Paclitaxel is a well known antitumor agent and has beer, approved by the Food and Drug Administration for treatment: of ovarian and breast cancer. This drug is also presently undergoing clinical trials for treatment of other types of cancer. The world-wide supply of paclitaxel, however, is limited to a finite number of yew trees and other yew species containing relatively small amounts of paclitaxel of which there is a serious shortage for human and animal tumor treatment, and as well as for use in routine bioactivity testing in the development of antitumor agents having paclicaxel-like antitumor activity. Thus, alternate sources of paclitaxel as well as alternate compounds having pacii_axei-like antitumor activity are highly desired.

Paclitaxel is most often present in combination with its well known and structurally similar taxane, cephalomannine. The structures of cephalomannine and paclicaxel are shown below (I) .

Paclitaxel

Cepha1omannine

Paclitaxel and cephalomannine are natural products found in the bark of the Pacific yew tree Taxus brevifolia , and other yew species including T. bacca ta, T. cuspidata , T. yunnane sis , T. chinensis, T. capi tata , T. brownii and T. dark green spreader . These compounds can also be found in Cephalotax s species such as Cephalotaxus mannii as well as cultured plant cells and fungi.

Cephalomannine has been reported to be effective in causing the remission of leukemic tumors. See U.S. Patent No. 4,206,221.

In accordance with this invention it has now been unexpectedly discovered that certain novel paclitaxel analogs, specifically 2", 3" side-chain halogenated cephalomannines show strong in vitro and in vivo paclitaxel-like efficacy in a variety of tumors thereby providing a viable alternative to paclitaxel and paclitaxel derivatives, such as Taxotere ^® . The chemical structures of both cephalomannine and paclitaxel contain eleven asymmetric carbon atoms, of which nine are in the taxane ring and two are in the side chain at carbon 13. Stereostructures of cephalomannine and paclitaxel are shown belcw (II) :

Stereoview of Taxanes

Paclitaxel R =

Cephalomannine; R =

The exocyclic 2", 3" side-chain double bond in cephalomannine along with the number of stereocenterε present in the structure of this compound suggests the possibility of the existence of numerous stereoisomers of this taxane. For example, cephalomannine can be distributed in two isomeric forms wherein the hydroxyl group at carbon 13 is acylated with phenylisoserine acylated in amino group by either (Z)- or ( E) - 2-methyl-2-butenoic acid leading to (Z) - and (J?) -cephaloman¬ nines, respectively. In addition, it is known that cephalo- mannine and paclitaxel can be epimerized at carbon 7 either thermally, during chromatographic procedures or in acidic or basic solutions to produce 7-epi-cephalomannine, which is shown below (III) . Miller, et al . , J. Org. Chem . , 40:1469 (1981) ; Chaudhary, et al . , J. Org. Chem . , 58:3978, (1993) ; and Wender, et al . , CRC Press, Inc., Boca Raton, Fla. , . (1995) . Thus, during halogenation the 2", 3" side chain positions can give rise to a mixture of diastereomeric products.

Therefore, in addition to that set forth above the invention provides isolated and purified diastereomers of 2" , 3" -dihalocephalomannine and 2" , 3"dihalo-7-epi- cephalo annine, which show strong antitumor efficacy.

Therefore, in addition to that set forth above the invention provides isolated and purified diastereomers of 2" , 3 "-dihalocephalomannme and 2" ,3 "dιhalo-7-epι- cephalo annine, which show strong antitumor efficacy.

7-epi-cephalomannine

Detailed Description of the Invention with Preferred EmbodimentsX

The present invention provides novel analogs of paclitaxel, specifically isolated and purified 2", 3" dihalocephalomannme and 2", 3" dihalo-7-epι-cephalommannιne diastereomers, which show strong m vi tro and m vivo paclitaxel-like antitumor activity in a variety of tumor cell lines. The invention a. so provides methods for the preparation of these compounds and their use in tumor treatment .

In accordance with this invention, the diastereomeric mixture of dihalocephalomannme analogs are prepared in good yields from either relatively refined sources of cephalomannine or from complex unpurified mixtures compris- ing cephalomanine, paclitaxel and other taxane compounds. The analogs are prepared by selective halogenation of the unsaturated side-chain of the cephalomannine molecule, while leaving other portions or moieties of the molecule or other important taxane compounds in the mixture, such as paclitaxel, intact.

Separation and purification of individual 2", 3"- dihalocephalomannine/dihalo-7-epi-cephalomannine diastereomers from the mixture is accomplished by conventional methods, and these compounds also show strong an i-tumor efficacy. The selective halogenation is carried out by reacting cephalomannine and/or 7-epi-cephalomannine under conditions inclusive of a temperature and time effective to selectively halogenate the 2", 3" side-chain portion of these compounds, and then separating the resulting less polar mixture of dihalocephalomannine/dihalo-7-epi-cephalomannine diastereomers from paclitaxel and other taxane compounds. Individual diastereomers can be isolated from the mixture and purified by standard chromatographic techniques and/or recrystallization. The synthetic methods of this invention are advantageously independent of the concentration of cephalomannine and 7-epi-cephalomannine present in various complex or more refined mixtures of taxane compounds and can utilize any source containing cephalomannine and/or 7-epi- cephalomannine as starting material. Representative examples of sources include the bark from various Taxus species, such as Taxus brevifolia , Taxus bacca ta , Taxus yunnanensis , Taxus chineneεis and Taxus wallichiana ; from Cephalotaxus species such as Cephalotaxus mannii , plant material; leaves, needles and twigs from various Taxus and Cephalotaxus species, extracts of biomass containing a complex mixture of taxane type compounds, as well as in the downstream purification of cephalomannine and 7-epi-cephalomannine produced from sources

such as cell cultures of Taxus and Cephalotaxus species and cephalomannine-producing fungi.

In one example of this invention, a mixture of taxanes comprising cephalomannine and/or 7-epi-cephalomannine in addition to paclitaxel is treated with stoichiometric quantities of halogen, such as for example, bromine or chlorine dissolved in an inert solvent, preferably a chlorinated solvent such as carbon tetrachloride, chloroform, methylene chloride or ethylene dichloride. In a typical treatment, for example, using a mixture containing approximately 30 wt. % cephalomannine with halogen in carbon tetrachloride results in a quantitative yield of a mixture of 2" , 3"-dihalocephalomannme diastereomers and the corresponding 2" , 3" -dihalo-7-epi-cephalomannine diastereomers. The general reaction scheme (IV) is as follows:

wherein,

[2"R, 3 "S) -dihalocephalomannine

II . (2" S , 3" R) -dihalocephalomannine

III . (2"R, 3 " S) -dihalo- 7 -epi -cephalomannine

IV. (2 " S, 3 "R) -dihalo-7-epi-cephalomannine

and X = halogen.

The resulting pure dihalo-diastereomers I-IV can be separated and their chemical structures elucidated by conventional analytical and physicochemical techniques. Further in accordance with this invention, for mixtures containing cephalomannine and/or 7-epi-cephalomannine and from about 0.01% wt . to about 95.05% wt . paclitaxel, the process is similar to that described above. The mixture is first dissolved in an inert solvent, preferably carbon tetrachloride, chloroform, 1,2-dichloroethane or methylene chloride which is reacted with a halogen, for example, a solution of bromine or chlorine in an inert chlorinated solvent, and the reaction stirred until cephalomannine is completely reacted. It is preferred that the reaction be run at temperatures between -20"C and 20°C and more preferably between -5°C and 5°C, preferably in the dark. The preferred halogen solution is bromine or chlorine in carbon tetrachloride of from 0.01M to 0.1M. To ensure that reaction conditions favor the production of the desired 2", 3"- dihalocephalomannine and/or 2" , 3"-dihalo-7-epi-cephalomannine diasteromeric reaction products, reaction progress can be conveniently monitored by conventional analytical techniques, for example, HPLC, and the appropriate reaction conditions maintained.

The reaction mixture containing taxane impurities can then be separated and purified by conventional methods such as chromotography and recrystallization and the individual separated and purified diastereomers made available for antitumor treatment.

Conventional wisdom would lead one to expect that the use of halogen in the presence of taxane compounds having several functional groups would result in undesired side reactions, thereby depleting the concentration of cephalomannine and/or 7-epi-cephalomannine and halogen without generating the desired dihalocephalomannines, or appreciable yields thereof. It would also be expected that other valuable taxanes such as paclitaxel would be degraded by such halo¬ genation. However, in accordance with this invention it has been found that selectivity for halogenation of the 2", 3"- side-chain double bond in cephalomannine and 7-epi-cephaloman¬ nine is very high under controlled conditions, with paclitaxel neither significantly degraded nor halogenated. As mentioned above, any undesired degradation or reaction products during halogenation can be avoided and the effective conditions adjusted appropriately without undue experimentation by monitoring the reaction, for example, by HPLC.

The molar equivalents of halogen used in this invention are dependent upon cephalomannine and/or 7-epi- cephalomannine content and presence or absence of other unsaturated compounds. In general, a less pure mixture, i.e. a mixture containing large amounts of unsaturated taxanes relative to cephalomannine and 7-epi--cephalomannine will require a higher molar equivalent of halogen to halogenate all or substantially all of the cephalomannine and/or 7-epi- cephalomannine present in the mixture. Structures of various other unsaturated taxanes typically present along with cephalomannine, 7-epi-cephalomannine and paclitaxel in plant extracts are shown below (V) :

_• ψΛQnMXO 2. ceρluuofl_____t_H-

(V)

I. pa ilaid 3. 2 ■ , 3--dihalocephalomannine

_-_p*i-(ma

The following examples are provided to illustrate preferred embodiments of the invention, specifically selective bromination and chlorination of samples containing cephalomannine, 7-epi-cephalomannine, paclitaxel and other taxanes, all present in varying amounts, and without significant undesirable reactions and/or degradation, for example, of paclitaxel. Examples are also provided which demonstrate the antitumor efficacy of the inventive dihalocephalomannine/dihalo-7-epi-cephalomannine compounds. These examples are only intended for the illustration of some preferred embodiments of this invention, and are not intended to limit the scope of the invention as defined by the claims.

EXAMPLE 1

BROMINATION OF A PARTIALLY PURIFIED MIXTURE CONTAINING CEPHALOMANNINE

A solution of 0.63g of 91.5% cephalomannine (0.0007 moles) , also containing about 6-7% paclitaxel, dissolved in 150 ml carbon tetrachloride was added to a 500ml three neck round bottom flask fitted with a 250 ml separatory funnel.

The flask was then immersed in an ice-salt bath. When the temperature reached -5°C, a solution of bromine (0.1221 g) in carbon tetrachloride (76.31 ml, 0.01 M) was added slowly with stirring at such a rate that the reaction temperature did not exceed 5 C. The cephalomannine to bromine ratio was 1:1.1 mole. This addition required about three hours and the resulting solution was light brown and cloudy.

The bromination was monitored by HPLC analysis every hour. The reaction was completed when all the cephalomannine present was converted to the 2", 3"-dibromoderivative, which, based on HPLC analysis, required approximately 8 hrs. The reaction mixture was light yellow to colorless, due to the consumption of the bromine, in contrast to the darker starting solution. The reaction mixture was then transferred to a one litre separatory funnel and first washed with 0.5% aqueous sodium sulfite (300 ml) , 0.5% aqueous sodium bicarbonate (300 ml) and then twice with deionized water (200 ml each) to a final pH 6.5. The combined aqueous layer was extracted once with CH ₂ C1 ₂ and the CH ₂ C1 ₂ layer mixed with the previous organic extract. The organic layer was next dried over Na ₂ S0 ₄ , filtered, and evaporated to dryness. The yield was 0.76 g of a light cream-colored solid which is approximately a 100% yield based on the starting material. The cream colored solid material was chromatographed on a column of silica gel (50g, ICN Silitech, 32-63 D, 60 A) using the solvent mixture acetone :CH ₂ C1 ₂ (10:90) as the eluent. Fifty ml fractions were collected and checked by TLC (Silica gel 60 F ₂₅₄ , Merck #5554, developed with acetone/CH,Cl ₂ : 20/80, detected using vanillin-sulfuric acid in methanol spray reagent) . The fractions with a single spot at R _£ = 0.64 (fractions #26 - #38) were mixed, concentrated to dryness to yield 0.485 g of a light cream powder, which was recrystal¬ lized to white crystalline solid, mp 158^, and identified as 2", 3" -dibromocephalomannine by physico-chemical methods (TLC, HPLC, UV, IR, NMR, MS) . The yield was estimated to be 70% on the basis of starting cephalomannine.

EXAMPLE 2

BROMINATION OF A CRUDE MIXTURE CONTAINING

CEPHALOMANNINE, PACLITAXEL AND

OTHER TAXANE-TYPE COMPOUNDS

Using similar apparatus as used in Example 1, a sample of crude paclitaxel (2.0 g) having a mixture of 51.2% paclitaxel 28.8% cephalomannine, and about 20% other taxanes or non-taxane impurities based on HPLC was dissolved in 150 ml carbon tetrachloride and 150 ml CH ₂ C1 ₂ , to yield a clear, light yellow solution. The flask was immersed in an ice-salt bath and stirred. When the temperature reached -5°C, a solution of 0.1332 g 100% bromine in 83.13 ml (0.01 M) of carbon tetrachloride (1 mole cephalomannine : 1.2 moles bromine) was added to the solution at such a rate that the temperature of the reaction mixture did not exceed 5°C. The addition required about three hours and resulted in a cloudy, brownish-yellow solution. After the addition of bromine was completed, the reaction was allowed to continue under the same conditions for an additional 8 hours, with HPLC analysis of the paclitaxel and cephalomannine performed every hour. The reaction was complete when the solution is colorless or light yellow and all the cephalomannine has been converted to the dibro o derivative. If after the additional 8 hours the solution still contained more than 1 - 2% cephalomannine, keeping the initial conditions, 10 ml 0.01 M bromine in carbon tetrachloride was added dropwise and allowed to react for 1 hour before analyzing again with HPLC.

Excess bromine from the reaction mixture was removed by washing with 0.5% aqueous Na ₂ S0 ₃ (300 ml), 0.5% aqueous NaHCOj (200 ml) . and deionized water (2x200 ml) . The reaction mixture was dried using anhydrous Na ₂ S0 ₄ and concentrated to dryness under high vacuum to yield 2.35 g of dry light cream to white powder. The dry material was then purified on a silica gel column under the conditions listed in Example 1. The ratio between the mixture to be separated and the silica gel was 1: 60, thus 120 g silica gel were used. Each fraction was checked by TLC and every third fraction by HPLC. Frac-

mixture was dried using anhydrous Na ₂ S0 ₄ and concentrated to dryness under high vacuum to yield 2.35 g of dry light cream to white powder. The dry material was then purified on a silica gel column under the conditions listed in Example 1. The ratio between the mixture to be separated and the silica gel was 1: 60, thus 120 g silica gel were used. Each fraction was checked by TLC and every third fraction by HPLC. Frac¬ tions with the same R _f in TLC and same retention time in HPLC were mixed to afford two combined fractions. Fractions (#25- #39) which showed a single TLC spot with R _f 0.64 represented dibromocephalomannine and fractions (#41 - #81) which showed a single TLC spot with R _f 0.49 represented paclitaxel.

Fractions #25 - #39, after concentration to dryness at about 40°C under high vacuum, yielded a white to light yellow solid, 0.460 g, (66.6% theoretical yield) with a m.p. 158' - 160°C (chromatographic purity 96.19%) as determined by TLC.

TLC materials were employed as follows: R _f = 0.64 (single spot) on Silica gel 60 F ₂₅₄ Plate (Merck, #5554) Solvent system: acetone : CH ₂ C1 ₂ (20:80)

Spray Reagent: vanilin/sulfuric Acid in methanol

Mass Spectrum [FAB] * of the obtained dibromocephalomannine:

[M + H] ^* = 990, 992, 994 [M + Na] ^* = 1014

[M + K] ^* = 1030

Concentration of the second combined fractions (#41 - #81) yielded 1.16 g (>100% theoretical yield) paclitaxel, which was recrystallized using 50 : 50 acetone/hexane, filtered, washed with the same ratio of cooled solvent and dried under high vacuum at 40°C for 24 hrs. The yield was 0.902 g (45.11% based on the starine material and 88.08% based on the HPLC analysis of paclitaxel in the starting material) of a white crystalline material with a m.p. of 214°C - 216°C.

TLC analysis materials: R _f = 0.49 in the presence of authentic sample on silica gel 60 F ₂₅₄ plate [Merck #5554] Solvent system: acetone/CH ₂ Cl ₂ (20:80) Spray Reagent: vanilin/sulfuric acid in methanol Both the UV and the IR spectra of the resulting material match those of pure paclitaxel thereby demonstrating the high selectivity of the bromination reaction for the 2", 3" unsaturated side chain positions of cephalomannine while leaving its close analog paclitaxel untouched.

EXAMPLE 3

SCALED-UP EXAMPLE ILLUSTRATING BROMINATION OF A CRUDE MIXTURE CONTAINING CEPHALOMANNINE

A solution of 10.00 g crude paclitaxel (on the basis of HPLC analysis the content was 28.8% cephalomannine, 51.2% paclitaxel and approximately 20% other taxane or non-taxane impurities) was dissolved in 1.5 1 carbon tetrachloride in a 2 1 three-necked flask fitted with a 500 ml separatory funnel, reflux condenser, thermometer and magnetic stirrer and immersed in an ice-salt bath. The reaction mixture was stirred until the temperature reached -5 ^U C and then 41.2 ml of 0.1 M bromine (0.665g bromine) in carbon tetrachloride was added dropwise for about 3 hours. The molar ratio between cephalomannine and bromine was 1 : 1.2. The temperature did not exceed 5"C. After the bromine addition was completed, stirring was continued while maintaining the temperature at -1°C to 5°C. The reaction was monitored by HPLC every hour until all the cephalomannine had been converted to the dibromo derivatives (approximately 8 hrs.) . The final color of the 1500 - 1600 ml of solution was light yellow or cream, depending on the color of the starting mixture and the possible presence of a small excess of bromine.

To remove any trace of bromine, the reaction mixture was washed with 0.5% aqueous Na ₂ S0 ₃ (500 ml) , 0.5% aqueous NaHC0 ₃ (500 ml) , and deionized water (2x500 ml) . The reaction mixture was next dried with anhydrous Na ₂ S0 ₄ and concentrated

to dryness under vacuum to yield 13.20 g of a light cream to white solid material .

This material was chro atographically separated on a silica gel column under the conditions listed above in Examples 1 and 2. A 100 x 5 cm glass column was prepared by the slurry method with 600 g silica gel (ratio 1:50) . The column was eluted with acetone/CH ₂ Cl ₂ (10 : 90) . One 1 of acetone/CH ₂ Cl ₂ (25 : 75) was used as a final column wash. Every fraction was analyzed by TLC and every third fraction by HPLC. Fractions #11 - #22 had a single spot at R _f = 0.64 and their combination, concentration and drying (40°C, high vacuum) , yielded 3.25 g (95%) of 2" , 3 " -dibromocephalomannine as a white to light yellow solid.

Analysis of this compound is as follows .- m.p. : 158 - 160oC.

R _f = 0.64 (single spot) on silica gel 60 F _2S4 plate [Merck #5554] .

Solvent system: Acetone/CH ₂ Cl ₂ (20 : 80) Spray Reagent : Vanilin/Sulfuric Acid in Methanol.

Elemental Composition and Molecular Weight (on the basis of HR FAB ⁺ )

C _4S H ₅₄ N0 ₁₄ ⁷⁹ Br ₂ [M + H] ⁺ :

Calculated: 990.191000 Found: 990.191103 Urn = 0.1 ppm)

C ₄₅ H ₅₄ N0 ₁₄ ⁷⁹ Br ^eι Br [M + H] ^* :

Calculated: 992.181000

Found: 992.189057 m = 8.1 ppm)

C ₄₅ H _S4 N0 ₁₄ ^βl Br ₂ [M + H] ⁺ : Calculated: 994.175000

Found: 994.187011 Um = 12.1 ppm)

C _4E H _S3 N0 ₁₄ Na ⁷⁹ Br ^fll Br [M + Na] ⁺ : Calculated: 1014.161000

Found: 1014.171002 (ΔΠI = 9.9 ppm)

C ₄₅ H _S3 N0 ₁₄ K ⁷⁹ Br ⁸¹ Br [M + K] ^* :

Calculated: 1030.097000

Found: 1030.144940 Um = 46.5 ppm)

[c_] _D ²⁵ =-40.207° (c 0.29, MeOH)

5 UV Spectrum in CH ₃ OH [λ _max nm, (e) ] : 274.2 (1550.8) ; 227.1 (18610.4) ; 221.8 (18325.1)

IR Spectrum in KBr (cm ^"1 ) 3500, 1105, 1070 (tert & sec OH)

3420, 1670, 1580 (-CONH-) 3110, 3060, 1605, 1505, 770, 710 0 (monosubt . aromatic cpds . )

3060, 2960, 2915, 2870, 1465,

1370

(-CH ₃ , -CH ₂ -, =CH-)

3020, 1670, 1310, 980 (double 5 bond)

1730, 1270 (aromatic esters) 1715, 1240 (>C=0) 1730, 1180 (acetates) 855 (epoxy rings) 0 520 (bromo compounds)

-H NMR in CDC1 ₃ (300 MHz) : 1.94 (d, 3H, -COC (Br) CH ₃ -5") (ppm; side chain protons only) 1.98 (d,3H, -HC(Br)CH ₃ -4")

4.63 (qt, IH, >CH(Br) -3")

¹³ C NMR (300 MHz) 170.21 and 170.25 (C-1' ) 5 (in ppm; side-chain C only)

72.76 and 72.90 (C-2' )

172.26 and 172.32 (C-1")

54.34 and 54.52 (C-3' )

69.71 and 69.88 (C-2") C 55.13 and 55.35 (C-3")

30.39 and 30.77 (C-4")

27.21 and 27.62 (C-5")

El -MS 568, 551, 509, 491, 449, 431, 405, 391, 386, 329,

(m/z)

(the main fragments)

326, 308, 278, 264, 245, 217, 200, 188, 159, 149, 122,105,91,83,77,55,43.

DCI-MS(_τ?/z) (the main fragments)

569,552,510,492,474,450,432, 424, 392, 387, 370, 329, 327, 309, 279, 265 264, 246, 218, 200, 188, 167, 149, 125, 124,

106,101,100, 1,83, 69.

FAB' - MS: 1030 [M + K] \- 1014 [M + Na] \- 992

(positive ion mode) [M + H] ^* (See Elem. Anal.) ; 974 [M-H ₂ 0] ( /z) 932 [M-AcOH] ^* ; 914 [M-AcOH-H ₂ 0] ^* ; 912 [M-HBr] ⁺ ; 870 [M- BzOH] \- 854 [870-

H ₂ 0-2H] ; 832 [M2-HBr]\- 705 [M-243- Ac]\- 569 [T]\- 551 [T-H ₂ 0] ; 509 [T- AcOH]\- 491 [T-AcOH-H ₂ 0] \- 448 [T- BzOH] ^* ;429;424 [SH ₂ ] ^* ; 413; 405 [S- H ₂ 0] ^* ; 391 [S-0-H ₂ 0]\-

387 [T-AcOH-BzOH] \-376;347 [S-0- CO-HCHO]\- 338:327 [387-T-AcOH] ⁴ ; 315; 284 [327 -Ac] \ 279 ; 264 [832 -T] ⁺ or [424-2HBr] \-246 [264 -H ₂ 0] ⁺ ; 231 ; 218 [264-HCOOH] \-188; 167 [S-C ₅ H ₈ 0NBr ₂ ] ⁺ ;

149 [167-H ₂ 0] ^* ;133 ;122 [BzOH] \- 113:105 [Bz] ⁺ ; 91[C ₇ H ₇ ] ^* ; 83; 77 [C ₆ H ₆ ] \- 76; 57; 55;

(T=taxane ring in the compound; S- acid (side) chain in the compound. )

HPLC: Condition 1: Column CN lOμ (250x4.6mmn)

Solvent System CH ₃ CN:H ₂ 0 (40 : 60)

Flow Rate 1 mL/min Detector Waters 490uv at 227nm

Injection volume 20μL T. _" , 3 "-dibromocephalomannine 26.06 m.

Condition 2: Column Curosil G 6μ (250x3.2mm) Solvent System CH ₃ CN:H ₂ 0 (45:55) Flow Rate 0.75 mL/min Detector Waters 490uv at 227 nm Injection Volume 20 μL

^RT 2 _" , 3 --dibromocep alom nnine 2 dias t ereomer i c f orms : RT _X = 23 . 53 RT = 24 . 50

Thermogravimetric Analysis (TGA) : 28°C (100.0%) , 100°C (99.64%) ,

(Temp, and % decomposition) 150°C (98.88%) , 175 ^C C,

(95.35%) ,180°C (86.74%) ,200°C (60.38%) ,250°C (45.03%) .

Differential Scanning Calorimetry (DSC) : 173.76°C, 187.73°C.

As demonstrated from the following analysis, bromination of the crude paclitaxel mixture shows surprisingly high selectivity for the 2", 3", positions of the unsaturated side chain of cephalomannine, while leaving paclitaxel untouched.

The fractions from #26 to #68 which had a single spot in TLC (R _f 0.49, the same as the authentic sample of paclitaxel) and a single peak in the HPLC, were combined, concentrated and dried, (40°C, high vacuum 1mm to 2mm) to yield 6.10 g of a white solid. This material was crystallized from 60 ml of a mixture of acetone/hexane mixture (50:50) , filtered, washed with the same ratio of cooled solvents and dried under high vacuum at 40°C (24 hrs.) to obtain 4.84 g (92%) of a white crystalline solid identified by comparison to an authentic sample as paclitaxel.

Analysis is as follows: m.p. : 214 - 216oC

R _f :0.49 (in the presence of the authentic sample) Silica gel 60 F _2S4 plate (Merck #5554) Solvent system: acetone/CH ₂ Cl ₂ (20:80) Spary Reagent: Vanillin/Sulfuric Acid in Methanol

%N 1.64 1.63

[cc] _D ^2b =-51.104 (c 0.33, MeOH) UV Spectrum in CH ₃ 0H:

(λ_ in nm, e) 227.2 (29824.1) 208.0 26256.3)

IR Spectrum (KBr) (cm ^"1 ) 3500, 1105, 1070 (tert. __ sec. OH) 3430, 1650, 1580 (-CONH-) 1610, 1520, 780, 710 (monosub. aromatic rings) 2950, 2910, 1480, 1450, 1370

(-CH ₃ , -CH ₂ -,>CH-groups) 3020, 1315, 980 (double bond) 1725, 1270 (aromatic esters) 1710, 1240 (>C=0) 850 (epoxy rings)

^* H NMR Spectrum 1.88 (S,10H,C-1) ; 5.66 (d,lH, C-2) ; (300 MHz; CDC1 ₃ ) 3.82 (dd,lH,C-3) ; 2.38 (S,3H, CH ₃ COO (ppm) at C-4) ; 4.94 (dd,lH,C-5) ; 1.88

(ddd,lH,C-6) ; 2.48 (ddd,lH,C-6) ; 2.53 (d,10H,C-7) ; 4.38 (dd,lH,C-7) ; 6.27 (S,1H,C-10) ; 2.23 (S,3H,CH ₃ COO at C-10) ; 6.20 (qt,lH,C-13) ; 2.27

(ddd,lH,C-14) ; 2.33 (dd, IH, C-14) ; 1.13 (S,3H,C-19) ; 1.23 (S,3H, C-18) ;1.78(S,3H,C-18) ; 1.68 (S,3H, C-19) ;4.20 (dd,lH,C-20) ; 4.30 (S,1H,

C-20) ;3.77 (S,1H, C-2' ) ; 4.78 (ddd,lH, C-2' ) , 5.20 (ddd,lH, C-3' ) , 7.10 (d, IH, N-l! 7.30-7.53 (m, 10H,p-&m-protons at aromatic rings A _lf B-, & CJ ,- 7.64 (t,lH,A--p) ;7.72 (dd,2H, C^o) ;8.11(dd,2H,A--o) .

¹³ C NMR Spectrum 79.1 (C-1) ;75.1 (C-2) ; 45.8 (C-3) ;81.2 (300 MHz, CDC1 ₃ ) (C-4) ; 84.4 (C-5) ; 35.6 (C-6) ; 72.1

(ppm) (C-7) ;56.7 (C- 8) ;203.6 (C-9) ,-75.6 (C- 10) ;

133.3 (C-ll) ;141.9 (C- 12) ; 72.3 (C-13) ;35.7 (C- 14) ;43.2 (C-15) ; 21.8 (C- 16) ;26.9(C-17) ; 14.7 (C-18) ; 9.5(C-19) ; 76.5(C-20) ; 73.3 (C-2' ) ;55.1 (C-3' ) ;20.7(CH ₃ CO) at C-10; 22.6(CH ₃ CO at C-4 ) ; 170.3 (CH ₃ £0 at C-10) ; 171.1 (CH ₃ CO at C-4) , -167.0 (ArCO-A ; 167.0 (ArCO-C ; 172.7(PhISCO-) ;129.3 (aC-A ;133.8 (aC-B ;138.1 (aC-C ; 130.3

(o-CA,) 127.0 (o-C,B.) ;127.0 (o-C,C _α ) 128.7 (m-C,A-) ; 128.6 (m-C,B _: ) 129.0 (m-C,C ;133.6 (p-CA

131.9 (p-C _. B.) ; 128.3 (p-C,C.)

EIMS: [M} ⁺ =853 568 [T]\-550[T-H ₂ O]\-508 [T-AcOH] ^* ;490 (m/z, the main [T-AcOH-H ₂ 0] ⁺ ; 448 [T-2AcOH] ⁺ or fragments) [T-BzOH] \-386 [T-AcOH-BzOH] ^* ;

326 [T-BzOH-2AcOH] \-308 [326-H-O] \-286 [M-T] or [S] ^* ;280;268 [S-O] \-240 [S-O-CO] \-

210 [S-O-C0-HC0H]\- 122 [BzOH] \-

105 [Bz]\-91 [C ₇ H ₇ ] \-

77[C ₆ H ₅ ]\-51;43 [Ac] ^* .

DC/MS: [M + H] ⁺ =854 569;551;509 ;492 ; 49 ; 387;327; (m/z; the main 311;287;269 ; 240 ;224 ;222 ;210 ; 165 ; fragments) 149;123 ; 105 ; 92; 71.

FAB MS: (positive ion mode) : (m/z; the main 892 [M+K] \- 876 [M+Na}\- 854 [M+H] \- 569 ; 551 ;523 ,- fragments) 509;495;369;327;286 ;240;210; 177;155;149;

119,-105; 85,-69.

FAB MS: (negative ion mode) :

852 - [M - H] ^" HPLC:

Column: μBondapak Phenyl

Solvent System: CH ₃ CN:CH ₃ 0H:H ₂ 0-132 : 20 :48 Flow Rate: lmL/min

Detector: Waters 490uv at 227 nm Injection volume: 20μL

TGA: 50°C (100.0%) , 205 C (99.86%) , 215°C (99.10%) , 220°C (92.19%) , 250°C (56.66%) ,275°C (45.92%) .

DSC: 210 ^U C.

Water content (%H ₂ 0) : 0.90% (Karl Fischer)

EXAMPLE 4

ISOLATION AND PURIFICATION OF 2" ,3"-DIBROMOCEPHALOMANNINE DIASTEREOMERS

4.1 Raw materials Batches of crude plant extracts from Taxus yunnanensis having approximately 15-40% cephalomannine, 50-70 % paclitaxel, and approximately 20-35 % other taxane/non- taxane components were obtained either from Seattle, Oregon, Western yew ( T. brevifolia) , or from the Peoples Republic of China ( T. yunnanensis or T. wallachiana) . Bromine reagent was obtained from Fisher Scientific. Silica gel used was ICN Silitech, 32-63 um, 60 A, ICN Biomedicals, Inc., Aurora, OH. All solvents used were either HPLC or ACS grade and were obtained from Spectrum Chemical Mfg. Corp. Purified water used was deionized in-house.

4.2 Bromination of Crude Plant Extract

Crude plant extract (10.0 g, 26.4 % cephalomannine) was dissolved in chloroform so that a total of 250 ml solution was obtained. To the solution cooled in an ice bath and continually stirred with a magnetic stirrer was added carbon tetrachloride (4750 ml) . To the cooled solution (4°C) was added dropwise 0.1 M bromine in carbon tetrachloride (40 ml) . HPLC analysis of this mixture indicated a ratio of paclitaxel to cephalomannine peak areas 2.6 to 1. The reaction mixture was stirred in the dark with the temperature gradually rising to 15°C. After 7 hrs of reaction, an additional 7 ml 0.1 M bromine in carbon tetrachloride was added and the reaction continued at 15°C. After an additional 8 hrs of reaction, the final portion of 7 ml 0.1 M bromine in carbon tetrachloride was added and the reaction continued at 15'C overnight (14 hrs) . Subsequent HPLC analysis of the mixture showed a ratio of paclitaxel to cephalomannine peak areas 11 to 1. This ratio increased to 12.3 to 1 after another 7 hrs of reaction. The mixture was then washed with 5000 ml 0.2% aqueous sodium

sulfite solution. The pH of the aqueous layer was 8.0. This was followed by two washes with water (2x5 1) .

The pH of the first and second water washes were 6.5 - 7.0 and 6.0 - 6.5 respectively. The combined aqueous layer 5 was reextracted with 5 1 chloroform. The organic layers were combined, dried with anhydrous sodium sulfate (500 g) , and evaporated to dryness using a rotary vacuum evaporator at 40°C. The solid residue (13.64 g) was purified by chromatography.

4.3 Chromatographic Purification 10 of Brominated Material

The thus obtained brominated material (13.64 g) was purified by medium pressure chromatography using a column (6.9 cm i.d., 70 cm long) packed with silica gel (ICN Silitech, 32- 63 um, 60 A) by the slurry method using 1.5% methanol in 1,2-

15 dichloroethane. The sample dissolved in the same solvent was loaded and eluted at the rate of 50 ml/min. Total 55 fractions (500 ml each) were collected. The fractions were analyzed by TLC, with the TLC plates developed with 10% methanol in 1, 2-dichloroethane and detected with 1% vanillin

20 in 50/50 sulfuric acid-methanol . Dibromo-7-epi- cephalomannines eluted in fractions 10-14 and yielded 1.42 g solids following evaporation of solvents. Likewise, the dibromocephalomannines eluted in fractions 24-28 and yielded 1.64 g solids following evaporation of solvent. Individual

25 diastereomers of dibromocephalomannine and the corresponding 7-epi-cephalomannine were subsequently separated and isolated by semi-preparative HPLC, discussed below in 4.4.

Evaporation of medium pressure chromatographic fractions 34-54 yielded 4.79 g pure paclitaxel, m.p. 214° -

30 216°C, with analytical data determined by UV, IR, HPLC, MS, NMR, which is the same as presented in U.S. Serial No. 08/571, 427.

4.4 Separation of 2", 3" -Dibromocephalomannine and 2", 3" -Dibromo-7-epi-cephalomannine Diastereomers

The final purification of dibromocephalomannine and dibromo-7-epι-cephalomannine diastereomers from other impurities was accomplished by semi-preparative HPLC (Waters Deltaprep 3000) using a Waters Deltapak C18 column, 100A, 19 mm x 30 cm with 50% acetonitrile in water as the mobile phase at a flow rate of 15 ml/mm. Peak elution was monitored using a Waters Lambda Max Model 481 UV detector set at 227 nm. Portions of 200 mg of material dissolved in methanol (2 ml) were injected into the column. Elution of dibromocephaloman¬ nine diastereomer I peaked approximately at 54 mm. and diastereomer II at 56 mm. Likewise, the dιbromo-7-epι- cephalomann e diastereomer III peaked at approximately 104 mm. and the corresponding diastereomer IV peaked at 112 mm. respectively Fractions collected from repeated injections were pooled and evaporated at 40"C under reduced pressure to remove the organic solvent. The crystallized solids were filtered, washed with water, and dried in a vacuum oven at 40°C to yield pure dibromocephalomannine and dιbromo-7-epι- cephalomannme diastereomers. The preparation, separation and structures of the obtained diastereomeric dibromo compounds,

(I) ( 2 " R , 3"S) -dibromocephalomannine, (DiBr-I)

(II) ( 2 " S, 3"i?) -dibromocephalomannine, (DiBr-II) (III) ( 2 " R, 3"S) -dibromo-7-epi-cephalomannine,

DiBr-III and

( IV) ( 2 " S, 3 " R) - dibromo - 7 - epi - cephalomannine , (DiBr- IV) is shown in VI :

(VT)

Paditax l Cephalomannine 7 • epi - cephalomannine

Br ₂

(CCW

Paditaxel 2", 3" • dibromo • cephaloπuaoinc 2", 3" • dibromo • 7 - epi • C-pbalomannin.

Separation

Analogue 1: (2"R , 3"5) - dibromocephalomannine Analogue 2: (2"5 , 3"R) • dibromocephalomannine

Paclitaxel

Analogue 3: (2"R , 3"_T) - dibromo-7-epi-cephalomannine Analogue 4: (2 ^* _ , 3"_?) - dibromo-7-eρi-cephalomannine

Paclitaxel Analogues (Brominated)

Analytical characterization of the diastereomers is as follows :

FIG. 1 is a TLC separation of 2" , 3 " - dibromocephalomannine and 2" , 3 "-dibromo-7-epi-cephalomannine diastereomers (DiBr-I-IV) as summarized below in Table 1.

TABLE 1 lane Compound

(1) DiBr-I

(2) DiBr-II (6) DiBr-III

(7) DiBr-IV

(T) paclitaxel plate: silicagel 60 F ₂₅₄ (Merck #5554) solvent system: a) 10% CH ₃ 0H 1, 2-dichloroethane b) hexane/chloroform/EtOAc/CH ₃ OH

20/60/15/5 reagent: a) UV light b) vanilin/H ₂ S0 ₄ in methanol FIG. 2 is an HPLC chromatogram of a mixture of diastereomers (I) DiBr-I; (II) DiBr-II ,- (III) DiBr-III ; and

(IV) DiBr-IV. Equipment and conditions employed in generating this chromatogram are the following:

Column: ES Industries FSP (pentafluorophenyl) 4.6 mm ID x250 mm, 5 um particle size, 60 A pore size

Solvent System: water/acetonitrile/methanol , 41 : 39 : 20 Flow Rate: 0.50 ml/min., isocratic Detector: Waters 990 photodiode array detector, monitored at 227 nm Injection Volume: 20ul

FIG. 3 are superimposed UV spectra of diastereomers DiBr-I, DiBr-II, DiBr-III and DiBr-IV in CH ₃ 0H. The spectra are summarized below in Table 2.

TABLE 2

ISOMER λ max (nm) 111

DiBr-I 226.0 14732

DiBr-II 226.0 12415

DiBr-III 219.4 37900

DiBr-IV 218.4 20013

Fig. 4 are superimposed IR spectra of diastereomers DiBr-I, DiBr-II, DiBr-III and DiBr-IV in KBr, which are summarized below in Table 3.

TABLE 3

Band, cm ^"1 Functional Groups

3500, 1105, 1070 tert . and sec . OH

3420, 1670, 1580 -CONH-

3110, 3060, 1605 monosubs. aromatic

1505, 770, 710 rings

2960, 2915 2870 -CH ₃ -; -CH ₂ -; -CH- in 1465,1370 aliphatic or cylic comps.

3020, 1670, 1310 double bonds 980

730, 1270 aromatic esters

1715, 1240 >c=o

1730, 1180 acetates

855 oxetane rings

FIG. 5 are El-MS mass spectra of diastereomers DiBr-

I; DiBr-II; DiBr-III and DiBr-IV, which are summarized below.

FIG. 5 EI-MS; [M] ^* =992 (m/z; the main fragments)

568 [T]\- 550 [T-H ₂ 0]\- 508 [T-AcOH]*; 490 [T-AcOH-H ₂ 0] ^* ; 448 [T-2AcOH] ^* ; or [T-BzOH] ^* ; 390 [S-0-H ₂ O]\- 386 [T-AcOH-BzOH] ⁺ ; 348 [S-0-CO-HCHO] ^* ;

326 [T-BzOH-2 AcOH] ^* ; 308 [T-326-H ₂ 0] ^* ; 284 [327-Ac]\- 264 [832-T] ^* ; or [424 - 2HBr] ^* ; 246 [264-H ₂ 0]\- 218 [264-HC00H]\- 188, 167 [S-C ₅ H ₈ ONBr ₂ ] ^* ; 148 [167-H ₂ 0] ⁺ ; 122 [BzOH] ^* ; 105 [Bz] ^* ;

91 [C ₇ H ₇ ] ^* ; 83 [C ₄ H ₇ C__0] ^* ; 77 [C ₆ H ₅ ] ^* ; 57,55. (T = taxane ring in the compound; S = acid (side) chain in the compound. )

FIG. 6 are FAB" - MS mass spectra of diastereomers DiBr-I, DiBr-II, DiBr-III and DiBr-IV, summarized below.

FIG. 6 FAB ^* - MS: (positive ion mode) (m/z)

1030 [M + K] ^* ; 1014 [M + Na] ⁺ ; 992 [M + H] ⁺ (See Elem. Anal.) ; 974 [M-H ₂ 0] \ ^* 932 [M- AcOH] ^* ; 914 [M-AcOH-H ₂ 0]\- 912 [M-HBr] ^* ; 870

[M-BzOH] ^* ; 854 [870-H ₂ O-2H] ; 832 [M-2HBr] \- 705 [M-243- Ac] ^* ; 569 [T] \- 551 [T-H ₂ 0] ; 509 [T-AcOH] ^* ; 491 [T-AcOH-H ₂ 0] ^* ; 448 [T- BzOH] \- 429; 424 [SH-]+; 413; 405 [S- H ₂ 0] ^* ; 391

[S-0-H ₂ O] ^* ; 387 [T-AcOH-BzOH] ^* ; 376; 347 [S-0-CO-HCHO] ⁺ ; 338:327 [387-T-AcOH] \- 315; 284 [327-Ac] ⁺ , 279 ; 264 [832 -T] ⁺ or [424-2HBr] ⁺ ; 246 [264-H ₂ 0] ⁺ ; 231; 218 [264- HCOOH] ^* ;

188; 167[S-C ₅ H _β 0NBr ₂ ]-; 149 [167-H ₂ 0] ⁺ ; 133; 122 [BzOH] ^* ; 113 : 105 [Bz] ^* ; 91[C ₇ H ₇ J ^* ; 83; 77 [C ₆ H ₅ ] ^* ; 76; 57; 55; (T-taxane ring in the compound; S-acid (side) chain in the compound.)

FIGs. 7-10 are ^α H-NMR spectra and FIG. 11 is ¹³ C-NMR spectra of the diastereomers which are summarized below.

DIBr-I 'H-NMR in CDCL ₃ (300 MHz in ppm: side chain and some important protons only) Chemical Shift (ppm) Assignments

I I

3.36 (d, IH) - (HO - C - C= O) (HO 2'a)

I

I 4.74 (d, IH) - (H C C = O) (HO

5.68 (d, IH) - (N - CH - ) (H - 3'

4.62 (qt, IH) >CH Br) -3")

1.81 (s, 3H) - (Br - C - CH ₃ ) (3H-4"

1.78 (d, 3H) (Br - C - CH ₃ ) (3H-5")

I I

- C = O

2.35 (s, 3H) - (- O - C - CH ₃ ) (3H-4) O

2.68 (m, IH) (- CH ₂ -) (H - 6a)

1.98 (m, IH) - (- CH ₂ -) (H - 6a)

4.41 (m, IH) - (- CH -) (H - 7a)

I I

OH

2.46 (d, IH) - (- C - ) (H - 7b)

I I

OH

6.28 (s, IH) - (- CH - O - OCCH ₃ ) (H - 10)

I I

2.22 (m, 2H) - (- CH ₂ - ) (2H - 14, a, b)

2.01 (s, 3H) - (- CH ₃ ) (3H - C - H)

4.22 (qt, 2H) (- CH ₂ - ) (2H - 20a, b)

DIBr-I ¹³ C-NMR

300 MHz in ppm; side chain and some important carbons only

Chemical Shift (ppm) Assignmen s

5 170.3 - (C - 1' ; C = O)

73.0 - (C - 2' )

54.6 - (C -3' )

172.4 - (C - 1"; C = O)

70.1 - (C - 2") 10 55.4 - (C - 3")

22.7 - (C - 4") 27.6 - (C - 5")

203.5 - (C - 9; C = O)

DIBr- II ^X H-NMR in CDCL ₃

(300 MHz in ppm; side chain and some important protons only

Chemical Shift (ppm) Assignments

3.42 (d, IH) (HO - C - C = 0) (HO - 2'a)

4.74 (d, IH) (- HC - C = O) (H - 2b) I I I I

5.68 (d, IH) N - CH) (H-3' )

4.62 (qt, IH) (>CH - Br) (H- 3'

1.81 3H) (Br CH ₃ ) (3H- 4")

1.78 (d, 3H) (Br - C - CH ₃ ) (3H-

- C = O

2.35 (s, 3H) - (O - C - CH,) (3H- 4)

0 I! ^~

2.68 ( , IH) - (-CH ₂ -) (H - 6a) 1.98 (m, IH) - (- CH ₂ -) (H - 6a)

4.41 (m, IH) (- CH) (H 7a)

OH

2.48 (m, IH) (H 7b)

OH

6.28 (s, IH) (- CH - o - OCCH ₃ (H- lo;

I I

1 2.22 (m, 2 H) ( - CH ₂ -) (2H - 14a, b) 2.01 (s, 3H) (- CH ₃ -) (3H - C -18) 4.22 (qt, 2H) (- CH ₂ - ) (2H - 20a, b)

DIBr-II ¹³ C-NMR

(300 MHz in ppm; side chain and some important carbons only

5 Chemical Shift (ppm) Assignments

170.3 - (C - 1' ; C = O) 72.9 - (C - 2')

54.6 - (C - 3')

172.4 - (C - 1") C = O) 10 70.1 - (C - 2")

55.2 - (C - 3")

22.7 - (C - 4") 27.9 - (C - 5")

203.5 - (C - 9; C = O)

DIBr-III ^X H-NMR in CDCL ₃

(300 MHz in ppm: side chain and some important protons only

Chemical Shift (ppm) Assignments

3.23 (d, IH) (HO - C - C = 0) (HO 2 'a)

I I

4.76 (d, IH) (- HC - C =- 0) (H 2b)

I I

I I

5.65 (d, IH) - N - CH) (H-3

4.62 (qt, IH) >CH Br) (H- 3 ¹

1.98 (s, 3H) iBr C CH ₃ ) (3H- 4")

I I

1.28 (s, 3H) (Br - C CH ₃ ) (3H- 5 ¹

C = 0

2.45 (s, 3H) (0 - C - CH ₃ ) (3H- 4) O I

1.72 (t, 2H) (- CH ₂ -) (H - 6a, b)

3.72 (m, IH) ( - CH -) (H - 7a)

I I

OH

4.62 (s, IH) (- C -) (H - 7b)

I OH 6.79 (s, IH) (- CH-) - OCCH ₃ ) (H- 10)

2.42 (m, IH) (- CH ₂ -) (2H - 14a, b)

2.05 (m, IH) (- CH ₂ -) (2H - 14a, b)

2.18 (s, 3H) (-CH ₃ - ) (3H - C - 18) 4.38 (s, 2H) (- CH ₂ - ) (2H - 20a, b)

DIBr-III ¹³ C-NMR

(300 MHz in ppm; side chain some important carbons only)

Chemical Shift (ppm) Assignments

5 169.3 - (C - 1' ; C = 0)

72.9 - (C - 2' )

54.0 - (C - 3' )

172.5 - (C - 1") C = 0)

57.7 - (C - 2")

10 54.5 - (C - 3")

22.6 - (C - 4")

29.4 - (C - 5")

207.1 - (C - 9; C = 0)

DIBr- IV __-NMR in CDCL ₃

(300 MHz in ppm; side chain and some important protons only)

Chemical Shift (ppm) Assignments

3.23 (d, IH)

4.76 (d, IH)

5.65 (d, IH)

4.62 (qt, IH)

1.98 (s, 3H)

1.28 (s, 3H)

2.45 (s, 3H)

1.72 (t, 2H)

3.72 ( , IH)

4.62 (s, IH)

6.79 (s, IH)

2.42 (m, IH) - (- CH ₂ -) (2H - 14a, b)

2.05 (m, IH) - ⁽ - CH ₂ ") (2H - 14a, b) 2.18 (s, 3H) - (-CH ₃ - ) (3H - C - 18)

4.38 (s, 2H) - (- CH ₂ " ) (2H - 20a, b)

DIBr- IV ¹³ C-NMR

(300 MHz in ppm; side chain and some important carbons only)

Chemical Shift (ppm) Assignments

5 169.2 72.1 54.1 172.5 57.8 10 54.3 22.6 29.4 207.1

Physico-chemical properties of dibromocephalo- 15 mannine/7-epi-cephalomannine diastereomers of this invention are summarized below in Table 4 :

TABLE 4 Physico-chemical Properties of Bromo-Analogues of Paclitaxel

* The IR spectra of DiBr-I-IV are superimposable.

** Solvent System A: Methanol-1, 2, -Dichloroethane- either (1:9) or (1:10) . Solvent System B: Hexane-Chloroform-Ethyl Acetate-

Methanol-(2:6:1.5:0.5)

*** Condition 1: Column: ES Industries FSP

(Pentafluorophenyl) 4.6 mm ID x 250 mm, 5 qm particle size, 60A pore size,- mobile phase - water - acetonitrile - methanol - (41:39:20) ; flow rate 0.50 ml/min; separation mode - isocratic; detector - Waters 990 Photodiode Array Detector; elution monitored at 227 nm,- injection volume - 20 ql .

Condition 2: Column: Phenomenex 4.6 mm ID x 250 mm, 5 urn particle size, 80A pore size; mobile phase - water - acetonitrile - methanol - (45:40:15) ; flow rate - 0.50 ml/min; separation mode - isocratic; detector - Waters 490 programmable multi avelength detector, elution monitored at 227 nm; injection volume - 80 ul total mixture.

EXAMPLE 5

In Vitro and In Vivo Studies Showing Antitumor Efficacty of A Mixture of Dibromo- Cephalomannine/Dibromo-7-epi-

Cephalomannine Diastereomers Which Correlate to Known Paclitaxel Antitumor Efficacy

As is known, paclitaxel and its derivative Taxotere ^* (Rhone-Poulenc Rhor) exhibit highly desirable antitumor efficacy against a number of tumors. These antineoplastic drugs act in a unique manner by preventing depolymerization of tubulin forming microtubules of the mitotic spindle which is essential for cell division, and thus cause cell division to cease along with tumor cell proliferation. The mechanism of action of paclitaxel, its pharmacology, etc. is described, for example, in Rowinsky et al . Taxol .- A Novel Invest igational Antimicrotuble Agent, J. Natl. Cancer Inst. , 82:1247 (1990) .

In accoradance with this invention, a mixture of novel dibromocephalomannine/dibromo-7-epi-cephalommanine diastereomers has been found to exhibit strong paclitaxel-like antitumor efficacy in vi tro and in vivo.

5.1 In Vi tro Studies (NCI)

The following in vi tro studies were conducted by the National Cancer Institute's Developmental Therapeutics Program, which demonstrate strong antitumor efficacy of the inventive dibromocephalomannine diastereomers which efficacy correlates closely to that of paclitaxel.

The Developmental Therapeutics Program provides as a service to the public an in vi tro anticancer drug discovery screen using a panel of sixty different human tumor cell lines over which candidate drugs are tested at defined ranges of concentrations. See Boyd et al . , Drug Development Research 34:91-109 (1995), the entirety of which is incorporated herein by reference. As discussed in Boyd et al . , the screen is designed and operated in such a manner that both relative and absolute sensitivities of each of the cell lines comprising the screen are reproducible to the degree that a characteristic profile ("fingerprint") of a respective cell lines' response to a drug candidate can be generated. Recent studies of the in vivo counterpart of the NCI in vi tro screen have indicated the in vi tro screen to be an effective selector of compounds with in vivo anticancer efficacy. See Grever et al . , Proc. Am. Assoc. Cancer Res. 35:369 (1994) . Operation and interpretation of the screen are discussed in detail in Boyd et al . , as well as in several other articles cited therein and thus need not be repeated here, except comparative results obtained from the screen between the novel 2"3"- dibromocephalomannine/dibromo-7-epi-cephalomannine diastereomic mixture represented as compound "XCLY-401759 analog" and that of the known antitumor compound, paclitaxel. Tn vitro antitumor efficacy of XCLY-401759 is shown in FIGs. 12 and 13, Testing Results and Mean Graphs, respectively.

In corresponding manner, in vi tro antitumor efficacy is shown in FIG. 14 by dose resonse represented by a mean graph of paclitaxel.

5.1.1 Discussion of Results (NCI)

In the NCI in vi tro anticancer drug screen the effect of an antitumor candidate, i.e. XCLY-401759 of the present invention, on a cell line, percentage growth (PG) , and calculated response parameters are discussed in detail in Boyd et al . , Data display and analysis stra tegies for the NCI - disease -oriented in vi tro anti tumor drug Screen, Cytotoxic Anticancer Drugs: Models and Concepts for Drug Discovery and Development, Kluwer Academic Publishers, Amsterdam, pp. 11-34 (1992) , and Monks et al . Feasibili ty of a high- flux anticancer drug screen utilizing a diverse panel of human tumor cell l ines in cul ture, J. Natl. Cancer Inst . 83:757-766 (1991), the entire disclosures of which are incorporated herein by reference. In general, in the screening data report, FIG. 12, and mean graphs, FIGs. 13 and 14, "GI _S0 " represents the 50% growth inhibition factor, "TGI" represents a total growth inhibition, or cytostatic level of effect, and "LC ₅₀ " represents a lethal concentration, or net cell killing or cytotoxicity parameter. Values accompanied by a "<" signify that the dosage level or real value is a value that is something less than the lowest tested concentration, and values accompanied by a ">" sign indicate that the effective dosage or real value is a level greater than the highest tested concentration. The mean graphs are obtained from GI _S0 , TGI and LC _S0 concentrations obtained for compounds tested against each cell line in the NCI in vi tro screen. A detailed discussion of mean graph construction is provided in Boyd et al. (1995) . In interpreting the mean graphs, in general a bar projecting to the right represents sensitivity of a particular cell line to an anticancer candidate in excess of the average sensitivity of all tested cell lines, while bars extending to the left represent cell lines which are less sensitive on average to the anticancer candidate. The bar scales are logarithmic, such that a bar which extends, for example, 2 or 3 units to the right of the vertical reference line in, say a GI ₅₀ mean graph, indicates that the anticancer candidate achieved a response parameter for a particular cell line at a

concentration one-hundredth to one-thousandth of the mean concentration required over all cell lines, thereby indicating that the particular tumor cell line is unusually sensitive to the tested candidate. Turning now to FIG. 13, XCLY-401759 shows a relatively high magnitude of effect in TGI, for example, on Leukemia cell line HL-60 (TB); Non-Small Cell Lung Cancer line NCI-H522; Colon Cancer cell lines COLO 205 and HT 29; CNS Cancer cell lines SF-539 and SNB-75; Ovarian Cancer Cell line OVCAR-3; Renal Cancer cell line RXF-393; and Breast Cancer cell lines MCF7, MDA-MB-231/ATCC, HS 578T, MDA-MB-435 and MDA- N.

In comparison with FIG. 14, analysis of paclitaxel, XCLY-401759 demonstrates an unusually high magnitude of response such as that of paclitaxel to Non-Small Cell Lung Cancer cell line NCI-H522 (<-8 v. <-10 for XCLY-401759 and paclitaxel, respectively) . Compare also the respectively high magnitude of response of both XCLY-401759 and paclitaxel on Colon Cancer Cell line COLO 205 (<-8 v. -7.97) ; on CNS cancer cell line SNB-75 (-7.30 v. -9.18) , and, for example, on Breast Cancer Cell line HS 5787 (-7.61 v. -9.91) .

The high magnitude of effect of XCLY-401759 on many cell lines is perhaps more pronounced in GI ₅₀ in which XCLY- 401759 demonstrates a high response level in many of the same cell lines as does paclitaxel, such as, for example, with various tested colon cancer cell lines, melanoma cell lines, ovarian cancer cell lines, and renal cancer cell lines, and thus falls within the footprint of paclitaxel-like antitumor activity thereby reproducibly demonstrating the high antitumor efficacy of the novel XCLY-401759 mixture.

The strong paclitaxel-like antitumor efficacy of XCLY-401759 is further shown in correlation data generated by the NCI, as summarized below in Table 5:

TABLE 5

NCI

COMPARE-CORR-GI50 XCLY-401759/LCONC-4.OOM (BV)

CORR.

NSC LCONC (MAX X) COEFF. (N) CHEM-NAME

*NSC-Test number

LCONC-Log of the highest concentration tested MAXX-Total number of tests COEFF. -Pearson correlation coefficient CORR. -

(N) -Total number of cell lines See Paull et al . , J. NCI (1989) 81:1088-1092

5.2 JN VITRO STUDIES (SOUTHERN RESEARCH INSTITUTE)

Additional in vi tro studies were performed by the Southern Research Institute, Birmingham, Alabama, an independent research group, of the biological anti-cellular activity of XCLY-401759 on four human tumor lines, MX-1 (breast carcinoma) , RXF-393 (renal cell carcinoma) , NCI- H522 (lung adenocarcinoma) and OVCAR-3 (ovarian carcinoma) . In these studies, the XCLY-401759 analog was shown to yield a range of activity comparable to paclitaxel. This testing was conducted using the aforementioned human tumor cell lines employing standard tissue culture techniques with semi-automated dye conversion assays. Selection of the human cell lines for testing was based at least in part on the following criteria: (1) histogenesis of clinical import, (2) adequate growth characteristics, and (3) the Institute's experience with particular cell lines. The materials, methods and results of this study follow.

5.2.1 Materials and Methods 5.2.1.1 Cell culture.

In the Southern Research Institute Study, human cell lines were propagated under sterile conditions in RPMI 1640 (Hyclone) with 10% fetal bovine serum (Sigma Chemical) , 2 mM L-glutamine, and sodium bicarbonate (complete medium) and incubated at 37°C in HEPA-filtered

Sterilcult C0 ₂ tissue culture incubators (Forma) with 5% C0 ₂ and 95% humidity. The cell lines were subcultured weekly to bi-weekly and used in experiments. All lines were screened for mycoplasma contamination using GeneProbe™ (Fischer) and positive cultures were cured of contaminants over three passages using constant treatment with BM- Cyclin™ antibiotic combination (Boehringer Mannheim) . Only lines confirmed as mycoplasma free were used in testing compounds for anticellular activity.

5.2.1.2 Anticellular activity experimental design.

For all experiments, cells were harvested and pelleted to remove the medium and then suspended in fresh complete medium. Samples were taken to determine cell density. The cell count was determined with a Coulter Model Z cell counter and viability was measured with propidium iodide staining followed by analysis on a Coulter EPICS Elite Flow cytometer. The cell samples were adjusted with complete medium to a density of 5 x 10 ³ cells/ml. Tissue culture cluster plates (96 well, cat No. 3595 Costar) were seeded with 100 ql cells (5 x 10 ³ ) and incubated as described.

On the day of treatment analog XCLY-401759 was dissolved in 100% ethanol, and then serially diluted in medium. The 0 dose control was mock treated with medium. The appropriate wells (columns of 8) were treated with 5 concentration levels (10 ^~4 , 10 ^"5 , 10 ^"6 , 10 ^~7 , and 10 ^"8 M) . The highest dose of initial vehicle (ethanol in media) was < _. 0.2% ethanol. A vehicle control was prepared at 0.2% to determine the effects of vehicle on the cell lines.

Paclitaxel supplied by XECHEM, Inc., New Brunswick, New Jersey, was dissolved in DMSO, serially diluted in medium and then added to the wells to achieve doses of 1 x 10 ^"8 and 1 x 10 ^"9 M. Each cluster plate contained a cell control (8 wells, mock-treated with complete medium), a medium control (7 wells with medium used to substract out signal generated by medium conditions) and an air blank (1 well, for calibrating the plate reader) . Once dosing was completed, the plates were stacked and wrapped in plastic film to reduce evaporation and incubated as described. Replicate sets of cluster plates had either 1 hour or 72 hour drug exposure. For the appropriate drug exposure, the plates were aseptically blotted on sterile towels and gently washed three times with medium. The samples were then fed with fresh medium, and the plates were wrapped in plastic wrap. The plates of both exposure sets were incubated to day 7 and then processed to analyze for anticellular activity using the sulforhodamine B (SRB) procedure.

5 . 2 . 1 . 3 Results

In the 1 hour exposure XCLY-401759 concentration dependent activity was demonstrated in all the tested cell lines. OVCAR-3 ovarian and NCI-H522 lung cell lines were the most sensitive to XCLY-401759. Paclitaxel activity was minimal at the two concentrations tested for MX-1, RXF 393 and OVACAR-3 tumor cell lines, with NCI-H522 showing sensitivity to paclitaxel. All cell lines showed increased sensitivity to both XCLY-401759 and paclitaxel when the exposure time was increased to 72 hours. MX-1 was relatively less sensitive than other lines to paclitaxel and XCLY-401759.

In summary, according to the Southern Research Institute's Study, XCLY-401759 yielded a range of anticellular activity comparable to paclitaxel in four human tumor cell lines of tested various neoplastic disease originans .

The results are summarized below in Tables 6 and 7.

TABLE 6 SOUTHERN RESEARCH INSTITUTE

TREATMENT DAY 1 POST PLATING-PLATES WASHED 1 HR AFTER Rx; PLATES READ ON DAY 7

% INHIBITION TREATMENT CELL LINE (CELLS PLATED 5.0E+03 CELLS/WELL)

AGENT (M) RXF 393 MX-1 OVCAR-3 NCI-H522

3.2 3.7 0.0 1.7 42.7

VEHICLE CONTROL 0.0 0.8 7.0 4.5

PACLITAXEL

TABLE 7

SOUTHERN RESEARCH INSTITUTE

TREATMENT DAY 1 POST PLATING-PLATES WASHED 72 HRS AFTER Rx; PLATES READ ON DAY 7

% INHIBITION TREATMENT CELL LINE (CELLS PLATED 5.OE+03 CELLS/WELL)

AGENT (M) RXF 393 MX-1 OVCAR-3 NCI-H522

XCLY-401759 29.4 45.4 76.5 75.4 98.3

VEHICLE CONTRO 2.3

PACLITAXEL 10.6 41.1

5.3 IN VIVO STUDIES

In vivo hollow fiber assays were performed by the NCI Developmental Therapeutics Program on the anti-cellular efficacy of the inventive XCLY-401759 analog on several neoplastic tumor cell lines.

This testing was performed by the Biological Testing Branch of the Developmental Therapeutics Program. In these assays, human tumor cells as indicated were cultivated in polyvinylidene fluoride (PVDF) hollow fibers, and a sample of each cell line implanted into each of two physiologic compartments (intraperitoneal and subcutaneous) in mice. Each test mouse received a total of six fibers (3 intraperitoneally and 3 subcutaneously) representing 3 distinct cancer cell lines.

Three mice were treated with potential antitumor compounds at each of 2 test doses by the intraperitoneal route using a QD x 4 treatment schedule. Vehicle controls consisted of 6 mice receiving the compound diluent only. The fiber cultures were collected on the day following the last day of treatment.

To determine antitumor efficacy, the viable cell mass was determined for each of the cell lines using a formazan dye (MTT) conversion assay. From this, the % T/C was calculated using the average optical density of the compound treated samples divided by the average optical density of the vehicle controls. The net increase in cell mass was determined for each sample .

The XCLY-401759 diastereomeric mixture/compound was tested against a minimum of 12 human cancer cell lines, amounting to a total of 4 experiments as each experiment contains 3 cell lines. The data are reported as %T/C for each of the 2 compound doses against each of the cell lines with separate values calculated for the intraperitoneal and subcutaneous samples. The results of this in vivo assay are summarized below in Tables 8-11.

TABLE 8

This experiment is within acceptable quality control parameteres and is considered valid.

TABLE 9 Capillary Hollow Fiber Assa for XCLY-401759

co x mm

33 c r m ro σ>

This experiment is within acceptable quality control parameters and is considered valid.

TABLE 10 Capillary Hollow Fiber Assay for XCLY-401759

NCI

Comments for HF594-0-HF

This experiment is within acceptable quality control parameters and is considered valid.

TABLE 11

x m m

H VEHICLES

3c Grp 3 -> 1 (Dose- 150.00) in Saline : Tween 80 (0.05%) (Unknown) Inj. Vol . :0.1 ml/lOgm body wt Grp 4 -> 1 (Dose- 100.00) in Saline : Tween 80 (0.05%) (Unknown) Inj. Vol. :0.1 ml/lOgm body wt

I 0)

Comments for HF595-0-HF

This experiment is within acceptable quality control parameters and is considered valid.

EXAMPLE 6

PREPARATION OF 2" ,3"-DICHLOROCEPH2\LOMANNINE DIASTEREOMERS AND BIOLOGICAL ACTIVITY STUDIES 6.1 Raw Materials

Batches of crude plant extracts from Taxus yunnanensis or from Taxus wallachiana containing approximately 15-40% cephalomannine, approximately 50-70% paclitaxel, and approximately 20-35% other taxane/non- taxane components were obtained from The People's Republic of China. Chlorine gas was obtained from Matheson Ltd. Silica gel used was ICN Silitech, 32-63 um, 60 A, ICN Biomedicals, Inc., Aurora, OH. All solvents used were either HPLC or ACS grade and were obtained from Spectrum Chemical Mfg. Corp. Purified water used was deionized in-house.

6.2 Chlorination of Crude Plant Extract in Oxidized Chloroform

6.2.1 Preparation of Oxidized Chloroform

Chlorine (3.12 g) was added dropwise to chloroform (4 1) in order to neutralize a stabilizer, amylene, present in the commercially available solvent. The solution was mixed vigorously and left standing at room temperature overnight, and then washed once with 1.5% sodium sulfite solution (1.0 1) , and twice with water (2x1.0 1) . Hydrogen peroxide solution (3%, 10 ml) was then added, mixed vigorously, and allowed to stand for 3-5 days. Chlorine content in the solvent was determined by volumetric analysis. Next, to the solvent sample (5 ml) was added 1.0 N HCl (10 ml) and water (50 ml. To this mixture was then added KI (2 g) , mixed well to dissolve, and the resulting dark brown solution titrated with O.l N sodium thiosulfate solution. As the color of solution turned light brown, 3-4 drops of starch indicator solution (0.5%, USP) were added. The dark blue

- purple solution was further titrated until the solution turned colorless. The volume of sodium thiosulfate solution used to arrive at the end point was noted and chlorine content calculated. The desired chlorine content was in the range of 0.01 - 0.1%. The solvent was dried with anhydrous sodium sulfate (100 g) and used for the following chlorination reaction.

6.2.2 Chlorination Crude plant extract (5.0 g, 28.8% cephaloman¬ nine, 62.2% paclitaxel) was dissolved in oxidized chloro¬ form (1 1) in a 31 flask cooled to 4"C using an ice bath. HPLC analysis of the mixture after 1 hour showed a paclitaxel to cephalomannine ratio of 8:1. The reaction mixture was then stirred at 15°C for 9 hrs. HPLC analysis of the reaction mixture at this point showed a paclitaxel to cephalomannine ratio of 19:1. The reaction mixture 5 ml sample washed with 5 ml deionized water had a pH of about 2.0. The mixture was then washed with 500 ml 1.0% aqueous sodium sulfite solution, and the pH of the aqueous layer was 7.5. This was followed by two washes with water (2x500 ml) . The pH values of first and second water washes were 7.0 and 6.5, respectively. The combined aqueous layer was re-extracted with 150 ml chloroform. The organic layers were combined, dried with anhydrous sodium sulfate (85 g) , and evaporated to dryness. The solid residue (5.85 g) was purified by chromatography. LCMS analysis of the chlorinated material indicated formation a diastereomer mixture of dichlorocephalomannine as the product of reaction along with paclitaxel present in the starting material .

6.3 Chromatographic Purification of Chlorinated Material

The chlorinated material (5.85 g) was chromatographically purified using a column (4.1 cm i.d., 62 cm long) packed with silica gel (300 g) by the slurry method using 10% acetone in 1, 2-dichloroethane. The sample was dissolved in 10% acetone in 1,2- dichloroethane. Following the first two 700 and 350 ml fractions, all subsequent fractions were limited to 50 ml each. The fractions were analyzed by TLC (TLC plates were developed with 20% acetone in 1, 2-dichloroethane, detected with 1% vanillin in 50/50 sulfuric acid- methanol) . Dichlorocephalomannines eluted in fractions 8-13 and yielded 1.6 g solids (-90%) following evaporation of solvents. This material was finally purified by semi-preparative HPLC.

6.4 Chlorination of Crude Plant Extract in 1,2-Dichloroethane

6.4.1 Preparation of Chlorine Solution in 1,2-Dichoroethane A solution of chlorine in 1,2-dichloroethane was prepared by slow bubbling of chlorine into 1,2- dichloroethane (1 1) precooled to 0 - 4°C using an ice bath. The bubbling was continued for several min. (approx. 10 min.) until the desired concentration of chlorine in 1, 2-dichloroethane was achieved. Samples of the solvent were withdrawn periodically and analyzed for dissolved chlorine content as follows: To the solvent sample (5 ml) in a 250 ml erlenmeyer flask were added 1.0 N HCl (10 ml) and water (50 ml) . To this mixture was added KI (2 g) , mixed well to dissolve, and the dark brown solution was titrated with O.l N sodium thiosulfate solution. As the color of solution turned light brown, 3-4 drops of starch indicator solution (0.5%, USP) were added. The dark blue - purple solution was further

titrated until the solution turned colorless. The volume of sodium thiosulfate solution used to arrive at the end point was noted and chlorine content was calculated. The desired chlorine content was in the range of 0.01-0.1%.

6.4.2 Chlorina ion Crude plant extract (5.0 g) dissolved in 1,2- dichloroethane (200 ml) was cooled to -4°C and added dropwise to the stirred solution of chlorine (0.06%) in 1, 2-dichloroethane (1250 ml) cooled to 4°C by using an ice bath. Following complete addition, the mixture was stirred at 4°C for 1 hr and a sample was analyzed by HPLC. HPLC analysis indicated that the cephalomannine peak was nearly completely eliminated. The mixture was washed with 1.0% sodium sulfite solution (1 1) and water (2 x 1 1) . The pH values of the aqueous layers were as follows: sodium sulfite wash, 7.5-8.0; first water wash, 6.0-6.5; second water wash, 5.5. The aqueous layers were extracted with 1, 2-dichloroethane (200 ml) . The organic layers were combined, dried with anhydrous sodium sulfate (50 g) and evaporated using a rotary evaporator at 40°C. The residual solids were dried in vacuum oven at 40°C for 2 hrs to yield 5.3 g chlorinated material. HPLC analysis of this material showed dichlorocephalomannine as the product of the reaction together with paclitaxel present in the starting crude plant extract.

6.5 Separation of Chlorinated Material From Paclitaxel by Crystalization

The chlorinated product mixture from 6.4.2 (5.30 g) was dissolved in acetone (50 ml) in a 250 ml Erlenmeyer flask. To this solution was added hexanes (65 ml) , mixed well, and let stand at room temperature until crystallization began to occur. The flask was then stored at 4°C for 60 hrs. The crystals were filtered, washed with cold 20% acetone in hexanes, and dried in

vacuum oven at 40°C for 3.5 hrs to yield 3.10 g paclitaxel (- 95%, crystals I) . The combined filtrate and washings were evaporated, and the residual solids dried in a vacuum oven at 40°C for 2 hrs to yield 1.96 g mother liquor material (mother liquor I) . The crystals I (3.10 g) were next dissolved in acetone (32 ml) . To this solution was added hexanes (40 ml) and the mixture stored at room temperature for 5 hrs and then at 4°C overnight. The crystals were filtered, washed with 20% acetone in hexanes, and dried in vacuum oven at 40°C for 3 hrs to yield 2.49g paclitaxel, 98.5% (crystals II) . The filtrate and washings were combined and evaporated. The residual solids were dried in a vacuum oven at 40°C for 2 hrs to yield 0.65 g mother liquor material (mother liquor II) . The crystals II (2.49 g) were again dissolved in warm acetone (25 ml) . To the solution was added hexanes (25 ml) and the mixture stored at room temperature for 5 hrs and then at 4°C overnight. The crystals were filtered, washed with 20% acetone in hexanes, and dried in a vacuum oven at 40°C to yield 2.01g paclitaxel (99.5%, crystals III) . The filtrate and washings were combined and evaporated. The residual solids were dried in a vacuum oven at 40 C for 2 hours to yield 0.47g mother liquor material (mother liquor III) . The mother liquors I, II, and III containing dichlorocephalomannines were then pooled and further separated by semi-preparative HPLC.

6.6 Final Purification of 2", 3" - Dichlorocephalomannine and

2 " ,3"-Dichloro-7-epi-cephalomannine Diastereomers

The final purification of dichlorocephaloman- nine and 7-epi-dichlorocephalomannine diastereomers from other impurities was accomplished by semi-preparative HPLC (Waters Deltaprep 3000) using a Waters Deltapak C18 column, 100 A 19mm x 30 cm with 45% acetonitrile in water

as the mobile phase at the flow rate of 15 ml/min. Peak elution was monitored using a UV detector set at 227 nm. Portions of 200 mg material dissolved in methanol (2 ml) were injected. Elution of dichlorocephalomannine diastereomer I peaked approximately at 86 min. and diastereomer II at 98 min. Likewise, the dichloro-7-epi- cephalomannine diastereomer III peaked at approximately 118 min and the corresponding diastereomer IV peaked at 124 min respectively. Fractions collected from repeated injections were pooled and evaporated at 40°C under reduced pressure to remove the organic solvent. The crystallized solids were filtered, washed with water, and dried in vacuum oven at 40°C to yield pure dichloro¬ cephalomannine and dichloro-7-epi-cephalomannine diastereomers. The dichlorocephalomannine diastereomer I isolated in this manner was associated with a contaminant and was repurified by collecting smaller fractions during peak elution following the described HPLC procedure.

The preparation, separation and structures of the obtained diastereo eric dichloro compounds,

(I) (2"R, 3"_?) -dichlorocephalomannine (DiCl-I) ;

(II) ( 2 " S, 3 " R) -dichlorocephalomannine (DiCl-II) ; (III) (2"i.,3"5) -dichloro-7-epi- cephalomannine (DiCl- III ) ; and ( IV) (2 " S, 3 "R) -dichloro- 7-epi - cephalomannine (DiCl - IV) , m VII .

(VII)

7 - epi - cephalomannine

CI ₂

(cα,)

(CHO.) (CH.Oι)

(C_H«C_.

Paclitaxel 2", 3" • dichloro • cephalomannine 2", 3" - dichloro - 7 • epi - cepbalomaniiiπe

Separation

Analogue 1: (2"R , 3"S) - dichlorocephalomannine Analogue 2: (2"5 , 3"_.) - dichlorocephalomannine

Analogue 3: (2"_. , 3"S) ■ dichlor o-7-epi-cephalomannine Analogue 4: (2"S , 3"R) - dichloro-7-epi-cephalomannine

Paclitaxel Analogues (Chlorinated)

Analytical characterization of these diastereomers follows.

FIG. 15 is a TLC separation of 2" , 3 " - dichlorocephalomannine and 2", 3 " -dichloro-7-epi- cephalomannine stereoisomers (DiCl-I-DiCl-IV) . A key to FIG.16 is set forth below in Table 12.

Plate: silica gel 60 F ₂₅₄ (Merck #5554) Solvent System: a) 10% CH ₃ OH in 1, 2-dichloroethane b) hexane/chloroform/EtOAc/CH ₃ OH 20:60:15:5

Reagent: a) UV light b) vanilin/H ₂ S0 ₄ in methanol

FIG. 16 is an HPLC chromatogram of a mixture of the dichlocephalo annine and dichloro-7-epi- cephalomannine diastereomers of this invention, with peaks identified below in Table 13.

Equipment and conditions employed in generating this chromatogram are as follows:

column: ES Industries, FSP (pentafluorophenyl) ; 4,6mm ID X 250 mm; 5 um; 60 A solvent system: water/acetonitrile/methanol, 41:39:20 flow rate: 0.50 ml/min. ; isocratic detector: waters 990 photodiode Array Detector, monitored at 227 nm injection volume: 20 ul

In FIG. 17 are shown superimposed UV spectra of the stereoisomers of this invention in CH ₃ OH. Spectra results are summarized below in Table 14.

TABLE 14

Peak No. Stereoisomer λmax. nm (e)

I DiCl-I 226.6 14,813

II DiCl-II 227.2 14,990

III DiCl-III 228.2 17,252 IV DiCl-IV 229.4 14,694

FIG. 15 shows superimposed IR spectra of the presently inventive stereoisomers, which are summarized below in Table 15.

TABLE 15 Band, cm ^"1 Functional Groups

3500,1105,1070 tert. and sec. OH 3420,1670,1580 -CONH-

3110,3060,1605 mono sub.aromatic rings

1505,770,710 2960,2915,2870 -CH ₃ -; -CH ₂ -; -CH-groups

1465,1370 (in aliphatic or cyclic compounds)

3020 , 1670 , 1310 double bonds 980

1730, 1270 aromatic esters 1715,1240 >=0 groups 1730,1180 acetates 855 oxetane rings

FIGs. 19-22 are proton spectra ('H-NMR) of DiCl-I,

DiCl-II, DiCl-III and DiCl-IV diastereomers in CDC1 ₃ (300

MH _Z ) , respectively, and FIG. 23 are ¹³ C-NMR (300 MHJ spectra of these diastereomers, which are all summarized below as follows :

DICL- I -N R in CDCL ₃

(300 MHz in ppm; side chain and some important protons only) Chemical Shift (ppm) Assignments

2.54 (m, IH) 1.92 (t, IH) 2.32 ( , 2H) 2.32 (m, 2H) 4.58 (q, IH)

1.55 (d, 3H)

CL

1.70 (s, 3H) C - CH ₃ - C ₅ ")

0 CL

DICL-I ¹³ C-NMR in CDCL ₃

(300 MHz in ppm; side chain and some important carbons only)

Chemical Shift (ppm) Assignments

170.2 - (C - 1' ; C = O)

73.1 - (C - 2' )

55.0 -(C - 3')

172.0 - (C - 1' ) C = O) 70.8 - (C - 2")

58.7 - (C - 3")

21.8 - (C - 4" ) 27.5 -(C - 5") 203.6 - (C - 9; C

DICL-II 4.-NMR in CDCL ₃

(300 MHz in ppm; side chain and some important protons only)

Chemical Shift (ppm) Assignments

2.56 (m, IH)

1.94 (t, IH)

2.34 (m, 2H)

2.34 (m, 2H)

4.58 (q, IH)

1.55 (d, 3H)

1.70 (s, 3H)

DICL-II"C-NMR

(300 MHz in ppm; side chain and some important carbons only)

Chemical Shift (ppm) Assignments

170.2 - (C - 1' ; C = O) 72.6 - (C - 2')

55.0 - (C - 3' ) 172.6 - (C - 1") C = 0)

70.6 - (C - 2")

58.7 - (C - 3")

21.8 - (C - 4")

27.7 - (C - 5") 203.5 - (C - 9; C = O)

DICL-III ^X H-NMR in CDCL ₃

(300 MHz in ppm; side chain and some important protons only) Chemical Shift (ppm) Assignments

2.35 (m, 2H)

2.35 (m, 2H)

2.54 (m, IH)

2.35 (m, IH) 4.50 (qt, IH)

1.52 (d, 3H)

1.28 (s, 3H)

DICL-III ¹³ C-NMR

(300 MHz in ppm; side chain and some important carbons only)

Chemical Shift (ppm) Assignments

170.2

73.0

54.8

172.2

62.7

55.3

21.6

29.3

203.5 - (C - 9; C = 0)

DICL- IV ^* Η-NMR in CDCL ₃

(300 MHz in ppm; side chain and some important protons only)

Chemical Shift (ppm)

2.35 (m, 2H)

2.35 ( , 2H)

2.54 (m, IH)

2.35 ( , IH)

4.50 (q, IH)

1.52 (d, 3H)

1.28 (s, 3H) Gt")

DICL- IV ^X C-NMR

(300 MHz in ppm; side chain and some important carbons only)

Chemical Shift (ppm) Assignments

170.2 - (C - 1' ; C = O) 72.9 - (C - 2' )

53.9 - (C - 3' )

172.2 - (C - 1") C = 0)

62.5 - (C - 2")

55.0 - (C - 3") 21.7 - (C - 4")

29.3 - (C - 5")

203.5 - (C - 9; C = O)

FIG. 24 is an EI-MS spectrum of the DiCl-IV diastereomer which is the same fragmentation pattern for diastereomers DiCl-I, DiCl-II and DiCl-III, and FIG. 25 is an MS-FAB ^* spectrum for these diastereomers, all of which are summarized below.

DiCl-I, DiCl-II, DiCl-III and DiCl-IV EI-MS; [M ^* ] =902 568 [T] ^* ; 550 [T - H ₂ 0] ^* ;

508 [T-AcOH] +; 490 [T-AcOH-H ₂ 0] \- (m/z, the main 480; 448 [T-2AcOH] + or [T-B ₂ 0Hj ^* ; 386 fragments) [T-AcOH-B ₂ OH] ^* 326 [T-B ₂ OH-

2ACOH] ^* ; 308 [T-326-H _a O] ^* ; 264 [832-T] + ;

246 [264-H ₂ 0] ^* ; 188; 148; 122 [B.OH] ^* ; 105

[B_] ^* ; 91[C ₇ H ₇ ] ⁺ ; 83 [C ₄ H ₇ C=0] ; 77[C ₆ H ₅ ]\-

57; 55; 43

DiCl-I, DiCl-II, DiCl-III and DiCl-IV (m/z, the main fragments) ,

940 ( [M+K] ^* ) ; 924 ( [M+Na] ^* ) ; 902 ( [M+ __] ⁺ ) ; 842 { [M-60 ^* ] ) ; 832 ( [cephal] ) ; 824 ( [M-60-18] ^* ) ;

569 ( [T] ^* ) ; 551 ( [T-18] ^* ) ; 527 ( [T-43] ^* ) ;

509 ( [T-60] ^* ) ; 491 ( [T-60-18] ^* ) ; 449/448 ( [T-122] ^* ) ; 405 ( [S-18] ) ;

387 ( [T-60 -122] ^* ) ; 327 ( [387-60] ^* ) ; 309( [327-18] ^* ) ; 264 ( [832 -T] ^* ) ;

246 ( [264-18] ^* ) ; 218 ( [264-46] ^* ) ; 105 ( [C ₆ H ₅ CO] ^* ) ; 91 ( [C ₇ H ₇ ] ^* ) ; 77 ( [C ₆ H ₅ ] ^* ) ;

Physico-chemical properties of the dichlorocephalomannine/dichloro-7-epi-cephalomannine diastereomers of this invention are summarized below in Table 16.

TABLE 16 Physico-chemical Properties of Chloro-Analocmes of

Paclitaxel

*The IR spectra of DiCl-I-IV are superimposable

**Solvent System A: Methanol-1, 2, -Dichloroethane-

(1:10) . Solvent System B: Hexane-Chloroform-Ethylacetate- Methanol- (2:6:1.5:0.5)

*** Condition 1: Column: ES Industries FSP

(Pentafluorophenyl) 4.6 mm ID x 250 mm, 5 um particle size, 60A pore size; mobile phase - water - acetonitrile - methanol - (41:39:20) ; flow rate 0.50 ml/min; separation mode - isocratic; detector - Waters 990 Photodiode Array Detector; elution monitored at 227 nm; injection volume - 20 ul .

Condition 2: Column: Phenomenex 4.6 mm ID x 250 mm, 5 qm particle size, 80A pore size; mobile phase - water - acetonitrile - methanol - (45:40:15) ; flow rate - 0.50 ml/min; separation mode - isocratic; detector - Waters 490 programmable multiwavelength detector, elution monitored at 227 nm; injection volume - 80 ul total mixture.

EXAMPLE 7 In Vitro NCI Studies Showing

Antitumor Efficacy of (2*'R,3"£?) - and (2" S.3 ⁿ R) -Dichloro¬ cephalomannine Diastereomers. In this NCI study, isolated and purified

(2"R,3"S) and (2"S,3"R) diastereomers of dichloro¬ cephalomannine are shown to exhibit strong paclitaxel- like antitumor efficacy in vi tro in the NCI's sixty human tumor cell line screen.

7.1 Discussion of Results The results of the NCI in vi tro study are summarized below in FIGs. 26, 27, 28 and 29. Mean graphs, FIGs. 27 and 29 for diastereomers ( 2 "R, 3 " S) and ( 2 "S, 3 "R) , respectively, show strong antitumor efficacy for both of these compounds.