WO2005095430A1

WO2005095430A1 - An improved process for the manufacture of high pure gemcitabine hydrochloride

Info

Publication number: WO2005095430A1
Application number: PCT/IN2004/000084
Authority: WO
Inventors: Ramesh Babu Polturi; Subramanian Hariharakrishnan Venkata; Ramesh Betini
Original assignee: Ramesh Babu Polturi; Venkata Subramanian Hariharakr; Ramesh Betini
Priority date: 2004-03-31
Filing date: 2004-03-31
Publication date: 2005-10-13

Abstract

A process for the preparation of Gemcitabine hydrochloride of formula (I) of extra high purity by the reaction of (R) -2,3-0-isopropylidene glyceraldehyde of formula (II) with ethyl bromo difluoroacetate of formula (III) followed by hydrolytic cyclization of the product of formula (IV) converting the product into a dibenzoyl derivative of formula (V) of high purity reducing the product of formula (V) and converting the resultant lactol into a mesylate of formula (VI) followed by coupling the mesylate of formula (VI) with bis-silyl acetyl cytosine of formula (X) and subsequently deblocking and purifying.

Description

AN IMPROVED PROCESS FOR THE MANUFACTURE OF HIGH PURE GEMCITABINE HYDROCHLORIDE

REFERENCES CITED: US Pat. No. 4526988 US Pat. No. 4808614 US Pat. No. 5233608 US Pat. No. 5945547 EP Pat. No. 306190

Other References Synthesis 565-570, 1992

BACKGROUND OF THE INVENTION:

Furanosyl nucleosides like cytarabine and zalcitabine have been used for the treatment of acute mylogenous leukemia. Introducing fluorine into nucleosides has also proved to be useful procedure for modifying the biological activity of these compounds. Gemcitabine, a 2'-deoxy-2, 2'-difluoro nucleoside has proved to be highly active against cancer and has been used for treating several types of tumors.

The synthesis of Gemcitabine has been described in many patents. The basic chemistry approach has remained the same, except variations in the protecting groups and improvements in the purification methodology.

CROSS REFERENCE TO THE RELATED APPLICATIONS

US Pat No. 4526988 and 4808614 describe the synthesis as given under scheme I

Formula - II Formula - III Formula - IV

Formula - X Formula - 1 Scheme-I In this process, (R)-2,3-0-isopropylideneglyceraldehyde was treated with ethyl bromodifluoroacetate under Reformatsky conditions to yield the product of formula IV, which contained R and S isomers in the ratio 3:1 and this mixture, was purified by column chromatography to get the pure R-isomer. This product was hydrolytically cyclized using ion-exchanged resin to give 2-deoxy-2, 2- difluoro-1-oxo ribose of formula VII, which was converted into bis-t-butyl dimethyl silylated product and purified by column chromatography to give a product of formula VIII

The lactone of formula VIII was reduced with DIBAL to give lactol of formula IX, which was again purified by column chromatography.

Reaction of this product with methane sulfonyl chloride in presence of a base yielded the product of formula X, which was coupled with bis silyl acetyl cytosine of formula XI and the coupled product was deprotected and purified by preparative HPLC to give product of formula I

The US Pat No. 5223608 is described in scheme II

OOOC_jHs

α + β Anomer

Scheme II

p Anomer As per the scheme II, the reaction sequence is same as followed by US Pat. No's 4526988 and 4808614 up to the formation of product of formula IV. This was first benzoylated without separation of R&S isomers to give a product of formula XII, which was subjected to hydrolytic cyclization to give a product of formula XIII. This product was benzoylated and purified by crystallization, preferably, from a mixture of dichloromethane and hexane/heptane or isopropyl alcohol to give a product of formula V viz. 2-deoxy-2,2-difuluoro-D-erythreo pentofuranose-1-ulose-3,5-dibenzoate in about 95% purity.

The product of formula V was reduced to give lactol of formula Va and then mesylated to give a product of formula VI. The mesylate, of formula VI, was coupled with bis silyl acetyl cytosine to give the protected nucleoside, which was hydrolyzed, converted into hydrochloride salt and crystallized to give Gemcitabine Hydrochloride of greater than 99% purity.

These are the two major approaches for the synthesis of Gemcitabine hydrochloride. The method described in US Pat No's 4526988 and 4808614 adopt laborious chromatographic purifications in different stages which make the process unviable industrially. The process described in US Pat.No.5223608 takes care of the shortcomings. In this procedure also, there are a few inconsistencies, which are listed below a) Isolation of 2-deoxy-2,2,-difluoro-D-erythropentofurano-1-ulose-3,5- dibenzoate b) Reduction of the lactone of formula V to the lactol of formula Va c) Isolation of the 1 : 1 α/β anomers and further purification

The crude product of formula V containing the erythro and threo isomers was crystallized, preferably, from a mixture of dichloromethane and hexane. When this method was followed, there was large variation in the proportion of erythro and threo for different batches and it was tedious to get a product mixture containing 95% rich in erythrose isomer The reduction of formula V to the lactol of formula Va was carried out using DIBAL or tri-tertiary butoxy lithium aluminum hydride. Both these reagents are expensive and very difficult to handle.

The crude Gemcitabine hydrochloride was isolated as 1 :1 α and β anomer mixture in aqueous isopropyl alcohol and subsequently crystallized from aqueous acetone to get the product of formula I in greater than 99% purity by HPLC

The present invention has addressed all the above given difficulties and come out with an excellent method for the isolation of Gemcitabine hydrochloride of extra high purity 99.8-99.9%

OBJECTS OF THE INVENTION

As per the procedure described in US Pat. No 5223608 the product of formula IV was first benzoylated, followed by hydrolytic cyclization and again benzoylated to get the 3,5-dibenzoate of formula V. The crude mixture contained erythro and threo isomers. The product enriched in erythro isomer was obtained by crystallization in a mixture of dichloromethane and hexane. When this procedure was tried out, the observation was that it was difficult to reproduce the earlier results with any consistency. Hence, it was decided to attempt some modification.

The purpose of the present invention was to develop a new methodology for the large-scale preparation of Gemcitabine hydrochloride. Yet, another purpose of the invention was to develop a methodology by which a product of extra high purity could be obtained.

BRIEF SUMMARY OF THE INVENTION

The invention describes a new process for the manufacture of highly pure Gemcitabine hydrochloride of formula I

Formula - 1 by the reaction of (R)-2,3-O-isopropylidene glyceraldehyde of formula

Formula - II with ethylbromodifluoroacetate of formula III F I Br — C COOC2H5 I F Formula - III to give a product of formula IV,

OH Formula - IV

which is hydrolytically cyclized, converted into a dibenzoyl derivative and purified to give a product of formula V

Formula - V Product of formula V is reduced to lactol and converted into a mesylate of formula VI

Formula-VI which is coupled with Bis(trimethylsilyl)-N-acetylcytosine of formula XI, under Verburggen condition, hydrolyzed and isolated to give product of formula I in 95% purity. This is further purified to yield a product of extra high purity.

DETAILED DESCRIPTION OF THE INVENTION

The procedure followed is represented in the scheme III.

OH Formula - VII Formula - IV

Purification

Scheme Ethyl 2,2-difluoro-3-hydroxy-3 (2,2-dimethyldioxalan-4-yl) acetate was hydrolytically cyclized using acetonitrile and water containing trifluoro acetic acid. The proportion of acetonitrile was about 5-10 volumes of the weight of the substrate, the amount of water was about 2 to 4 times to the moles of substrate and trifluoro acetic acid was about 0.2 to 0.3 times moles of the substrate.

The reaction mixture was stirred under reflux for about 3 to 6 hours. Subsequently, acetonitrile was distilled off and simultaneously toluene was added into the reactor. The addition of toluene was continued till the mass temperature reached to 110°C. The progress of the reaction could be monitored by IR spectroscopy. The CO absorption for compound of formula IV is at 1760 cm^"1 and for the product of formula VII is at 1812 cm^"1. At that point, toluene was distilled off completely under reduced pressure. The residue was dissolved in ethyl acetate and reacted with benzoyl chloride in the presence of an organic base like pyridine and a catalyst like 4-dimethylamino pyridine. The addition of benzoyl chloride to a mixture of the substrate, pyridine and 4- dimethyl amino pyridine was preferably carried out at 50 °C -60 °C and after the addition, the reaction mixture was stirred at 65 °C -70 °C for 3-4 hours. The reaction mixture on HPLC analysis (see examples for condition) indicated the presence of erythro isomer in 60-65%, threo in 20-25% and an unknown product in about 10-20%. A crude product was obtained on working up the above reaction mixture and this on crystallization from a mixture of dichloromethane and hexane in the ratio of 1:2 yielded a product containing a mixture of erythro, threo and an unknown product. The unknown product, in all probability could be the lactone reverting back to the γ-hydroxy acid represented by the formula XIV

Formula - XIV This was confirmed by stirring the crude product of formula V with aqueous hydrochloric acid. This process converted the mixture completely into the product of formula XIV, which was characterized by IR, NMR, MS and elemental analyses. The pure compound of formula XIV was compared with the unknown product obtained during the hydrolytic cylization by HPLC and found to be the same. The product of formula XIV on treatment with benzene/toluene containing benzene/toluene sulphonic acid under azeotropic distillation condition yielded a product containing the erythro and threo mixture of the product of formula V.

Hence it was decided to subject the crude mixture to an azeotropic distillation using preferably toluene containing p-toluene sulphonic acid. After the water collection had ceased, the reaction solution was cooled to about 60 °C and washed with hot water. Toluene layer was separated, dried with anhydrous sodium sulphate and concentrated under reduced pressure.

Since the crystallization by dissolution in dichloromethane and precipitation by the addition of hexane did not yield consistent results, it was decided to attempt the purification in different composition. For example, the crude product was dissolved in a solvent like benzene or toluene and under hot condition (60 °C to 70 °C) solvent like petroleum ether (60-80 fraction) or hexane or cyclohexane was added to precipitate the pure product. After several experimentations, it was observed that the preferable combination was toluene and hexane. The residue, when dissolved in about 3 to 6 volumes of toluene and on addition of an equal volume of hexane, precipitated a product, which was >99% pure in the erythro isomer. [α]²⁵ _D= +47.4 (C=1.0, CHCI₃) mp 119-120°C and the overall yield starting with product of formula IV was 35 %.

Another variation of the process described above was also tried out. This was an effort to avoid an additional solvent like acetonitrile. The crude product of formula IV was treated with about 5 to 10 volumes of water. It was preferable to use 6 to 7 volumes of water. The mixture was warmed to about 60 °C to 70 °C in the presence of an acid catalyst like benzene/p-toluene sulphonic acid. The acid catalyst was used to the extent of 20-mole % of the substrate. The reaction mixture was heated, at 60 °C to 70 °C for 3 to 6 hours. After that, the reaction mixture was diluted with toluene and subjected to azeotropic distillation of water. When the water separation had ceased, toluene was distilled off completely and the residue was dissolved in ethyl acetate and subjected to benzoylation with benzoyl chloride using pyridine as base and 4-dimethylamino pyridine as catalyst. After the benzoylation and work up, the residue was dissolved in about 3 volumes of toluene, warmed and diluted with equal volume of hexane and stirred to room temperature and later cooled to 10°C-15 °C. The precipitated solid was filtered, washed with a mixture of toluene and hexane and dried. The purity of the product by HPLC was >99%, [α]²⁵ _D= +47.4° (C=1.0, CHCI₃), mp 119-120°C and overall yield starting with product of formula IV was35%

An attempt was made to reduce the lactone to lactol with reagent other than DIBAL or tri-tertiary butoxy lithium aluminum hydride. The reducing agent, Vitride i.e. sodium bis (2-methoxyethoxy) aluminum hydride, which is commercially available as a 65% w/w solution in toluene is an excellent reagent. The reagent is easy to handle and comparatively cheaper. The reaction was tried in solvents like ethylene glycol dimethyl ether, diethylene glycol dimethyl ether, THF, 1 ,4-dioxane etc. THF was found to be most preferable solvent. The mole ratio of the substrate to the reagent was preferably 1 :0.5 to-0.75moles. It was preferable to conduct the reaction at about -30 °C to +30 °C. The duration of the reaction was preferably 5 to 10 hours. The progress of the reaction was followed by HPLC (Zorbax CN hexane: PrOH 92:8 at λ254). The product showed up as two peaks. The lactol of formula Va was isolated and converted to the mesylate of formula VI using methane sulfonyl chloride/triethyl amine in dichloromethane. Progress of the reaction was monitored by HPLC (Zorbax CN hexane: PrOH 92:8 at λ254). The product showed up as two peaks. The reaction mixture was worked up as per standard method and the residue was subjected to Verbruggen protocol with bis-silyl acetyl cytosine using trimethyl silyl trifluoro methane sulfonate as the reagent in dichloroethane. After the coupling, the product was subjected to deprotection and the nucleoside was converted to its hydrochloride and isolated. The isolated product usually contained α and β-anomers and it was purified by crystallization. The present invention describes a method by which the crude hydrochloride is obtained in nearly 95% rich in β-anomer. This was further purified to get β-anomer in extra high purity.

The dibenzoate derivative of the nucleoside - Gemcitabine - was hydrolyzed in methanol using preferably ammonia. After the hydrolysis, the methanolic solution was concentrated and the residue was dissolved in 2 to 4 volumes of water. The aqueous solution was washed with an organic solvent. It was preferably washed with ethyl acetate and later with petroleum ether. The aqueous layer was diluted with about 3 to 4 volumes of isopropyl alcohol and concentrated. This process was repeated, 2 to 3 times. Subsequently the residue was dissolved in 3 to 6 volumes of isopropyl alcohol. It was preferable to dissolve the residue in about 4 to 5 volumes of isopropyl alcohol. The solution was warmed to about 60°C to 80 °C and hydrochloric acid, about 0.5 volumes to that of the residue, was added. The reaction mixture was stirred at about 65 °C to 75 °C for 30 to 60 minutes, allowed to cool to room temperature under stirring and further stirred for about 10 to 14 hours. Then the reaction mixture was cooled to about 0 °C to 5 °C and preferably stirred for 2 to 4 hours. The precipitated solid was filtered, washed with acetone and dried. The sample was about 95% rich in the β-anomer. The yield at this step was 12%.

The solid, obtained above, was further purified by first stirring with water. The ratio of the solid to water was 1 :1 to 2. The slurry was stirred for about 1 hour at room temperature, filtered, washed with acetone and dried. Dissolving in water and precipitating with a solvent like acetonitrile or isopropyl alcohol or acetone further purified the product. The solid was dissolved in 3 to 6 volumes of water. The solid was dissolved more preferably in 4 to 5 volumes of water. The aqueous solution was heated to about 70°C, filtered and diluted with one of the organic solvents mentioned earlier. The ratio of water to that of the organic solvent was preferably 1 :8 to 12. The ratio was preferably 1 :9 to 11. After dilution, the mixture was stirred at 60 °C to 70 °C for 2 to 4 hours, slowly cooled to room temperature, later stirred at 0°C to 5°C filtered, washed with acetone and dried. The overall yield after the two purifications was about 10-11 %. The obtained Gemcitabine Hydrochloride was 99.8 to 99.9% pure by HPLC. It showed [α]²⁵ _D +48° (c=1.0 D₂O) and mp288-291°C.

The following examples illustrate the specific aspects of the present invention

EXAMPLE 1

Preparation of 2-deoxy-2,2-difluoro-D-erythro pento furano-1-ulose-3,5- dibenzoate

Method A

Ethyl 3(RS) 2,2-difluoro-3-(2,2-dimethyldioxalan-4-yl) propionate was prepared by the methods described in the literature. The product was rich in R-isomer by about 75% and this product was used in the following reactions

Ethyl-2,2-difluoro-3-(2,2dimethyl dioxalaln-4-yl) propionate 100gms, was charged into a 2 lit four-necked round bottom flask fitted with a stirrer, condenser, dropping funnel and a stopper. Acetonitrile, 1 lit, trifluoroacetic acid 5.0ml and DM water 28ml were also charged into the flask and the reaction mixture was stirred to reflux. The reaction mixture was stirred at reflux temperature for about 3 hours. Then acetonitrile was slowly distilled from the reaction mixture. Simultaneously toluene was charged dropwise into the RB flask at the rate acetonitrile was distilling off. This process was continued until the mass temperature rose to about 110°C and the reaction mixture was stirred at 110 °C for about 15minutes. Later toluene was distilled off completely under reduced pressure at about 60 °C and the residue was used for the next reaction. The residue, obtained above, was dissolved in ethyl acetate, 500ml and transferred into a 2 lit four-necked flask fitted with a stirrer, condenser carrying a guard tube, an addition funnel carrying a nitrogen bubbling system and a stopper. 4-Dimethyl amino pyridine, 10.3gms and pyridine, 86ml, were charged into the flask. Benzoyl chloride, 112ml was dissolved in ethyl acetate, 500ml, and charged into the addition funnel. The reaction mixture was warmed to about 60 °C to 65°C under stirring and the solution of benzoyl chloride was added into the flask dropwise in about 3 hours. After the addition, the reaction mixture was stirred at 60 °C to 65 °C for about 3 hours and then cooled to 25 °C to 30 °C and filtered through a bed of hyflow. Then the hyflow bed was washed with ethyl acetate, 200ml. The combined ethyl acetate extract was washed with 10% hydrochloric acid, 300ml, 10% sodium bicarbonate solution, 300ml, saturated sodium chloride solution, 300ml and finally dried over anhydrous sodium sulphate. The ethyl acetate solution was filtered and concentrated under reduced pressure. The residue was dissolved in dichloromethane, 150ml and transferred into a three-necked round bottom flask fitted with a stirrer, addition funnel and a thermometer socket. Hexane, 300ml was charged into the addition funnel and this was added dropwise into the stirred dichloromethane solution. After the addition, the solution was stirred at 25°C to 30 °C for about 1 hour, when solid started precipitating out. The reaction mixture was then cooled to about 10°C to 15°C and stirred at that temperature for about one hour. The solid was filtered, washed with a mixture of dichloromethane and hexane at 10°C to yield 30 gms of title compound having mp 116-118°C [α]²⁵ _D= +40° (C=1.0, CHCI₃)

The solid was checked for its purity by HPLC [Zorbax CN 4.6x25 cm :eluent, hexane/l-PrOH (92:8)]. It was about 85% pure in the D-erythro, about 8% of threo and 7% of an unknown substance. Method B

The experiment was repeated as given under Method A. After the benzoylation the residue was crystallized from isopropyl alcohol to give 35gms of the title compound in 24% yield., having mp 115-116 °C [α]²⁵ _D= +39° (C=1.0, CHCI₃) HPLC analysis (Zorbax CN 4.6x25 cm :eluent, hexane/l-PrOH (92:8) flowrate, 5ml/min UV 254) indicated the product in 83% rich in D-erythro 8% D-threo and about 9% unknown compound.

EXAMPLE 2

Preparation of 2,2-difIuoro-3,5-dibenzoyloxy-4-hydroxy butyric acid

The product, as obtained in example 1 method A, 20gms, was charged into a 500ml three-necked round bottom flask fitted with a stirrer, condenser and a stopper. A solution of 4N hydrochloric acid, 200ml was also charged in the flask and the reaction mixture was stirred at 45°C to 50°C for about 3 hours and on cooling a solid precipitated out. The solid was filtered and dried. The product, on HPLC analysis, was matching with the retention time of the unknown compound found along with the product of formula IV and the purity was 99%. The compound was characterized by IR 'HNMR, MS and elemental analysis m p 104.5°C - 106.5°C

[ ]_D ²⁵ +30.71° (C = 1 , MeOH)

Elemental analysis as CιgHι₆F₂O Calculated C 57.90; H 4.10 Found C 57.85; H 4.12 IR (KBr, cm^"1) 3351, 1761, 1694, 1451, 1286, 1266, 1146, 1091714 NMR (DMSO_D6)δ 4.25 (M, 2H, C-5), 4.378 (m, 1H, C-4), 5.88 (m, 1H, C-3) 7.513 (m.6H, Ar-H), 7.96 (m, 4H, Ar-H) EXAMPLE 3

Preparation of pure 2-deoxy-D-erythro pento furanose-1-ulose-3,5-dibenzoate

Method A

The experiment was repeated with quantities as given in example 1 , method A. The residue obtained, after benzoylation, was analyzed by HPLC, which indicated erythrose 82%, threose 8% and the hydroxy acid 10%. This mixture was dissolved in toluene 500ml and transferred into a 1 lit three-necked round bottom flask fitted with a stirrer, Dean and Stark apparatus carrying a condenser and a stopper. p-Tolunesulfonic acid, 2gms was added and the mixture was stirred at reflux and the water collected azeotropically. When the water collection had ceased, the reaction mixture was analyzed by HPLC. An aliquot was transferred into a test tube, cooled to about 60°C and washed with hot water 60 °C. Toluene layer was separated and concentrated and the residue was subject to HPLC analysis. The product indicated erythrose 90% and threose 10%. The hydroxy acid was absent. The reaction mixture was cooled to about 60°C and hot water (60°C) 100ml was added and the mixture was stirred for 15 minutes. Toluene layer was separated and concentrated under reduced pressure at 60°C. The residue was taken in 350ml toluene and warmed to about 50°C. To the stirred solution hexane 350ml was added dropwise. At the end of the addition, the mixture was cooled first to 25°C and subsequently to about 10°C to 15°C. The cooled mixture was stirred at 10°C to 15°C for about an hour. The solid was filtered, washed with a 1 :1 (v/v) mixture of toluene and hexane. The solid was dried at 60°C. It showed HPLC purity of 99.8% in erythrose, having mp 120-121 °C [α]²⁵ _D= +47.8°. Yield 50 gms (34%). Method B

Ethyl 3(RS)-2,2-difluoro-3-(2,2-dimethyldioxalan-4-yl)propionate 100gms was charged into a 2 lit three-necked round bottom flask fitted with a stirrer, condenser and a stopper. Water 300ml was charged into the flask, followed by p-toluenesulfonic acid, 2gms. The reaction mixture was stirred at 75°C to 80°C for 3 hours. Subsequently toluene, 600ml, was charged into the flask and the set up was fitted with a Dean and Stark apparatus. The reaction mixture was stirred to reflux and water was distilled off as an azeotrope. Toluene, equivalent to the volume distilled off, was replaced into the flask. This process was continued until the distillation of the water was completed. The reaction mixture was cooled to about 60 °C and hot water (60 °C) 600ml, was added and the reaction mixture was stirred for 15 minutes. Toluene extract was separated and concentrated under reduced pressure at 60 °C. The residue was dissolved in ethyl acetate and subjected to benzoylation as given in example 1 , method A. The residue, after distillation of ethyl acetate, was dissolved in toluene 350ml and warmed to and 50 °C and to the stirred toluene solution at 50 °C, hexane, 350ml, was added dropwise from dropping funnel. After the addition, the solution was cooled to about 25 °C under stirring. Subsequently, the solution was slowly cooled to 10 °C to 15 °C, maintained for one hour and the solid was filtered. The solid was washed with 1 :1 (v/v) toluene-hexane mixture and dried at 60 °C to yield 50 gms of product. The HPLC purity of the compound was >99% of erythrose with mp120-121°C and [α]²⁵ _D= +47.8°.

Method C

100gms of solid obtained as given under example 2 was charged into a 2 lit three-necked round bottom flask fitted with a stirrer, Dean Stark apparatus carrying a condenser and thermometer socket. Toluene, 600ml, was added along with p-toluenesulfonic acid, 2gm and the mixture was stirred to reflux. Water formed was collected by azeotropic distillation. When the water collection had ceased, the reaction mixture was cooled to 60 °C and washed with hot water (60 °C, 300ml). Toluene layer was separated and concentrated under reduced pressure at <60 °C. Subsequently the residue was taken in toluene, 350ml and warmed to about 40 °C to 50 °C and hexane, 350ml, was added dropwise. After the addition, the mixture was cooled to 25 °C under stirring and then cooled to 10 °C to 15 °C. It was stirred at 10 °C to 15 °C for one hour, the solid was filtered, washed with a mixture of toluene - hexane (1 :1 ), dried at 60 °C to give the product >99% rich in the erythrose, which showed [α o +48° (C=1 , CHCI₃) mp = 121-122°C .

EXAMPLE - 4

Preparation of 2-deoxy-2,2-difluoro-D-ribo furanose-3,5-dibenzoate

In a 2 lit four-necked round bottom flask fitted with a stirrer, condenser carrying a guard tube, addition funnel carrying a nitrogen inlet system and thermometer socket, was taken 2-deoxy-2,2-difluoro-D-erythropentofuranose-1-ulose-3,5- dibenzoate 100gms, and THF 1 lit. The solution was cooled to about -20°C under stirring. Vitride, 60ml (65% w/w in toluene), was taken in the addition funnel and added dropwise into the flask at such a rate so as to maintain the mass temperature at -20 °C. After the addition, the reactor mixture was stirred at -20 °C for 1 hour. TLC (Mobile phase-Ethyl acetate: Pet. Ether: 2:8) was checked for completion of the reaction. HPLC analysis in Zarboax CN using hexane + isopropyl alcohol 92ml + 8ml indicated the product as two peaks. On completion, the reaction mixture was allowed to warm upto 0 °C and treated with methanol, 15ml. Later 10% aqueous hydrochloric acid, 300ml, was added into the reaction mass and stirred for 30min. The organic layer was separated. The aqueous layer was extracted with ethyl acetate 500ml. The combined organic layer was washed with 10% NaHCO₃ solution, 300ml. The organic layer was separated and then washed with saturated sodium chloride solution, 300ml. Finally organic layer was separated, dried over anhydrousNa₂SO4, filtered and concentrated under reduced pressure at 50°C. This was used in the next stage to prepare the mesylate of formula VI. EXAMPLE - 5

Preparation of 2-deoxy-2,2-difluoro-D-ribofuranose-3,5-dibenzoate-1 -methane sulfonate

In a 2-lit four-necked round bottom flask fitted with a stirrer, condenser, addition funnel and a thermometer socket, was added the lactol, 100gm, prepared as given in example 4, and dichloromethane, 1 lit. The solution was cooled to 0 °C under stirring and triethyl amine, 54ml, was added. The solution was stirred at 0 °C for 30min and a solution of methane sulfonylchloride, 23.3ml, in dichloromethane, 70ml, was added dropwise maintaining the mass temperature at 0 °C. Later the reaction mass was allowed to warm upto 20 °C to 25 °C and stirred further for 12-14hours. The reaction mixture was analyzed by TLC (Mobile phase-Ethyl acetate: Pet. Ether: 2:8). It was also analyzed by HPLC in Zarboax CN using hexane + isopropyl alcohol 92ml + 8ml indicated the product as two peaks. The reaction mixture was washed with 5% aqueous sodium bicarbonate solution, 325ml, saturated sodium chloride solution, 325ml and separated. The dichloromethane extract was dried over anhydrous sodium sulphate, filtered and then concentrated to give 110 gms of the title product. This was used for the Verbruggen protocol.

EXAMPLE - 6

Preparation of 2'-Deoxy-2',2'-difluorocytidine

In a 5 lit three-necked round bottom flask fitted with a stirrer, condenser and a stopper, acetyl cytosine 97.36gms, 1 ,2 dichloroethane 2lit, hexamethyl disilazane, 169ml and trimethyl silyl chloride, 5ml, were charged and the reaction mixture was heated to reflux under stirring. The reaction mixture slowly attained clarity and from that point the mixture was stirred at reflux for another 3 hours. Then dichloroethane was completely distilled off and fresh dichloroethane, 1250ml, was charged into the flask and the reaction mixture was cooled to 20 °C to 25 °C. The reaction system was maintained under nitrogen atmosphere and trimethyl silyl trifluoromethane sulfonate, 113ml, was added. Then 2-deoxy-2,2-difluoro-D-ribofuranose-3,5-dibenzoate, -1 -methane sulfonate, 100gms, was also added into the flask. The reaction mixture was heated to 70 °C to 75 °C and maintained for 12-14 hours. Subsequently the reaction mixture was stirred under reflux for 2 hours. The progress of the reaction was checked by TLC (dichloromethane: methanol 9:1 ). The reaction mixture was cooled to 20 °C to 25 °C and 5% aqueous hydrochloric acid, 600ml, was added dropwise in about 30minutes. After the addition, the reaction mixture was stirred for 15minutes and allowed the layers to separate. The aqueous layer was washed with dichloroethane 2000ml and the combined dichloroethane extract was washed with a saturated solution of sodium chloride, 300ml. Then the organic layer was separated and dried with anhydrous sodium sulphate. The dichloroethane solution was filtered and concentrated under reduced pressure at 50 °C. The residue was dissolved in 1.Slit methanol and cooled to about 0 °C to 5 °C. Ammonia gas was bubbled into the methanolic solution for about 8 hours. Then the solution was brought to about 20°C to 25°C and ammonia gas was bubbled for further 8 hours. The reaction was checked by TLC for completion. The methanolic solution was treated with carbon, 15gms and filtered. The filtrate was concentrated under reduced pressure at 45°C. The residue was dissolved in about 400ml water and the aqueous solution was washed first ethyl acetate, 100ml and then with hexane, 100ml. The ethyl acetate and hexane extracts were separately washed with 50ml each of water and the water extracts were combined with the main aqueous solution. The aqueous extract was stirred with carbon, 10gms, filtered through a bed of hyflow and the hyflow bed was washed with water, 50ml. The aqueous solution was then concentrated. The residue was dissolved in 1500ml isopropyl alcohol and then the solution was subjected to distillation. The addition of isopropyl alcohol and distillation was repeated two more times by adding 750ml of alcohol each time. Finally isopropyl alcohol, 700ml, was added to the residue and the solution was warmed to about 70°C. To the warm aqueous isopropyl alcohol, concentrated hydrochloric acid, 75ml, was added drop wise in about 30min. The solution was stirred at 70°C to 75°C for about 30min, then cooled to RT and stirred at RT for 12 hours. Later the solution was cooled to 0°C to 5°C, stirred at that temperature for 3 hours, the solid formed was filtered, washed with acetone and dried to give gm of product, which was >95% rich in the β-anomer and showed [α]o²⁵ = +46° (C= 1.0 D ₂O)

EXAMPLE - 7

Purification of 2'-deoxy-2',2'-difluorocytidine

Method A

50gms of the product of about 95% rich in β-anomer was stirred with 50ml DM water at room temperature for about 30minutes in a 250ml three-necked round bottom flask. The solid material was filtered, washed with acetone and pressed dry to give 40gms of product.

The solid, obtained above, was taken in about 190ml water in a 500 ml three- necked round bottom flask fitted with a stirrer, condenser and a stopper. The slurry was heated to 70°C to 75°C in order to dissolve the solid. The solution was filtered to remove any insoluble material and the filtrate was transferred into a 3 lit three-necked flask fitted with a stirrer, addition funnel and stopper. The aqueous solution was stirred at 25°C to 30°C and acetone 2.3 lit was added drop wise from the addition funnel. After the completion of addition, the slurry was stirred at 25°C to 30°C for 1 hour and the solid was filtered. It was washed with acetone and dried at 60°C to give 32 gms of very pure product. HPLC purity 99.9%, mp 288-291 °C, [α]_D ²⁵ = +48° (c= 1.0 D ₂O)

Method B

50gms of the product of about 95% rich in β-anomer, was stirred with 50ml DM water at room temperature for about 30minutes in a 250ml three-necked round bottom flask. The solid material was filtered, washed with acetone and pressed dry to give 40gms of product

The solid, obtained above, was taken in about 190 ml of water in a 500 ml three-necked round bottom flask fitted with a stirrer, condenser and a stopper. The slurry was heated to about 70°C to 75°C in order to dissolve the solid. The solution was filtered to remove any insoluble material and the filtrate was transferred into a 3 lit ml three-necked round bottom flask fitted with a stirrer, addition funnel and stopper. The aqueous solution was stirred at 25°C to 30°C and acetonitrile, 2.3 lit was added drop wise from the addition funnel. After the completion of addition the slurry was stirred at 25°C to 30°C and the solid was filtered, washed with acetone, pressed dry and dried at 60°C to give 36gms of very pure product. HPLC purity 99.92%, mp 287-290°C, [α]_D ²⁵ = +48 ° (c= 1.0 D ₂O)

Method C

50gms of the product about 95% rich β-anomer was stirred with 50ml DM water at room temperature for about 30minutes in a 250ml three-necked round bottom flask. The solid material was filtered, washed with acetone and pressed dry to give 40gms of product.

The solid, obtained above was later dissolved in about 190 ml water in a 500 ml three-necked round bottom flask fitted with a stirrer, condenser and a stopper. The slurry was heated to about 70°C to 75°C in order to dissolve the solid. The solution was filtered to remove any insoluble material and the filtrate was transferred into a 500ml three necked round bottom flask fitted with a stirrer, addition funnel and stopper. The aqueous solution was stirred at 25°C to 30 °C and isopropyl alcohol 2.2 lit was added dropwise from the addition funnel. After completion of addition, the slurry was stirred at 25°C to 30°C for one hour and the solid was filtered, washed with first with isopropyl alcohol, then with acetone and dried at 60°C to give 36gms of very pure product. HPLC purity 99.92%, mp 287-290°C, [α]_D ²⁵ = +48° (c= 1.0 D ₂O)

Claims

We claim 1. A process for the preparation of Gemcitabine hydrochloride of formula I of extra high purity by a) Hydrolytic cyclization of the product of formula IV. b) Converting the product of formula IV to the dibenzoyl derivative of formula V. c) Purifying the product of formula V by a suitable process. d) Reducing the product of formula V and converting the resultant lactol into a mesylate of formula VI e) Followed by coupling the mesylate of formula VI with bis-silyl acetyl cytosine of formula X and subsequently deblocking and purifying.

2. A process for the preparation of Gemcitabine hydrochloride as claimed in claim 1 , wherein the product, after benzoylation, consisting of a mixture of erythro, threo and a γ hydroxy acid of formula XIV (isolated and characterized) is purified by treatment with benzene/toluene containing benzene or p-toluene sulphonic acid under azeotropic condition, followed by crystallization in a mixture of benzene / toluene and hexane to give a product of formula V

3. A process for the preparation of Gemcitabine hydrochloride as claimed in claim 1 , wherein the reduction of the product of formula V is carried out preferably using Vitride in a solvent like glyme, diglyme, THF or dioxane and at a temperature of -30°C to +30 °C to give a product of formula Va

4. A process for the preparation of Gemcitabine hydrochloride as claimed in claim 1 , wherein the mesylation of the reduced product is carried out in a solvent like dichloromethane or ethyl acetate using methanesulfonylchloride in the presence of a base like triethylamine/N- methyl morpholine/diisopropyl ethylamine at a temperature of 0°C to50 °C

5. A process for the preparation of Gemcitabine hydrochloride as claimed in claim 1 , wherein the mesylated product of formula VI is coupled with bis (trimethylsilyl)-N-acetylcytosine of formula XI in a solvent like dichloroethane using trimethyl silyl trifluoro methane sulfonate at 60 °C to 85 °C for a duration of 3 to 24 hours

6. A process for the preparation of Gemcitabine hydrochloride as claimed in claim 1 , wherein the coupled product is deblocked using a base like ammonia in a polar solvent like methanol.

7. A process for the preparation of Gemcitabine hydrochloride as claimed in claim 1 , wherein the purification is carried out by isolation of the crude product in aqueous isopropyl alcohol containing hydrogen chloride

8. A process for the preparation of Gemcitabine hydrochloride as claimed in claim 1 , wherein the purification is carried out by leaching the crude product with water and crystallizing the water leached product in a mixture of ethanol, isopropyl alcohol, acetonitrile or acetone and water.