Trimethyl Siloxane Steroid Intermediates

Steroids characterized by presence of an 11.beta.-hydroxy substituent are widely employed in the pharmaceutical arts, including for example dexamethasone, betamethasone, flucortilone, triamcinolone and many other potent anti-inflammatories. However, once the 11.beta.-hydroxy substituent is present in a steroid molecule, its presence complicates transformation of that steroid into many of the desired pharmaceutically active compounds. In fact, many synthesis procedures which are high yielding conversion sequences for pregnane series steroids cannot be employed on a large scale with cortico-steroids. This situation has been one factor which heretofore has caused workers in the art to prefer diosgenin rather than hecogenin as source material for many cortico-steroids. However, an ever growing unavailability of diosgenin is forcing renewed consideration of hecogenin (a by-product from sisal cultivation) as source material for preparation of potent cortico-steroids.

One instance of such attention can be found in U.S. Pat. No. 3,876,633, wherein preparation of many corticoids from hecogenin are described, the sequences disclosed there proceeding by way of 3.beta.-acyloxy-5.alpha.-pregna-9(11), 16-dien-20-one.

It has now been discovered that efficient conversion sequences are possible even when the 11.beta.-hydroxy substituent (also present in the final cortico-steroid products) is on the molecule at the earliest stages of the conversion sequence.

Briefly stated, the present invention involves conduct of synthesis procedures on 11.beta.--OH steroids while the 11 position is blocked off by a trimethyl siloxy group, i.e. 11.beta.--O--Si(CH.sub.3).sub.3. It has been discovered that the trimethyl siloxy group can be added and removed, selectively and quantitatively. It has also been discovered that presence of this blocking group is advantageous for reaction sequences that transform the steroid molecule and place therein one or more of the following groups: 6.alpha.F; 16,17--OH; 16.alpha.--CH.sub.3, .DELTA..sup.1,4.

A convenient starting compound for practice of the present invention is hydrocortisone itself or its 21-acetate, and in this connection it may be noted that British Pat. No. 1,227,992 discloses the formation of 11.beta.--O--Si(CH.sub.3).sub.3 hydrocortisone acetate by reacting hydrocortisone acetate with trimethyl chlorosilane in pyridine solution. The British patent suggests that the 17.alpha.OH position can then be esterified, after which the 11.beta. hydroxy is reconstituted. However, the British patentees do not appear to have appreciated that use of the trimethyl siloxy blocking group might be advantageous in multi-step synthesis sequences leading to more active compounds than hydrocortisone esters, or to synthesis sequences for active compounds which do not proceed through hydrocortisone (as an intermediate).

The present invention relates to a group of reaction sequences made feasible by forming an 11.beta.--0--Si(CH.sub.3).sub.3 derivative of certain intermediates. Individual conversion reactions known to be effective can then be employed on the 11.beta.--0--Si(CH.sub.3).sub.3 compounds including notably:

i. converting a 3-keto-4-ene steroid to the corresponding 3-acetoxy-4,6-diene steroid;

ii. reacting a 3-acetoxy-4,6-diene steroid with perchloryl fluoride to form the 3-keto-4-ene-6-F steroid;

iii. formation of a 16-ene steroid through elimination of a 17.alpha.-acetoxy substituent;

iv. permanganate oxidation of a 16-ene steroid to form the corresponding 16.alpha.OH, 17.alpha.OH steroid;

v. DDQ dehydrogenation to form a 1,4-diene;

vi. conversion of a 16-ene steroid to the corresponding 16.alpha. methyl steroid.

These various reactions are employed in advantageous sequences which lead to converting an 11 .beta.-hydroxy steroid into one with one or more of the following substituents thereon: 6.alpha.F; 16.alpha.OH,17.alpha.OH; 16.alpha.CH.sub.3 ; .DELTA..sup.1,4.

In all fairness, it is noted that not each reaction is superior (on a step-by-step comparison to a comparable reaction in the prior art sequences) but the sequences herein described are considered advantageous as a whole. In particular the reaction sequences of this invention permit facile conversion of hecogenin derived intermediates into potent corticoids having one or more the above listed substituents therein.

An important aspect of this invention is selective hydrolysis of trimethyl siloxanes through reaction with 40-60% aqueous HF.

For understanding of the practice of this invention reference is made to the drawing illustration of reaction sequences wherein:

FIG. 1 illustrates the conversion of hydrocortisone to fluprednisolone;

FIG. 2 illustrates a conversion to fluocortilone acetate and the fluandrenolone;

FIG. 3 illustrates a conversion to betamethasone or dexamethasone; and

FIG. 4 illustrates how different source compounds may be converted into triamcinolone and its derivatives.

The 11.beta.-trimethyl siloxanes employed for practice of this invention are formed by reacting the corresponding 11.beta.--OH steroid with trimethyl chlorosilane in pyridine solution. Formation of the 11-.beta. silyl ether of hydrocortisone (compound 1, FIG. 1) is known to the art (c.f. British Pat. 1,227,992). The British patent discloses that the silylation does not change the 17.alpha.OH group, and that (within limits ) the 11.beta.--0--Si(Me).sub.3 group is inert to the strong acetylating conditions required to acetylate the 17.alpha.OH.

As indicated above, it has now been found that the 11.beta. trimethyl siloxy group is also stable against the conditions employed for enolizing steroids, and for fluorinating. FIG. 1 illustrates conversion of hydrocortisone to fluprednisolone.

Referring now to FIG. 1, it may be seen that subjecting the silyl ether of hydrocortisone 21-acetate (compound 1) to strong enolizong conditions forms, thereby, the enol acetate (compound 2). Thereafter reacting the enol acetate with perchloryl fluoride forms the 6-fluoro-.DELTA..sup.4 -3-keto-11.beta.-trimethyl-siloxy (compound 3). Unexpectedly, the ratio of the 6.alpha./6.beta. fluoro in the product was found to be about 1:1. When the same enolizing and fluorinating reaction sequence is carried out on 11-desoxy steroids essentially 100% of the 6.beta.-fluoro steroid is formed. Normally a separate acid catalyzed isomerization is then employed to convert the steroid to the 6.alpha.F configuration.

According to preferred practice of this invention, the blocking silyl group is thereafter removed by acid catalyzed methanolysis, regenerating the 11.beta.-hydroxyl group and concommitantly isomerizing the 50--50 epimer mixture entirely to the 6.alpha.F configuration of 6.alpha.Fluoro hydrocortisone 17,21-diacetate (compound 4).

The 6.alpha.fluoro hydrocortisone 17,21-diacetate is then dehydrogenated with DDQ (2,3-dichloro-4,5-dicyanobenzoquinone) to 6.alpha. fluoroprednisolone, 17.alpha.,21-diacetate (compound 5) which in turn is saponified to fluprednisolone (compound 6).

Referring now to FIG. 2, whereon is illustrated the reaction sequence preparing fluocortilone, the starting compound 11 is prepared by silylating 6.alpha.-fluoro hydrocortisone 17.alpha.,21-diacetate (compound 4). Thereafter the 17.alpha.OH is eliminated (by reaction with potassium acetate in dimethylformamide) to produce the .DELTA..sup.16 -steroid (compound 12) in excellent yield. Dehydrogenation with DDQ forms the triene (compound 13). The trimethyl siloxy blocking group is inert to the elimination reaction conditions and to the DDQ dehydrogenation conditions. In addition the silyl blocking group has been found to be inert to the reaction with lithium-dimethyl-copper employed to introduce the 16.alpha. methyl group (compound 14). This alkylation reaction is known for other 16-ene-20-keto steroids (R. Philippson, B. Acksteiner, P.E. Schulze, Offenlegungschrift 2,230,287, filed June 16, 1972). When the silyl blocking group is hydrolyzed from compound 14, fluocortilone (compound 15) is obtained.

FIG. 2 also illstrates conversion of the 16-ene (compound 12) to fluorandrenolone. Subjecting the 16-ene to potassium permanganate in a tube reactor resulted in the 16.alpha.,17.alpha. diol (compound 16) with the 11.beta. silyl blocking group remaining intact. Acid hydrolysis of the blocking group and saponification of the 21 ester resulted in the tetrol compound 17 (fluorandrenolone), which in turn may be made into Cordran by known esterification procedures.

The reaction sequences illustrated in FIGS. 1 and 2 involve conversions which incorporate into an 11.beta. hydroxy steroid one or more of the following groups: 6.alpha.F, 16.alpha.,17.alpha.hydroxy, 16.alpha. methyl, .DELTA..sup.1,4. Other reaction sequences hereinafter illustrated involve incorporation of 16.alpha.,17.alpha. isopropylidenedioxy and 16.beta. methyl groups.

The inertness of the trimethyl silyl ether blocking group to the individual reactions involved is desirable, since mostly, reactions of choice are employed. Heretofore presence of an 11.beta.-hydroxy group on the intermediate has precluded (commercial scale) utilization of what are believed to be the best reaction sequences for introducing the above listed functional groups into the readily available 11.beta.--OH steroids such as hydrocortisone and hecogenin derived intermediates.

Thus with specific regard to introduction of the 6 F group on steroid molecules (illstrated in FIG. 1), the best reaction sequence is believed start with a .DELTA..sup.4 -3 ketone, and:

i. form the .DELTA..sup.3,5 enol acetate;

ii. react with perchloryl fluoride to form the 6F derivative;

iii. isomerize to the 6.alpha. F configuration

The above direct and efficient pathway is not feasible with steroids containing an 11.beta.-hydroxy group. The 11.beta.--OH is coacetylated during the conditions used for forming the .DELTA..sup.3,5 enol acetate. As a rule, a 11.beta.-acetoxy group cannot be hydrolyzed successfully, for which reason many important products containing the 6.alpha.-fluoro-11.alpha.-hydroxy groupings (fluprednisolone, fluorandrenolide, flucortolone) have not been obtained from hydrocortisone by the direct efficient method referred to above. Protection of 11.beta.-hydroxyl groups by the trimethylsilyl ether derivative now makes many important 6.alpha.F steroid products available from hydrocortisone via a simple and efficient pathway.

In carrying out the conversions illustrated in FIGS. 1 and 2 certain process advantages were observed:

i. Formation of the enol acetate and 17.alpha.-acetate.

While it is known that the 17.alpha.-hydroxy group of the 11.beta. trimethyl siloxane blocked hydrocortisone 21-acetate can be acetylated (the BP 1,227,992 referred to above) without undesirable reaction about the 11.beta.-trimethylsiloxy group, it was, perhaps, surprising that the much stronger reaction conditions of isopropenyl acetate and presence of 10 times as much paratoluenesulfonic acid catalyst would give the enol acetate trimethylsilyl ether (compound 2) in close to quantitative yield.

ii. Fluorination of the enol acetate with perchlorylfluoride.

Mention has already been made that the immediate reaction product of the enol acetate with FC10.sub.3 (compound 3) exhibited the 6.alpha.F/6.beta. epimeric ratio of about 1:1 according to TLC analysis. The like reaction on steroids with either no 11.beta. substituent, or with one of small bulk results in up to 100% 6.beta.F. The .alpha.F preference found to exist may be due to shielding of the .beta.-side of the molecule from attack. Whatever the reason for the more favorable 6.alpha./6.beta.F ratio, the desirable consequence is that a cumbersome isomerization reaction step employed to convert the 6.xi.-fluoro epimeric mixture of the steroid to the desired 6.alpha.-epimer can be avoided (as a separate step). Conversion of the epimeric mixture resulting from the fluorination (compound 3) to 6.alpha.F-epimer takes place concomitant with the acid catalyzed methanolysis employed to regenerate the 11.beta.-hydroxy group and produce compound 4. Such favorable kinetics is of value to a large scale process sequence. The fluoro atom is small, and low energy differential exists between steroid 6.alpha. and 6.beta. fluoro epimers. For that reason the usual isomerization reaction carried out is slow, stops at some sort of equilibrium, and mother liquor material rich in 6.alpha.-fluoro is recycled back to the isomerization reaction. Starting with a 50-50 epimeric mixture constitutes a considerable advantage, particularly when isomerization is made to occur during hydrolytic regeneration of the 11.beta.--OH group.

iii. Dehydrogenation reaction with DDQ.

When a 2,3-dichloro-4,5-dicyanobenzoquinone dehydrogenation is run on an 11.beta.-hydroxy steroid under the usual conditions (dioxane, sometimes with added acids) employed to favor production of the 1,4-dienoic-3-keto structure, some coreaction about the 11.beta.-hydroxy occurs. The nature of this side reaction is not known, but is thought to be a dehydration at C-11 and C-12 leading to an enol which is rearranged to an 11-keto artifact.

In particular, inferior performance occurs if a negative group such as F is attached to C-9. For example 11.beta.-hydroxy-9.alpha.-fluoro is not a good substrate for introduction of the 1,4-dien-3-keto system. However, when the 11-hydroxy group has been converted to the trimethylsilyl ether group, the side reaction of DDQ dehydration (leading to formation of an 11-keto artifact) is completely inhibited. A substantial improvement occurs in the reaction yield for 9.alpha.-fluoro derivatives which is important to synthesis of such products as betamethasone and dexamethasone. It is not known whether the improvement is related to an inability to form the 11-keto, or to the generally more favorable ionics about the C-9 and C-11 positions. It may be noted that a .DELTA..sup.4 -3-ketone converts to the 1,4-dien with DDQ in better yield than does a 5.alpha.-pregnane-3-ketone.

iv. Introduction of the 16-double bond. The trimethylsiloxy blocking group was inert to the usual conditions of eliminating the 17.alpha. -acetoxy group to form the 16-double bond (L. Salce, G. Hazen, E. Schoenewaldt, J. Org. Chem., 35, 1681 (1970). These conditions, i.e. potassium acetate in dimethylformamide at 110.degree.-120.degree. C., gave excellent conversion of compound 11 to 12, probably because of unique solubility properties of the silyl ether intermediate in the solvent combination used for trituration, permitting facile isolation of a pure product without loss of material in the mother liquors.

v. Hydroxylation of the 16-double bond.

Introduction of the 16.alpha., and 17.alpha. dihydroxy groups, e.g. conversion of compound 12 to 16, is a very important aspect of the synthesis of certain corticoids (e.g. Triamcinolone, Fluocinolone). These groups form acetals with acetone very easily. The so called acetonides are themselves important corticoids. For carrying out the hydroxylation reaction of the 16-double bond, potassium permanganate is the reagent of choice (to react with the .DELTA..sup.16 -steroid dissolved in acetone). The trimethylsiloxy blocking group was found to be inert to this reagent even in the presence of added acids such as formic acid so long as a short time contact period is provided for the steroid solution with the acid containing permanganate solution, e.g. a tube reactor, 3 second contact time, with the reaction product continuously quenched by running into a large reservoir of aqueous sodium sulfite or bisulfite, which serves to keep the pH about neutral; manganese dioxide precipitates out.

vi. Formation of the acetonide.

Although the reaction to form the acetonide derivative of the 16.alpha., 17.alpha. -diol is an acid catalyzed reaction (perchloric acid), it may be carried out before removing the silyl blocking group, since no co-reaction takes place about the silyl blocking group. Acetone is the reaction milieu and there is not enough water around to bring about the hydrolysis of the silyl blocking group. The solubility and crystallizing ability of the 11-trimethylsiloxy-16,17-acetonide products are favorable properties resulting in maximal yield and purity of the product.

vii. Methylation with lithium-dimethyl-copper.

It was found possible to apply the methylation procedure described in Deutsches Patentamt Offenlegungsschrift 2,230,287 (by R. Philippson and coworkers, June 16, 1972) to the 16-ene, 11.beta.-trimethylsiloxy compound 13 without concomitant loss of the blocking group to produce the 16.alpha. methyl steroid of compound 14. As an advantage over the process described by the German workers, it was found not necessary to reacetylate during the workup, indicating a shielding of the .beta.-side of the steroid molecule by the bulky trimethylsiloxy group. The 16.alpha.methyl,11.beta.-trimethylsiloxy compound 14 exhibited superior crystallization and solubility characteristics, as did all of the 11.beta.-trimethylsiloxy derivatives.

viii. Regeneration of the 11.beta.-hydroxy group.

It is essential for practice of this invention to appreciate limits to the stability of the trimethylsiloxy ethers during the course of the intended reactions (so as to avoid any undesired cleavage of the silyl ether resulting in equally undesired coreaction of the 11-hydroxy functions). It is of course important for practice of this invention to possess selective methods for regeneration of the 11.beta.-hydroxy function, after the intended reactions have been successfully carried out. Without a dependable selective procedure capable of quantitatively cleaving the siloxy group with no co-reaction at other sensitive points on the steroid molecule, the 11.beta.-trimethylsiloxy ethers would be of limited use as reaction intermediates. Unfortunately, previously known cleavage reactions adapted to remove the siloxy group were unsatisfactory in many instances. Fortunately, the trimethylsiloxy blocking group was found to be sufficiently stable to carry out all the reactions discussed above. However, in some instances the usual reaction conditions were modified. For instance (Example 1 below) acetylation of the hindered 17.alpha.-hydroxy function has been achieved concommitantly with enolization of the .DELTA..sup.4 -3-keto moiety to the .DELTA..sup.4,6 diene without any transacetylation at C-11 by using the isopropenyl acetate/paratoluenesulfonic acid reagent, but a conventional workup after 17.alpha.-acetylation with the usual acetic anhydride/paratoluenesulfonic acid reagent, (Example 21 below) could not be employed. Resort was had to a dimethyl formamide, sodium acetate buffer, to avoid cleavage of the silyl blocking group.

As a whole it has been found that the rate of formation and the stability of the silyl ether blocking group was dependent upon the substituent in the 9.alpha.-position and on whether ring A was a 1,4-diene-3-one, a 4-en-3-one or a saturated 5.alpha.H-3-one. Conditions which were suitable for acetylation of the 17.alpha.-hydroxy group in the 9.alpha.-fluoro series, caused acetolysis at the trimethylsiloxy blocking group in the prednisolone series (1,4-diene-3-one).

An important additional aspect of this invention is the discovery of a novel cleavage method, which briefly is to expose the 11.beta.-trimethyl-siloxy ether to 40-60%, aqueous hydrofluoric acid, at room temperature, and preferably about 47-50%.

This cleavage reaction cannot be explained. It is heterogenous (the steroid reactant is in suspension). The reaction is fast and requires small particle size. The reaction may be controlled by TLC in a small pilot run of the same material and acid, and letting the reaction run until just under 100%, e.g. about 95%, of the starting material has reacted. If the cleavage reaction is run longer, other reaction products are formed. The total material is filtered off from a neutralized aqueous suspension, and for some uses of the steroid hydrolysis product no recrystallization is necessary. (As may be seen from Examples 10, 12, 16, 19 and 27 below there are differences in reaction time for complete cleavage). The cleavage reaction is very specific. One demonstration of the specificity of this cleavage reaction is that hydrolysis of compound 3 with HF, cleaved off the silyl blocking group but the 6.alpha.F/6.beta.F ratio in the steroid product remained the same. A different cleavage reaction, i.e. HCl/chlf/Ethanol is required for combined cleavage and isomerization. The aqueous HF method of cleavage is however exceedingly satisfactory where specificity is desirable. Thus, very easily hydrolyzable 21-acetoxy groups, as well as 17,21-diacetoxy groups of 20-ketones are left completely intact by aqueous HF cleavage. Similarly 16,17-acetonide groups are left untouched when the cleavage contact time is kept to less than 100% completion, i.e. neutralize when 95+% of the starting material has been shown reacted by TLC test on the pilot run.

Reference can now be made to FIGS. 3 and 4 for further illustrations of practice of this invention. These figures show how the already described reactions have been combined into reaction sequences well adapted to synthesis of widely used corticoids, and moreover, illustrate how hecogenin derived intermediates may be employed as starting materials, such intermediates generally being characterized by presence of saturated rings A and B connected trans (i.e. 5.alpha.H).

As is shown in FIG. 3, compound 22 (tetrahydrobetamethasone) is silylated to form the 11.beta. trimethyl siloxane, compound 23. Then dehydrogenation with 2,3-dichloro-4,5-dicyanobenzoquinone (DDQ) in dioxane solution provides the 1,4-dien, compound 24. Thereafter aqueous HF hydrolytic cleavage forms betamethasone, compound 25.

In a comparable sequence, also shown in FIG. 3, the 16 methyl fluorohydrin, compound 32 is transformed into the silyl ether, compound 33, then DDQ dehydrogenated to the .DELTA..sup.1,4 -dien, compound 34, and thereafter hydrolyzed to dexamethasone, compound 35.

Still another area of practice of this invention is production of triamcinolone from either hecogenin, or from prednisolone. FIG. 4 illustrates the reaction sequences involved.

One starting compound 41 is 11.beta.,17.alpha.-dihydroxy-21-acetoxy-9.alpha.-fluoro-5.alpha.-pregnane- 3,20-dione, a compound reported by J. Elks, G. H. Phillipps, W. F. Wall; J. Chem. Soc. 1958, 4001. This compound is converted into the trimethyl silyl ether, compound 42, then the 17.alpha.hydroxy group is acetylated to form compound 43 using conditions so that no enol acetate formation occurs around the 3-keto group. Thereafter the 17.alpha.,21-diacetate is subjected to a potassium acetate in dimethyl formamide elimination reaction to introduce the 16-double bond (efficiently), forming compound 44. The 1,4-diene system was obtained by dehydrogenation with DDQ, forming compound 45.

Compound 45 is converted to the 16.alpha.,17.alpha. diol, compound 46 by short time, e.g. three seconds, tube reactor contact with potassium permanganate. In turn compound 46 can be hydrolyzed to either of compounds 47 (i.e. 21-acetoxy or 21-hydroxy) or reacted with acetone to arrive at the acetonide of compound 48, which in turn may be hydrolyzed to form either the 21-acetoxy or 21-hydroxy of compounds 49. The two compounds 47 can be reacted with acetone to form appropriately the acetonide of compound 49a or 49b.

Alternatively compounds 49 can be arrived at from prednisolone by first converting prenisolone, 21-acetate to the 11.beta., 9.alpha. fluorohydrin thereof, compound 51, then silylating the 11.beta.--OH to form compound 52, followed by acetylating the 17.alpha. OH steroid to form the 17.alpha. acetoxy compound 53, followed by elimination of the 17.alpha. acetoxy group (potassium acetate in dimethyl formamide) to form the 16-ene compound 45, then proceeding to compounds 47 or 49 as already described.

The same sequence of reactions (i.e. of compounds 51 to 45) can be employed directly on prednisolone, 21-acetate to produce compound 55 (FIG. 4) and the compound 45-49 sequence already described then employed to form prednacinolide or its 21-acetate.

For further understanding of the present invention, the following specific examples thereof are presented.

A mixture of 22 g of 11.beta.-trimethyl-siloxy-17.alpha.-hydroxy-21-acetoxy-pregn-4-ene-3,20-di one, compound 1 in FIG. 1 (prepared according to J. Lens, A. F. Marx, BP 1,227,992 of Apr. 15, 1971, Example III), 2.2g of p-toluenesulfonic acid monohydrate, and 88 ml of isopropenyl acetate was heated under reflux for a period of 1.5 hr, and cooled. Then 3 g of anhydrous sodium acetate was added and the reaction mixture diluted with 140 ml of methylisobutyl ketone. The supernatant was decanted from the salt crystals and the resulting solution was washed twice, each time with 200 ml of water, then once with 200 ml of saturated aqueous sodium chloride solution, dried over anhydrous sodium sulphate, and evaporated under reduced pressure to a semi-crystalline residue. This residue can be dried to constant weight at 40.degree. C./20 mm of Hg then employed for starting material for later conversions. From the residue pure 11.beta.-trimethylsiloxy-3,17.alpha.,21-triacetoxy-pregna-3,5-diene-20-one (compound 2) can be prepared in the following manner:

The residue was dissolved in the minimum amount of benzene and charged on a column of silica gel. The product was eluted with benzene/ethylacetate (19:1), and gave a crystalline residue, which upon trituration with methanol containing a trace of pyridine gave the pure enol acetate with melting point about 148.degree. -160.degree. C.

IR: 1760, 1740, 1240, 1070, 845 cm.sup.-.sup.1

To a solution of 26 g of the enol acetate residue described in Example 1 (11.beta.-trimethyl-siloxy-3,17.alpha.,21-triacetoxy pregna-3,5-diene-20-one) in 400 ml acetone was added a solution of 10 g of anhydrous potassium acetate in 240 ml of absolute ethanol. The resulting solution was cooled to -30.degree. treated with a stream of perchloryl fluoride for 1 hr, maintaining the temperature within the range -30.degree. to -20.degree.. The reaction mixture was maintained within this temperature range for 60 hours, and then precipitated by pouring into 8 liters of iced water containing 20 g sodium sulphite. After 1 hr. standing the fine precipitate was collected, washed with water and dried under vacuum to constant weight (22g). A thin layer chromatographic study (benzene-ether 7:3) showed the presence of two products in about equal amounts, which were identified as the 6.alpha.- and 6.beta.- epimeric pair of compound 3 (i.e. 11.beta.-trimethyl-siloxy-17.alpha.,21-diacetoxy-6.alpha.-fluoro-pregn- 4-ene-3,20-dione).

24 g of the 6.alpha..beta.-epimeric mixture product described in Example 2 was dissolved in 1.2 liters of chloroform (which the supplier had stabilized by the addition of 0.6% ethanol). This solution was cooled to -40.degree. C., saturated with anhydrous hydrogen chloride, and maintained at -30.degree. for a period of 2 hours. The solution was then washed with water (3 liters) and with 1 liter of aqueous saturated sodium chloride solution, dried over anhydrous sodium sulphate, and evaporated to dryness under reduced pressure. The crystalline product, compound 4, (i.e. 11.beta.-hydroxy-17.alpha.,21-diacetoxy-6.alpha.-fluoro-pregn- 4-ene-3,20-dione) was obtained from ethyl acetate.

m.p.: about 231.degree.-234.degree. C.; optical rotation not taken.

Salient bands IR: 3520, 1750, 1730, 1670, 1620, 1240 cm.sup.-.sup.1

A mixture of 3 g of 11.beta.-hydroxy-17.alpha.,21-diacetoxy-6.alpha.-fluoro-pregn-4-ene-3,20-d ione [compound 4, the product of Example 3] , 2.5 g of 2,3-dichloro-4,5-dicyanobenzoquinone, and 30 ml of dioxan was heated under reflux for 2 hours. The resulting dark colored suspension was cooled, filtered from dichloro-dicyanohydroquinone, and then evaporated to dryness. The residue so obtained was dissolved in chloroform and filtered through neutral alumina (50 g). The solvent was evaporated under reduced pressure. The residue crystallized from methanol as colorless prisms, m.p.: 235.degree.-240.degree., .lambda.max, 241 nm (.epsilon.= 14,200). The analytical sample of 11.beta.-hydroxy-17.alpha.,21-diacetoxy-6.alpha.-fluoro-pregna-1,4-diene-3 ,20-dione (compound 5) was obtained by an additional recrystallization from methanol.

This product of Example 4 can be converted into fluoprednisolone according to the following procedure (Bjarte Loken and George Rosenkranz, U.S. Pat. No. 2,860,149, CA 91959, 14/1/57): To a suspension of 700 mg of this 17.alpha.,21-diacetate in 7 ml of the methanol while stirring under pure nitrogen at 23.degree. was added a solution of 130 mg of anhydrous potassium carbonate in 1.5 ml of water (this solution had been boiled to remove oxygen and then cooled under pure nitrogen). The reaction mixture was stirred under a nitrogen atmosphere at 23.degree.-25.degree. for a period of 60 minutes. 2-3 drops of acetic acid was added and the agitation continued until carbon dioxide evolution stopped. The product was precipitated with water containing salt, the solids filtered, washed on the filter with water, collected and dried.

The weight of the crude product corresponded to slightly less than the quantitative yield. Recrystallization from acetone gave the pure triol of compound 6: 11.beta.,17.alpha.,21-trihydroxy-6.alpha.-fluoro-pregna-1,4-diene-3,20-dio ne known as fluprednisolone M.P., 207.degree.-210.degree., [.alpha.].sub.D + 92.degree..

To a solution of 7.0 g of 11.beta.-hydroxy-17.alpha.,21-diacetoxy-6.alpha.-fluoro-pregn-4,ene-3,20-d ione (compound 4) obtained according to Example 3, in 16 ml of dry pyridine was added 3 ml of trimethyl chlorosilane. This mixture was stirred for 2 hours at room temperature and then diluted with 200 ml methyl isobutyl ketone, and 32 ml of 6N sulfuric acid was added with agitation and cooling. The mixture was allowed to solidify, the organic layer separated and washed successively with dilute sulfuric acid, 10% aqueous sodium bicarbonate and finally with water. Concentration to a small volume under reduced pressure gave a crystalline slurry from which the 11.beta.-trimethylsiloxy-17.alpha.,21-diacetoxy-6.alpha.-fluoro-pregn-4-en e-3,20-dione (compound 11) could be collected.

Obtained: 8.05 g.

A mixture of 4.0 g of 11.beta.-trimethylsiloxy-17.alpha.,21-diacetoxy-6.alpha.-fluoro-pregn-4,en e-3,20-dione (from Example 5) 2 g anhydrous potassium acetate and 10 dimethyl formamide was heated for a period of 4 hours at 110.degree.-120.degree. in an atmosphere of nitrogen. The reaction mixture was cooled and poured into 60 ml of water and ice. The solids were filtered after 2 hours of standing, washed on the filter with water and dried. Trituration of this material with acetone ether gave the 11.beta.-trimethyl siloxy-21-acetoxy-6.alpha.-fluoro-pregna-4,16-diene-3,20-dione in a crystalline form weighing 3.5 g (compound 12). TLC spotting showed the substance to be homogenous.

A mixture of 3.5 g of 11.beta.-trimethylsiloxy-21-acetoxy-6.alpha.-fluoropregna-4,16-diene-3,20- dione obtained according to Example 6, 2.65 g of 2,3-dichloro-4,5-dicyanobenzoquinone (DDQ) and 30 ml of dioxan was heated under reflux for 2 hours. The resulting dark colored suspension was cooled, filtered from dichloro-dicyanohydroquinone, and evaporated to dryness. The residue was dissolved in chloroform and passed through a short column of neutral alumina (50 g). The solvent was evaporated under reduced pressure. The residue was crystallized from methanol. The analytical sample of 11.beta.-trimethylsiloxy-21-acetoxy-6.alpha.-fluoropregna-1,4,16-triene-3, 20-dione (compound 13) was obtained by an additional recrystallization from methanol. The yield was 2.65 g.

750 mg cuprous iodide was suspended in 10 ml dichloromethane in an atmosphere of nitrogen and cooled to -10.degree. C. To this suspension under stirring and cooling was added 9 ml of a 2% solution of lithium-methyl by drop wise addition maintaining the temperature below -5.degree.. To the so obtained lithium-dimethyl-copper solution was added 425 mg 11.beta.-trimethylsiloxy-21-acetoxy-6.alpha.-fluoropregna-1,4,16-triene-3, 20-dione obtained according to Example 8 dissolved in 3 ml dichloromethane. The temperature of the reaction mixture was kept below -1.degree., and the stirring was maintained for a period of 15 minutes. The reaction mixture was poured into an aqueous ammonium chloride solution, the organic phase separated and the aqueous phase re-extracted with dichloromethane. The organic extracts were combined, washed with water, dried over anhydrous sodium sulfate and concentrated to dryness under reduced pressure. Trituration with acetone afforded a crystalline product 11.beta.-trimethylsiloxy-21-acetoxy-6.alpha.-fluoro-16.alpha.-methyl-pregn a-1,4-diene-3,20-dione (compound 14) yield 320 mg TLC: homogeneous.

1 g of 11.beta.-trimethylsiloxy-21-acetoxy-6.alpha.-fluoro-16.alpha.-methylpregna -1,4-diene-3,20-dione obtained according to Example 8 was powdered finely and suspended in 10 ml of 49% aqueous hydrofluoric acid with stirring at 0.degree.. After 5 minutes of contact about 5% of the starting material remained in a pilot experiment (by TLC, BPA). The reaction product was then neutralized by pouring into 12% sodium carbonate solution (200 ml), the solids were filtered, washed with water, collected, dried and recrystallized from acetone. There was obtained 600 mg of 11.beta.-hydroxy-21-acetoxy-6.alpha.fluro-16.alpha.-methylpregna-1,4-diene -3,20-dione (compound 15). Fluocortilone--a well known anti-inflammatory corticosteroid--can be obtained by hydrolysis of compound 15 (ref: B. Loken, G. Rosenkranz, U.S. Pat. No. 2,860,143, C.A. 91959, Jan. 14, 1957).

A solution of 200 g of 11.beta.-trimethylsiloxy-21-acetoxy-6.alpha.-fluoro-pregna-4,16-diene-3,20 -dione obtained according to Example 6 (compound 12) in a mixture of 40 liters acetone was charged into the holding tank of a tube reactor. In the other holding tank was charged a solution of 150 g potassium permanganate in 12.5 liter water, and 50 ml formic acid. With the help of the dosage pumps the two solutions were combined into the reaction tube at a feed rate of 80 and 25 ./hour respectively. The reactor tube was dimensioned so that the two solutions were in contact at -5.degree. for a period of 3 seconds with the reaction product flowing directly from the reaction tube into 10 liters of an aqueous sodium bisulfite solution containing 500 gms of sodium bisulfite (which stops the reaction). The suspension was filtered to remove manganese dioxide, the filtrate was concentrated under reduced pressure to about 2 liters and diluted with 2.5 liters of water. The crystalline slurry was left for 4 hours then filtered and the solids washed on the filter with water, collected and dried. Recrystallization from acetone/ether gave 190 g of purified 11.beta.trimethylsiloxy-16.alpha.,17.alpha.-dihydroxy-21-acetoxy-6.alpha. -fluoro-pregna-4-ene-3,20-dione (compound 16).

3.5 g of 11.beta.-trimethylsiloxy-16.alpha.,17.alpha.-dihydroxy-21-acetoxy-6.alpha. -fluoro pregn-4-ene-3,20-dione (from Example 10) was treated with 35 ml 49% aqueous hydrofluoric acid exactly as described in Example 9. Obtained was 11.beta.,16.alpha.,17.alpha.-trihydroxy-21-acetoxy-6.alpha.-fluoropregn-4- ene-3,20-dione, which after mild saponification gave 2.0 g of 11.beta.,21,16.alpha.,17.alpha.-tetra-hydroxy-6.alpha.-fluoropregn-4-ene-3 ,20-dione (compound 17), which is Flurandrenolone, a well known anti-inflammatory steroid. Reaction of this compound with acetone under acid catalysis according to known methods leads to formation of the 16,17-acetonide derivative (known as Cordran, Drenisone and Drocort).

When 7.0 of 11.beta.,17.alpha.-dihydroxy-21-acetoxy-9.alpha.-fluoro-16.alpha.-methyl-5 .alpha.-pregnane-3,20-dione, prepared according to Carrington et al J. Chem. Soc. 1961, 4560, was treated exactly as in Example 5 with trimethylchlorosilane in pyridine solution there was obtained 1.7 g of 11.beta.-trimethylsiloxy-17.alpha.-hydroxy-21-acetoxy-9.alpha.-fluoro-16.b eta.-methyl-5.alpha.-pregnane-3,20-dione (compound 23).

When 3.5 g of the product of Example 12 (11.beta.-trimethylsiloxy-17.alpha.-hydroxy21-acetoxy-9.alpha.-fluoro-16.b eta.-methyl-5.alpha.-pregnane-3,20-dione) was treated in all details as in Example 7 with 2.65 g of 2,3-dichloro-4,5-dicyanobenzoquinone in dioxane solution, there was obtained 1.7 g of 11.beta.-trimethylsiloxy-17.alpha.-hydroxy-21-acetoxy-9.alpha.-fluoro-16.b eta.-methyl-pregna-1,4-diene-3,20-dione (compound 24).

To a slurry of 1.7 g of 11.beta.-trimethylsiloxy-17.alpha.-hydroxy-21-acetoxy-9.alpha.-fluoro-16.b eta.-methyl-pregna-1,4-diene-3,20-dione (obtained according to Example 13) in 16 ml of methanol was slowly added a solution of 170 mg potassium hydroxide in 4 ml of methanol. This mixture was stirred in inert atmosphere at 10.degree. for 1 hour. TLC check indicated that the starting material had disappeared and the more polar 21-hydroxy compound had been formed. The reaction product was poured into a large volume of water (450 ml), neutralized by addition of a little acetic acid, and most of the clear supenatant solution was decanted. The remaining slurry of solids was filtered the solids washed on the filter with abundant water, collected and dried. This resulting 21-hydroxy substance was powdered and treated with 48% aqueous hydrofluoric acid (17 ml) exactly as described in Example 10, but it took 8 minutes of contact until the starting material had disappeared. The crude water precipitated product was recrystallized from acetone and gave 1.2 g of 11.beta.,17.alpha.,21-trihydroxy-9.alpha.-fluoro-16.beta.-methylpregna-1,4 -diene-3,20-dione (Betamethasone, a strong topical anti-inflammatory agent).

The starting material for Compound 32 ("tetrahydrodexamethasone acetate") is 17.alpha.hydroxy-21-acetoxy-9.beta.,11.beta.-epoxy-16.alpha.methyl-5.al pha.-pregnane-3,20-dione, disclosed in patent 3,876,633 which may be prepared according to Example 20 of the patent. The steroid was reacted with a THF/HF complex prepared by absorbing 40 g of HF in 80 ml of tetrahydrofuran (precooled to -30.degree. C.) at temperatures controlled to at least -20.degree. C.

10 g of the steroid was added to the fresh (cold) THF/HF complex and vigorous agitation was maintained while the temperature was permitted to climb to 0.degree. C. over a 2-3 hour period. The reaction mixture was maintained at 0.degree. until TLC showed disappearance of the starting steroid.

The reacted complex was then poured into 800 ml of water containing 150 g of sodium acetate and 300 g of ice. The mixture was agitated at 0.degree. -5.degree. for several hours, then filtered, washed on the filter and dried. Recrystallization from ethyl acetate gave 8.5 g of pure 11.beta.,17.alpha.-dihydroxy-21-acetoxy-9.alpha.-fluoro-16.alpha.-methyl-5 .alpha.-pregnane-3,20-dione. When 7.0 g of "tetrahydro dexamethasone acetate", e.g. 11.beta.,17.alpha.-dihydroxy-21-acetoxy-9.beta.-fluoro-16.beta.-methyl-5.a lpha.-pregnane-3,20 -dione (compound 32) was treated exactly as in Example 5 with trimethylchlorosilane in pyridine solution, there was obtained 8.5 g of 11.beta.-trimethylsiloxy-17.alpha.hydroxy-21-acetoxy-9.alpha.-fluoro-16.al pha.-methyl-5.alpha.-pregnane-3,20-dione (compound 33).

When 3.5 g of the (compound 33) product of Example 15 (11.beta.-trimethylsiloxy-17.alpha.hydroxy-21-acetoxy-9.alpha.-fluoro-16.a lpha.-methyl-5.alpha.-pregnane-3,20-dione) was treated in all details as in Example 7 with 2.65 g of 2,3-dichloro-4,5-dicyanobenzoquinone in dioxane solution there was obtained 1.85 g of 11.beta.-trimethylsiloxy-17.alpha.hydroxy21-acetoxy-9.alpha.-fluoro-16.alp ha.methyl-pregna-1,4-diene-3,20-dione (compound 34).

When 1.7 g of 11.beta.-trimethylsiloxy-17.alpha.-hydroxy-21-acetoxy-9.alpha.-fluoro-16.a lpha.-methylpregna-1,4-diene-3,20-dione from Example 16 was treated exactly as described in Example 14, there was obtained 1.3 g of 11.beta.,17.alpha.,21-trihydroxy-9.alpha.-fluoro-16.alpha.-methyl-pregna-1 ,4-diene-3,20-dione (Dexamethasone, a strong topical anti-inflammatory agent).

To a solution of 7 g of 11.beta.,17.alpha.dihydroxy-21-acetoxy-9.alpha.-fluoro-5.alpha.-pregnane-3 ,20 -dione (which was prepared from hecogenin according to J. Elks, G. H. Phillipps and W. F. Wall: J. Chem. Soc., 1958, 4001) (Compound 41) in 16 ml of dry pyridine was added 3 ml of trimethyl chlorosilane This mixture was treated in all details exactly as described in Example 5; there was obtained 8.11 g of 11.beta.-trimethylsiloxy-17.alpha.-hydroxy-21-acetoxy9.alpha.-fluoro-5.alp ha.-pregnane-3,20-dione (compound 42).

A mixture of 8 g of 11.beta.-trimethylsiloxy-17.alpha.hydroxy-21-acetoxy-9.alpha.fluoro5.alpha .pregnane-3,20dione obtained according to Example 18, 16 ml of acetic anhydride and 150 mg paratoluenesulfonic acid was heated under stirring at 55.degree. -70.degree. for 45 min. The reaction mixture was cooled to about 35.degree. -40.degree., 15 ml of dimethyl formamide and 3 ml of a 10% aqueous solution of sodium acetate added, and the temperature was controlled to remain below 45.degree. -50.degree.. Finally, the reaction mixture was poured into 350 ml of ice and water and a close to quantitative yield (8.65 g) of the diacetate (compound 43) was collected by filtration. The analytical sample of 11.beta.-trimethylsiloxy-17.alpha.,21-diacetoxy9.alpha.-fluoro5.alpha.-pre gnane-3,20-dione was obtained by recrystallization from dichloromethane-methanol.

A mixture of 8.0 g of 11.beta.-trimethylsiloxy-17.alpha.,21-diacetoxy9.alpha.fluoro-5.alpha.-pre gnane3,20-dione obtained according to Example 19, 4 g anhydrous potassium acetate and 20 ml of dimethyl formamide was heated for a period of 4 hours at 110.degree. -120.degree. in an atmosphere of nitrogen. The reaction mixture was cooled and poured into 120 ml of water and ice. The mixture was left stirring for 30 minutes, then lft standing for a period of 2 hours, filtered, washed on the filter with abundant water, collected and dried. Trituration of this crude product with acetone/ether gave 11.beta.-trimethylsiloxy-21-acetoxy-9.alpha.-fluoro-5.alpha.-pregn-16-ene- 3,20-dione (compound 44). Yield: 7.0 g.

When 3.5 g of the product of Example 20 (11.beta.trimethylsiloxy-21-acetoxy9.alpha.-fluoro-5.alpha.-pregn-16-ene-3 ,20-dione) was treated in all details as in Example 8 with 2.65 g of 2,3-dichloro 4,5-dicyanobenzoquinone in dioxane solution, there was obtained 2.0 g of 11.beta.-trimethylsiloxy-21-acetoxy-9.alpha.fluoro-pregna-1,4,16-triene-3, 20-dione (compound 45).

A solution of 11.beta.-trimethylsiloxy-21-acetoxy-9.alpha.-fluor-pregna-1,4,16-triene-3, 20-dione (1 g) from Example 21 (or from Example 28) in acetone (50 ml) and formic acid (1 ml) was cooled to 0.degree. and mixed quickly with an ice-cold solution of potassium permanganate (0.7 g) in water (20 ml) and acetone (30 ml). After 5 seconds the reaction was quenched with a solution of sodium sulfite (1.5 g) in water (10 ml). The reaction mixture was filtered and the colorless filtrate was concentrated until no acetone remained. After 30 minutes standing, the precipitate was collected, washed with water and dried under vacuum. There was obtained 0.94 g of (compound 46) 11.beta.-trimethylsiloxy-16.alpha.,17.alpha.-dihydroxy-21-acetoxy-9.alpha. -fluoropregna-1,4-diene-3,20-dione. m.p. about 203.degree. -204.degree.; TLC, homogenous (BPA).

IR bands at 3450, 3400, 1760, 1630, 1615, 1260, 1240, 1080, 845 cm.sup..sup.-1.

Larger amounts of the 1,4,16-triene starting material is better treated in a tube reactor as described in Example 9.

0.94 g of 11.beta.-trimethylsiloxy-21-acetoxy-9.alpha.-fluoropregna-1,4,16-triene-3, 20-dione (from Example 22) in 10 ml 49% aqueous hydrofluoric acid was stirred at 0.degree. for 10 minutes. TLC check in BPA indicated that no starting material was left after this relatively short period of contact, and that the only product was triamcinolone 21-acetate (compound 47a). The reaction product was neutralized by pouring into 12% sodium carbonate solution (200 ml) the solids were filtered, washed with abundant water, collected and dried. The pure triamcinolone tetrol (compound 47b) was obtained by saponification in methanol suspension (ref. B. Loken. G. Rosenkranz, U.S. Pat. No. 2,860,149, CA 91959, Jan. 14, 1957). Reaction of the triamcinolone tetrol (which in itself is a well known anti-inflammatory compound) with acetone under acid catalysis according to known methods leads to the formation of the Triamcinolone Acetonide (11.beta.,21-dihydroxy-16.alpha.,17.alpha.isopropylidene-dioxy-9.alpha.flu oro-pregna-1,4-diene3,20-dione) which is extensively used as an anti-inflammatory steroid (compound 49 b).

To a suspension of 10 g of 11.beta.-trimethylsiloxy-16.alpha.,17.alpha.-dihydroxy-21-acetoxy-9.alpha. -fluoropregna-1,4-diene-3,20-dione (compound 46, from Example 22) in 100 ml of acetone, was added a solution of 0.2 ml perchloric acid (70%) in 10 ml of acetone. The mixture was stirred for 16 hours at room temperature, 20 ml of a saturated sodium bicarbonate solution added, and the mixture was concentrated to about 15 ml under reduced pressure. Complete precipitation was obtained by addition of 200 ml water. The product was filtered off, washed well with water, collected and dried. There was obtained (compound 48) 11.beta.-trimethylsiloxy-21 -acetoxy-16.alpha.,17.alpha.-isopropylidenedioxy-9.alpha.-fluoro-pregna-1, 4-diene-3,20-dione, in close to quantitative yield (10.75 g). Recrystallization from acetone-hexane provided the analytical sample which was pure on TLC (BPB).

IR bands are at 1760, 1730, 1670, 1635, 1615, 1260, 1230, 1085, 845 cm.sup.-.sup.1.

1 g of 11.beta.-trimethylsiloxy21-acetoxy-16.alpha.,17.alpha.-isopropylidenedioxy -9.alpha.-fluoro-pregna-1,4-diene-3,20-dione (from Example 24) was powdered finely and suspended in 10 ml 49% aqueous hydrofluoric acid by stirring at 0.degree.. After 7 minutes contact period about 5% of the starting material in a model experiment remained (by TLC (BPA)). The reaction product, after exactly 7 minutes contact, was neutralized by pouring into 12% sodium carbonate solution (200 ml), the solids were filtered, washed with water, collected and dried. Triamcinolone Acetonide, 21-acetate was obtained (compound 49a). The pure product was obtained in about 80% yield by recrystallization from chloroform-ethyl acetate. (If desired, Triamcinolone Acetonide can be obtained by saponification of the 21-acetate group.)

To a solution of compound 51, namely 9.alpha.-fluoroprednisolone, 21-acetate (18 g) in pyridine (54 ml) was added trimethylchlorosilane (8 ml). The mixture was stirred at 25.degree. for a period of 60 hours, diluted with 200 ml of methylisobutyl ketone and washed with 10% aqueous sulfuric acid (2 .times. 200 ml) and water (260 ml). After drying with sodium sulfate, the solution was concentrated to a small volume under reduced pressure, to give a colorless crystalline residue of the 11-trimethylsilyl ether (compound 52) in quantitative yield. TLC indicated a homogenous product (System BPB, silica gel plate 254 )

Salient IR bands: 3400, 1760, 1730, 1665, 1620, 1610, 1260, 1240, 1080, 845 cm.sup.-.sup.1.

There was obtained 21.0 g of 17.alpha.-hydroxy-21-acetoxy-11.beta.-trimethylsiloxy-9.alpha.-fluoropregn -1,4-diene-3,20-dione. m.p. about 203.degree.-204.degree..

A mixture of the 17.alpha.-hydroxy-21-acetoxy-11.beta.-trimethylsiloxy-9.alpha.-fluoro-preg n-1,4-diene-3,20-dione from Example 26 (20.6 g) dichloromethane (50 ml) and acetic acid (20 ml) was treated with trifluoro-acetic anhydride (20 ml) for a period of 5 hours at 25.degree.. The reaction mixture was diluted with methyl-isobutylketone (200 ml) and extracted with sodium bicarbonate solution (500 ml) and then with water (200 ml). The organic layer was dried over sodium sulfate, and evaporated to dryness under reduced pressure. The residue was amorphous, homogenous in TLC (BPB, silica plates No. 254). Crystalline 17.alpha.,21-diacetoxy-11.beta.-trimethylsiloxy-9.alpha.-fluoropregna-1,4- diene-3,20 -dione (compound 53) was obtained directly from the residue by trituration with methanol.

Salient IR bands: 1760, 1740, 1670, 1635, 1610, 1260,1240, 1080, 840 cm.sup.-.sup.1.

Crude yield (residue): 22.2 g; crystalline (triturated with methanol): 19.8 g with m.p. 162.degree.-163.degree.).

A mixture of 17.alpha.,21-diacetoxy-11.beta.-trimethylsiloxy-9.alpha.-fluoropregna-1,4- diene-3,20-dione (4 g, from Example 27), anhydrous potassium acetate (4 g) and dimethylformamide (20 ml) was heated at 115.degree. for 1.5 hours. The reaction mixture was cooled and poured into cold water, and dried under vacuum. A sample was crystallized from methanol. Homogenous on TLC (benzene-ether, 1:1).

Salient IR bands: 1710, 1670, 1640, 1615, 1590, 1225, 1070, 845 cm.sup.-.sup.1.

Yield: 90-95%.

This product, which is 11.beta.-trimethylsiloxy-21-acetoxy-9.alpha.-fluoro-pregna-1,4,16-triene-3 ,20-dione m.p. 212.degree.-214.degree. (compound 45) was treated according to Example 22 and the product subsequently converted to triamcinolone according to the method described in Example 23, and also to the triamcinolone acetonide as described in Example 25.

7.0 g of prednisolone 21-monoacetate was treated exactly as in Example 5 with trimethylchlorosilane in pyridine solution, there was obtained 8.0 g of 11.beta.-trimethylsiloxy-17.alpha.-hydroxy-21-acetoxy-pregna-1,4-diene-3,2 0-dione. m.p. 180.degree.-184.degree. and 203.degree.-204.degree., two allotropic forms.

IR: 3300, 1755, 1735, 1665, 1620, 1605, 1270, 1240, 1060, 845 cm.sup.-.sup.1.

A mixture of 8 g of 11.beta. -trimethylsiloxy-17.alpha.-hydroxy-21-acetoxy-pregna-1,4-diene-3,20 -dione (from Example 29), 8 g of paratoluenesulfonic acid monohydrate and 800 ml of isopropenyl acetate was allowed to stand for one week at room temperature. By this time TLC control indicated that almost all the starting material had reacted. 15 g of anhydrous sodium acetate was added and the mixture left stirring for another day. The supernatant was decanted from the salt crystals and the resulting solution -- after addition of 400 ml methyl-isobutyl ketone -- was washed twice, each time with 500 ml of water, then with 200 ml of saturated aqueous sodium chloride solution, dried over anhydrous sodium sulfate, and evaporated under reduced pressure to dryness. This residue was found to be 11.beta.-trimethylsiloxy-17.alpha.,21-diacetoxypregna-1,4-diene-3,20-dione , about 90% pure, there was about 10% of a less polar artifact believed to be the result of a slight aromatization side reaction about ring A, followed by acetylation. The identity was proved by hydrolyzing off the silyl ether blocking group to give prednisolone 17,21-diacetate, identified by TLC with a sample obtained by the Ortho ester route.

4 g of 11.beta.-trimethylsiloxy-17.alpha.,21-diacetoxypregna-1,4-diene-3,20-dione (from Example 30) was treated with potassium acetate and dimethylformamide exactly as described in Example 28. There was obtained 11.beta.-trimethylsiloxy-21-acetoxypregna-1,4,16-triene-3,20-dione (compound 55).

The conversion sequence described in Examples 29, 30, 31 of prednisolone 21-monoacetate to compound 55 has not been illustrated on the drawing since it parallels the conversion of 51.fwdarw.52.fwdarw. 53.fwdarw.45, absent only presence of the 9.alpha.F substituent. The examples which follow describe a conversion sequence for compound 55 which parallels the conversions of compounds 45.fwdarw.49 again absent only the 9.alpha.F substituent.

A solution of 11.beta.-trimethylsiloxy-17.alpha.,21-diacetoxy-pregna-1,4,16-triene-3,20- dione (1 g of compound 55 from Example 31) in acetone (50 ml) and formic acid was treated with a solution of potassium permanganate and worked up exactly as described in Example 24.There was obtained 11.beta.-trimethylsiloxy-16.alpha.,17.alpha.-dihydroxy-21-acetoxy-pregna-1 ,4-diene-3,20-dione. TLC was homogenous, developed in BPA. Yield: 0.95 g.

0.95 g of 11.beta.-trimethylsiloxy-16.alpha.,17.alpha.-dihydroxy-21-acetoxy pregna-1,4 -diene-3,20 -dione from Example 32 was treated with 40% aqueous hydrofluoric acid exactly as described in Example 32, only this time with the 9.alpha.-fluoro substituent absent, the reaction was faster. The starting material had all reacted after 5 minutes of contact. No hydrolysis took place at the 21-acetate group, nor was there any aromatization side reaction about ring A, or any homo D annulation about ring D. There was obtained close to quantitative yield of 11.beta., 16.alpha., 17.alpha.-trihydroxy-21-acetoxy-pregna-1,4-diene-3,20-dione which was converted to the acetonide according to the known methods. Acetone, perchloric acid as a catalyst. This acetonide (11.beta.-hydroxy-21-acetoxy-16.alpha.,17.alpha.isopropylidene-dioxypregna -1,4-diene-3,20-dione) was saponified in methanol suspension (ref. B. Loken, G. Rosenkranz, U.S. Pat. No. 2,860,143, CA 91959, January 14, 1957) to give Prednacinolone Acetonide which is a well known anti-inflammatory compound (see Am. Med. A. 213 1325 (1970)) i.e. 11.beta.,21-dihydroxy-16.alpha.,17.alpha.-isopropylidendioxy-pregna-1,4-di ene-3,20-dione.