soluble fraction Average Test no. 1 11% 17% Test no. 2 23%

(18) This unexpected result therefore led to optimization of the patented method by: 1) operating from the centrifugation pellet containing MGL, 2) carrying out two successive crystallization steps for improving the removal of the impurities after loading on a DEAE column, 3) optimizing chromatography on a DEAE column.

(19) For this last step, it is found that the DEAE sepharose FF resin is finally not a sufficiently strong exchanger in the tested buffer and pH conditions. After different additional optimization tests, the selection was finally directed to 1) replacement of the phosphate buffer used in the initial method with Tris buffer pH 7.6 for improving the robustness of the method and 2) carrying out a second passage on DEAE in order to substantially improve the endotoxin level and the protein purity without any loss of MGL (0.8 EU/mg according to Takakura et al., 2006 versus 0.57 EU/mg for the modified method).

(20) Finally, in order to obtain a method compatible with the requirements for large scale GMP production, a polishing step on a membrane Q was added in order to reduce the residual endotoxins and HCP levels. This final step of polishing avoids the implementation of the S200 gel filtration chromatography which is a difficult step to be used in production processes at an industrial scale (cost and duration of the chromatography).

(21) The different purification steps of the method from EP 0 978 560 B1 as well as of the method of the present application are given in FIG. 1.

(22) The following table gives the possibility of checking that the provided adaptations have led to obtaining a purification method with a yield at least equivalent to the one described in the initial method.

(23) TABLE-US-00003 Patent EP 978 560 B1 Method of the application Amount of Yield Amount of Yield Step enzyme (g) (%) enzyme (g) (%) Solubilised pellet 125 100 70 100 before DEAE Concentrated solution.sup.$ 80 64 46 65 .sup.$post sephacryl S-200 HR (EP 978 560) or post Membrane Q (method of the invention).

EXAMPLE 2. CO-ENCAPSULATION OF MGL AND PLP IN MURINE ERYTHROCYTES

(24) Whole blood of CD1 mice (Charles River) is centrifuged at 1000×g, for 10 min, at 4° C. in order to remove the plasma and buffy coat. The RCs are washed three times with 0.9% NaCl (v:v). The freeze-dried MGL is re-suspended in water at a concentration of 78.7 mg/ml and added to the erythrocyte suspension in order to obtain a final suspension with a hematocrit of 70%, containing different concentrations of MGL and of the PLP. The suspension was then loaded on a hemodialyzer at a flow rate of 120 ml/h and dialyzed against a hypotonic solution at a flow rate of 15 ml/min as a counter-current. The suspension was then resealed with a hypertonic solution and then incubated for 30 min at 37° C. After three washes in 0.9% NaCl, 0.2% glucose, the suspension was taken up in a preservation solution SAG-Mannitol supplemented with 6% BSA. The obtained products are characterized at D0 (within the 2 h following their preparation) and at D1 (i.e. after ˜18 h-24 h of preservation at 2-8° C.). The hematologic characteristics are obtained with a veterinary automaton (Sysmex, PocH-100iV).

(25) Results:

(26) In the different studies mentioned hereafter, the MGL activity in the finished products is assayed with the method described in example 4 against an external calibration range of MGL in aqueous solution. These results, combined with explanatory studies, show that MGL activity in the finished products increases with the amount of enzyme introduced into the method and that it is easily possible to encapsulate up to 32 IU of MGL per ml of finished product while maintaining good stability.

(27) In another study, three murine finished products RC-MGL-PLP1, RC-MGL-PLP2 and RC-MGL-PLP3 were prepared according to the following methods: RC-MGL-PLP1: co-encapsulation of MGL and of PLP from a suspension containing 3 mg/ml of MGL and ˜30 μM of PLP. The final product was taken up in SAG-Mannitol, 6% BSA supplemented with final 10 μM PLP. RC-MGL-PLP2: co-encapsulation of MGL and of PLP from a suspension containing 3 mg/ml of MGL and ˜30 μM of PLP. The finished product was taken up in SAG-Mannitol 6% BSA. RC-MGL-PLP3: this product stems from a co-encapsulation of MGL and PLP from a suspension containing 3 mg/ml of MGL and ˜124 μM of PLP. The final product was taken up in SAG-Mannitol 6% BSA.

(28) In a third study, a murine finished product RC-MGL-PLP4 was prepared from a new batch of MGL according to the following methods:

(29) RC-MGL-PLP4: co-encapsulation of MGL and the PLP from a suspension containing 5 mg/ml of MGL and ˜35 μM of PLP. The finished product was taken up in SAG-Mannitol 6% BSA.

(30) Finally in a fourth study, a murine product RC-MGL-PLP5 was prepared from a third batch of MGL according to the following methods: RC-MGL-PLP5: co-encapsulation of MGL and PLP from a suspension containing 6 mg/ml of MGL and ˜100 μM of PLP. The finished product was taken up in SAG-Mannitol 6% BSA.

(31) The hematologic and biochemical characteristics of the three finished products at D0 (after their preparation) are detailed in the table below. The encapsulation yields are satisfactory and vary from 18.6% to 30.5%.

(32) TABLE-US-00004 RC- RC- RC- RC- RC- MGL- MGL- MGL- MGL- MGL- PLP1 PLP2 PLP3 PLP4 PLP5 Hema- Hematocrit (%) 50.0 49.6 50.0 50.0 50.0 tological Corpuscle volume (fl) 46.3 46.5 46.8 42.4 45.6 data Corpuscle hemoglobin (g/dl) 24.7 24.0 24.2 27.4 25.1 RC concentration (10.sup.6/μl) 6.5 6.9 6.6 7.2 6.0 Total hemoglobin (g/dl) 14.8 15.4 15.0 16.6 13.8 Extracellular Hb (g/dl) 0.1 0.1 0.1 0.2 0.05 mgl Intra-erythrocyte concentration 0.97 0.94 0.79 1.01 1.36 of MGL (mg/ml of RC) Intra-erythrocyte activity of MGL 12.8 12.4 8.8 5.0 8.6 (IU/ml of RC)* Extracellular activity (%) 0.92% 0.97% 1.32% 1.18% 2.23% Intracellular activity (%) 99.08%  99.03%  98.68%  98.82%  97.77%  Encapsulation yield of MGL (%) 18.6% 30.5% 22.6% 19.4% 22.7% PLP Intra-erythrocyte concentration ND 13.4 71.4 10.2 ND of PLP (μmol/l of RC) Intracellular PLP fraction (%) ND 99.5 98.7 98.1 ND Extracellular PLP fraction (%) ND 0.5 1.3 1.92 ND PLP encapsulation yield (%) ND 44.8 57.4 30.7 ND *Calculated from the specific activity of each batch.

EXAMPLE 3. PRODUCTION OF HUMAN RCS ENCAPSULATING METHIONINE GAMMA LYASE AND PLP ACCORDING TO THE INDUSTRIAL METHOD

(33) A pouch of leukocyte-depleted human RCs (provided by the “Etablissement Français du Sang”) is subject to a cycle of three washes with 0.9% NaCl (washer Cobe 2991). The freeze-dried MGL is re-suspended with 0.7% NaCl and added to the erythrocyte suspension in order to obtain a final suspension with a hematocrit of 70%, containing 3 mg/ml of MGL and ˜30 μM of PLP (stemming from the formulation of MGL). The suspension is homogenized and it is proceeded with encapsulation according to the method described in EP 1 773 452. The suspension from the resealing is then incubated for 3 h at room temperature in order to remove the most fragile RCs. The suspension is washed three times with a 0.9% NaCl, 0.2% glucose solution (washer Cobe 2991) and then re-suspended with 80 ml of preservation solution (AS-3). The encapsulated MGL level is assayed like in Example 4.

(34) TABLE-US-00005 J0 J1 J7 Hematocrit (%) 52.0 51.6 52.7 Corpuscle volume (fl) 91.0 92.0 88.0 Corpuscle hemoglobin (g/dl) 30.3 29.8 31.6 RC concentration (10.sup.6/μl) 6.00 5.92 5.98 Total hemoglobin (g/dl) 16.4 16.2 16.6 Extracellular Hb (g/dl) 0.119 0.197 0.280 Osmotic fragility (g/l) 1.17 Hemolysis (%)  0.7%  1.2% 1.7% Total MGL concentration (mg/ml) 0.36 0.35 MGL supernatant concentration (mg/ml) 0.01 0.01 MGL intra-erythrocyte concentration (mg/ml, 0.68 0.67 100% Ht) Extracellular activity (%)  1.3%  1.4% Intracellular activity (%) 98.7% 98.6% Encapsulation yield (%) 19.7%

EXAMPLE 4. ASSAY OF ENCAPSULATED MGL IN THE RCS

(35) The assay of the MGL activity in cell suspensions (total RCs) and in the supernatants is based on a measurement of NH.sub.3 produced by MGL. The NH.sub.3 ions are assayed indirectly by enzymatic action of glutamate dehydrogenase (GLDH) according to the kit marketed by Roche Diagnostics (11877984).

(36) Preparation of the standards: MGL standards at different concentrations were prepared in matrices (total or supernatant RCs) or in an aqueous solution. For standards in an aqueous solution, MGL is prepared at concentrations varying from 0 to 12 μg/ml in the presence of 20 μM PLP in a phosphate buffer 100 mM at a pH of 7.2. For total RC matrix standards, 10 μl of RC-LR are lysed with 90 μl of a solution containing 260 μM of PLP and of MGL at concentrations varying from 0 to 100 μg/ml. The “total RC” standards are then diluted 20 times with phosphate buffer 100 mM, pH 7.2. For supernatant matrix standards, 10 μl of supernatants of RC-LR are lysed with 50 μl of a solution containing 6.4 μM of PLP and of MGL at concentrations varying from 0 to 20 μg/ml.

(37) Pre-treatment of the samples: the samples to be assayed (10 μl) are pre-treated in the same way as the standards (addition of PLP and identical dilutions but without addition of MGL).

(38) Assay of MGL: 7.5 μl of standards (STD) or of samples are introduced into the wells of a UV plate. 94 μl of reagent R1 (Roche kit) and 56 μl of reagent R2 (Roche kit) containing α-ketoglutarate in a buffer solution, NADPH and GLDH are added in order to remove the endogenous NH.sub.3 ions of the samples. After 10 min of incubation, 75 μl of L-methionine at 78.3 mM are introduced and the reaction mixtures are incubated for 30 min. Degradation of NADPH into NADP.sup.+ is continuously tracked by measuring the optical density at 340 nm. For the standards and the samples, the value of ΔOD/min is calculated over the linear domain of the O.D. curves obtained at 340 nm. A calibration curve ΔOD/min=f (MGL concentration or activity in the standards) is then plotted. The regression parameters allow determination of the MGL concentration in the samples. This result may be expressed in mg/ml or in IU/ml (the specific MGL activity being evaluated for each batch). The intra-erythrocyte MGL level is obtained by a calculation with the following formula: [MGL].sub.intra-erythrocyte=([MGL].sub.total−([MGL].sub.supernatants×(1-hematocrit/100))/(hematocrit/100).

EXAMPLE 5. ASSAY OF PLP IN BLOOD SAMPLES BY A HPLC METHOD

(39) The assay of PLP in cell suspensions (total RCs) and in the supernatants is an adaptation of the method described by Van de Kamp et al. Nutrition Research 15, 415-422, 1995. The assay is carried out with RP-HPLC (Shimadzu UFLC) with detection by fluorimetry (RF-10AXL instrument, excitation: 300 nm, emission: 400 nm). The PLP contained in the samples is extracted with trichloro-acetic acid (TCA) at a final 6%. After centrifugation (15,000×g, 10 min), the supernatants are collected and then diluted in a mobile phase A. A 50 μl sample volume is injected on a 5 μC-18 Gemini column, 250×4.6 mm (Phenomenex). The mobile phase A consists of 33 mM of monobasic potassium phosphate, of 8 mM of sodium 1-octanesulfonate supplemented with sodium bisulfite (0.5 g/l) for intensifying the signal of the PLP and of the mobile phase B, of 33 mM of monobasic potassium phosphate and of 17% (v:v) of 2-propanol. The gradient used is the mobile phase A (100%) with increasing proportions of mobile phase B: an increase from 0% to 8% of B over a period of 8 min. The flow through the column is maintained at 1 ml/min. The PLP concentration in the samples is determined with an external standard range of PLPs subject to the same TCA treatment as the samples. The retention time of PLP is ˜3.4 min. The intra-erythrocyte PLP level is obtained by calculation with the following formula: [PLP].sub.intra-erythrocyte=([PLP].sub.total−([PLP].sub.supernatants×(1-hematocrit/100))/(hematocrit/100).

EXAMPLE 6. INCREASE IN THE PLP LEVEL IN RCS BY CO-ENCAPSULATION OF PLP WITH MGL

(40) Suspensions of murine RCs are subject to the method for encapsulating MGL and PLP as described in Example 2. The assay of the intracellular PLP is carried out according to the method described in Example 5.

(41) A suspension of human RCs is subject to the method for encapsulating MGL and PLP as described in Example 3. Before the incubation step at room temperature, a portion of the human RC-MGL-PLPs is sampled in order to carry out an assay of the intracellular PLP according to the method described in Example 5.

(42) The following table compares the physiological levels of PLP in human or murine erythrocytes with the level attained by co-encapsulation of the latter with MGL.

(43) TABLE-US-00006 Human RCs 0.11 μM Murine RCs (Natta & ~2.4 μM* Physiological level of PLP Reynolds) (Fonda) PLP level Conditions before ~3.90 μM ~13.4 μM in RC-MGL-PLP1s or dialysis: RC-MGL-PLP2s 3 mg/ml MGL  ~30 μM PLP PLP level Conditions before ~71.4 μM in RC-MGL-PLP3s dialysis: 3 mg/ml MGL ~125 μM PLP *detail of the calculation: 7.5 nmol/g Hb ~2.4 μM (by assuming a CCMH of 32 g/dl).

EXAMPLE 7. DEMONSTRATION OF THE INCREASE IN THE PLP CONCENTRATION IN THE RC-MGLS BY INCUBATION IN VITRO WITH PYRIDOXINE

(44) A suspension of human RCs is subject to the method for encapsulating MGL as described in Example 3. Before the 3 h incubation step, a portion of the RCs is sampled and separated into three for volume-volume incubation with pyridoxine at different concentrations (0 mM, 2 mM and 4 mM). After homogenization, these suspensions are incubated at room temperature (RT). After 3 h and 24 h of incubation, samples of the cell suspensions and of the supernatants (obtained after centrifugation of the suspensions at 1000×g, at 4° C., for 10 min) are prepared and frozen for a measurement of the PLP concentration by HPLC as described in Example 5.

(45) The obtained results are shown in FIGS. 2 and 3.

(46) In the absence of pyridoxine, the intra-erythrocyte PLP level is 3.9 μM (PLP stemming from the co-encapsulation of MGL and PLP). This PLP concentration remains constant after 3 h and 24 h of incubation. A slight decrease in the PLP concentration is observed at 24 h and is concomitant with occurrence of extracellular PLP which may be explained by hemolysis at the end of the incubation.

(47) In the presence of pyridoxine (at 2 mM or at 4 mM), the RC-MGLs are enriched in PLP with intra-erythrocyte concentrations increased by a factor 2 after 3 h of incubation (˜8 μM of PLP) and by almost a factor 3 after 24 h of incubation with occurrence of a dose effect (11 μM and 14 μM for respective pyridoxine concentrations of 2 mM and 4 mM). These results show that an incubation of a RC suspension encapsulating a PLP enzyme dependent on PLP with pyridoxine (PN) is capable of increasing the intracellular PLP level in a long lasting way.

EXAMPLE 8. PHARMACOKINETICS OF RC ENCAPSULATING MGL-PLP IN MICE

(48) The murine products RC-MGL-PLP2, RC-MGL-PLP3, RC-MGL-PLP4 and RC-MGL-PLP5 are labeled with CFSE (fluorescent) and injected intravenously into CD1 mice. After various times (D0+15 min, D1, D2, D5, for the three products with additionally D14 and D28 for the RC-MGL-PLP4 and D14 for the RC-MGL-PLP5 product), the mice are sacrificed and the blood is collected on a lithium heparinate tube kept at +4° C. away from light for determining the pharmacokinetics. The proportion of red blood cells labeled with CFSE in the whole blood is determined by a flow cytometry method. Five microliters of whole blood are diluted in 1 ml of PBS 0.5% BSA and each sample is passed in triplicate (counting of 10,000 cells in FL-1; cytometer FC500, Beckman Coulter). The evaluation of the survival of red blood cells loaded with MGL is obtained by adding the proportion of RCs labeled with CFSE at different times to the proportion of RCs labeled with CFSE at T0+15 min (100% control). The different obtained percentages for each time are copied onto a graph (FIG. 4) illustrating the proportion of RCs loaded with MGL in circulation versus time.

(49) The determination of the proportion of RCs marked with CFSE in circulating blood at different times shows its excellent stability of the four products in vivo in mice, up to 120 h post-injection (83.5±0.6%, 94.7±0.6%, 87.3±5.6% and 76.8±1.3% survival rate, respectively). For the product RC-MGL-PLP4, the pharmacokinetic study over 29 days showed that the half-life of the red blood cells encapsulating MGL is ˜12.6 days.

EXAMPLE 9. L-METHIONINE DEPLETION AT 24 H

(50) The murine products RC-MGL-PLP1, RC-MGL-PLP2 and RC-MGL-PLP3 prepared and characterized as in Example 2 are injected intravenously to CD1 mice at a dose of 8 ml/kg. After 6 h, ˜0.09 mg of pyridoxine (i.e. 150 μL of a 2.9 mM pyridoxine hydrochloride solution) were injected to mice receiving RC-MGL-PLP2. The L-Met plasma level was evaluated at 24 h by HPLC-MS-MS (Piraud M. et al., Rapid Commun. Mass Spectrum. 19, 3287-97, 2005). The following table shows the depletions obtained in the various groups of injected mice.

(51) TABLE-US-00007 L-Met Administered Methods for providing plasma % of product the PLP co-enzyme level (μM) depletion — Feeding 82.7 ± 22.5 — RC-MGL- Feeding 46.3 ± 3.5  44% PLP1 PLP in the finished product (~5 μmol/l, RC) PLP in the preservation solution of the finished product (10 μM) RC-MGL- Feeding 22.3 ± 4.9  73% PLP2 + PLP encapsulated in the finished pyridoxine product (~13.4 μmol/l RC) IV injection of ~0.09 mg of pyridoxine RC-MGL- Feeding 29.7 ± 4.6  64% PLP3 PLP encapsulated in the finished product (~71.4 μmol/l RC)

(52) The L-Met plasma level was evaluated to be 82.7±22.5 μM in control mice. The product RC-MGL-PLP1 containing encapsulated MGL with a low PLP concentration leads to 44% depletion of L-Met, 24 h after administration of the product. We put forward the assumption that the PLP added into the preservation solution of the product is not available for the enzyme MGL since 1) it is in majority bound to the BSA present in the preservation solution and 2) it cannot pass through the membrane of the RC.

(53) The results show that a more consequent provision of PLP in the red corpuscle either by IV injection of pyridoxine (RC-MGL-PLP2) or by encapsulation of PLP at a stronger concentration gives the possibility of obtaining L-Met depletions ˜1.5 times greater (73% and 64% depletion respectively).

EXAMPLE 10. PHARMACODYNAMICS OF RC-MGLS

(54) The MGL enzyme in its free form is injected intravenously to CD1 mice at a dose of 8 ml/kg. Two series of injections were made, the first with an enzyme concentration at 0.45 mg/ml (product MGL-L1), the second at a twice higher concentration (0.90 mg/ml; product MGL-L2). Six hours after injection, ˜0.09 mg of pyridoxine (i.e. 150 μL of a 2.9 mM pyridoxine hydrochloride solution) are injected into mice receiving MGL-L1 and MGL-L2. The L-Met plasma level is evaluated by HPLC-MS-MS at 15 min, 24 h, 48 h, and 120 h post-injection of MGL-L1 and at 15 min, 24 h, 48 h, 120 h and 144 h post-injection of MGL-L2. FIG. 5 shows the depletions obtained in the various groups of injected mice.

(55) The results show that in both experimental groups, a very strong L-methionine depletion (≈4 μM) and rapid overtime (15 min post-injection). However, this depletion is transient and not maintained over time, the L-methionine levels returning into the control values 24 h after injection, and this in spite of the initial provision of P5P (present in the dilution buffer but also in the formulation of the enzyme taken up in water) and the supplementation with vitamin B6 identical with the one carried out for the RC-MGL-PLP2 product. The activity of free MGL is therefore lost between 15 min and 24 h post-injection, probably due to rapid removal of the circulating enzyme.

(56) In a second phase, the murine products RC-MGL-PLP2 and RC-MGL-PLP3 prepared and characterized as in Example 2 are injected intravenously to CD1 mice at a dose of 8 ml/kg. After 6 h, ˜0.09 mg of pyridoxine (i.e. 150 μL of a 2.9 mM pyridoxine hydrochloride solution) are injected into the mice receiving RC-MGL-PLP2. The L-Met plasma level is evaluated to be at 15 min, 24 h, 48 h, and 120 h post-injection by HPLC-MS-MS. FIG. 6 and the following table show the depletion obtained in the various groups of injected mice.

(57) The results show that in both experimental groups, an L-methionine depletion stabilized at ˜35 μM and maintained over time (from 48 to 120 h post injection). These results indicate that supplementation with PLP or with its precursor (vitamin B6) gives the possibility of maintaining an activity of the MGL encapsulated in RCs for at least 120 h after injection in mice. As an indication, the L-methionine concentrations in plasma 24 and 120 h post-injection for the various products (free form of MGL or co-encapsulated in red blood cells with PLP) are given in the following table:

(58) TABLE-US-00008 L-methionine level (μM) At T0 (ctrl) At 24 h At 120 h Free MGL MGL-L1 68.2 ± 21.7 57.0 ± 8.0  62.5 ± 12.0 MGL-L2 68.2 ± 21.7 57.3 ± 5.1 51.0 ± 8.5 Encapsulated RC-MGL-PLP2 82.7 ± 22.5 29.7 ± 4.6 36.0 ± 2.6 MGL RC-MGL-PLP3 82.7 ± 22.5 22.3 ± 4.9 34.3 ± 7.4

(59) In order to assess the pharmacodynamics over times of more than 120 h, the murine product RC-MGL-PLP4 having a concentration of encapsulated MGL of 0.5 mg/ml of enzyme in the finished product is injected intravenously to CD1 mice with a dose of 10 ml/kg. After 6 h, ˜0.09 mg of pyridoxine (i.e. 150 μL of a 2.9 mM pyridoxine hydrochloride solution) are injected into the mice receiving RC-MGL-PLP4. The L-Met plasma level is evaluated by HPLC-MS-MS. FIG. 7 shows the depletions obtained at 15 min, on D1, D2, D5, D14 and D28 after injection.

(60) The results show a significant L-methionine depletion (≈10 μM against ≈68 μM for the control) and rapid depletion over time (15 min post-injection). However, this depletion is slightly stabilized between 24 h and 48 h to a value of ˜25 μM and increases up to ˜40 μM after 5 days so as to finally attain the control values at about 12 days after injection of RC-MGL-PLP4.

(61) Finally, the residual activity of the injected MGL enzyme is determined according to the assay method described in Example 4 in the presence of PLP. FIG. 8 hereafter lists the residual activities versus time for the free enzyme MGL-L2, the various products RC-MGL-PLP and for the TPL enzyme (data from the literature).

(62) The results show that by encapsulating MGL in murine red blood cells it is possible to retain strong enzymatic activity at 24 h (residual activity comprised between ˜60 and 100%). This residual activity slightly decreases at 48 h (˜35 to 100%) and is maintained up to 120 h, i.e. 5 days after injection, at values comprised between ˜20 and ˜65%. The residual activity of the MGL in its free form drastically drops in the first minutes post-injection so as to be almost zero at 24 h (residual activity<10%). By comparison, the residual activity of the TPL injected in its free form is copied on the graph and 5 h after injection, the latter is only at most 37% (Elmer et al., 1978). The measurement of the residual activity clearly shows the benefit of encapsulation of the PLP enzymes in red blood cells for maintaining their enzymatic activity.