The present invention relates to a method for the determination of Trilostane. In particular, but not exclusively, the invention relates to a method for the determination of Trilostane in a mixture containing Trilostane and structurally-related impurities formed during the preparation of Trilostane.

Trilostane ((4α,5α,17β)-4,5-epoxy-3,17-dihydroxyandrost-2-ene-2-carb onitrile; 1) is used in human and veterinary medicine to treat endocrine disorders. Its primary indication is as an anticancer agent but it is also used in the treatment of Cushing's disease and other hyperadrenocortical conditions.

(1)

TLC and HPLC have been used to investigate Trilostane and structurally-related compounds. For example, Mori et al} describe the use of silica gels in preparative TLC methods for the separation and purification of Trilostane metabolites. Powles et al? and Robinson et al? disclose methods for the determination of Trilostane and a metabolite of Trilostane (the 17-keto analogue of Trilostane) in human plasma by HPLC employing a Hypersil octadecyl-substituted silica (ODS) column. Powles et al? also carried out TLC analysis of human plasma containing Trilostane using TLC plates coated with a silica gel. Brown et al.4 describe an HPLC method for the determination of Trilostane and its 17-keto analogue using a silica-based Partisil column and a silica-based Phenyl/Corasil precolumn. A study of Trilostane and ketotrilostane in rat plasma is described by McGee et al.5 involving the use of HPLC employing an ODS column. In addition to use in academic and clinical investigations, silica-based stationary phases are the only type of stationary phase employed in the analysis of compound mixtures obtained during the industrial preparation of Trilostane.

The applicant has however unexpectedly found that the use of a silica or silica-based stationary phase in TLC methods for the determination of Trilostane leads to degradation of the target compound (i.e. Trilostane). Furthermore, the results of HPLC analyses of mixtures containing Trilostane and structurally-related compounds indicate that a conventional silica or silica-based stationary phase does not always provide adequate separation of Trilostane from the other compounds present. Consequently, uncertainties surround the accuracy and reliability of previous methods used to analyse Trilostane- containing mixtures.

According to a first aspect of the present invention there is provided a method of determining Trilostane in a compound mixture comprising effecting a chromatographic separation of said mixture and analysing the eluate from the separation to determine Trilostane, characterised in that the chromatographic separation is effected using a chromatographic stationary phase comprised of a cross-linked resin incorporating aromatic groups.

Use of a stationary phase comprised of a cross-linked resin incorporating aromatic groups avoids the previously unidentified problems outlined above in relation to the use of a silica or silica-based stationary phase, hi particular, using a stationary phase in accordance with the present invention ensures that Trilostane is not degraded during the chromatographic separation process. An unforeseen advantage of the present invention is that, whereas previous methods employed a relatively polar silica or silica-based stationary phase, a relatively non-polar stationary phase comprised of a cross-linked resin incorporating aromatic groups facilitates sufficient separation of Trilostane from further compounds (including structurally-related compounds) to allow the accurate and reliable determination of Trilostane.

A second aspect of the present invention provides a method of separating Trilostane in a compound mixture comprising effecting a chromatographic separation of said mixture using a chromatographic stationary phase comprised of a cross-linked resin incorporating aromatic groups.

It is envisaged that a range of different stationary phases comprised of a cross-linked resin incorporating aromatic groups may be used in the method of the invention. It is preferred that the resin comprises styrene moieties. More preferably the resin comprises divinylbenzene moieties. Yet more preferably the resin is a cross-linked styrene- divinylbenzene resin.

The chromatographic separation may be HPLC. In this case, a stationary phase resin having any suitable bead size may be used. For example, the resin may have a bead size of up to approximately 10 μm, although it is preferred that the resin has a bead size of approximately 3 μm or approximately 5 μm. Furthermore, the resin may have a pore size of up to approximately 10 μm, for example the pore size of the resin may be approximately 3 μm, approximately 5 μm or approximately 7 μm. When HPLC is used, the HPLC column may have any appropriate length and/or diameter. Particularly suitable columns have a length of less than around 300 mm. By way of example, the HPLC column may have a length of about 50 mm, about 150 or about 250 mm. Moreover, the diameter of the HPLC column may be in the range 1 - 5 mm, more preferably in the range 2 - 4.5 mm. For example, the HPLC column may have a diameter of approximately 2.1 or approximately 4.1 mm.

When the method of the invention employs HPLC, the chromatographic separation may be effected at any desirable flow rate to ensure satisfactory separation of Trilostane over a reasonable timescale. It is preferred that the chromatographic separation is effected at a flow rate of up to approximately 1.0 ml/min, more preferably approximately 0.7 ml/min. The chromatographic separation according to the present invention can be carried out at ambient temperature or at an elevated temperature. Preferably the chromatographic separation is effected at a temperature of up to approximately 35 0C, more preferably approximately 28 0C.

The compound mixture will typically be introduced onto the chromatographic stationary phase as a solution (e.g. in methanol) and be eluted under the influence of a suitable mobile phase, preferably an aqueous mobile phase. Preferably the mobile phase has a pH in the range 4 - 10, more preferably 7.5 — 9.0, and most preferably a pH of approximately 8.3. Thus, the mobile phase may comprise a buffer to maintain the pH of the mobile phase at a desired level. A preferred buffer is an acetate salt, such as ammonium acetate. The composition of the mobile phase should be selected to optimise the separation process for a particular application. To this end, it is preferred that the mobile phase incorporates a water miscible polar organic liquid, e.g. acetonitrile. Under certain conditions acetonitrile may be the sole constituent of the mobile in which case the mobile phase consists of 100 vol. % acetonitrile. hi a preferred embodiment of the invention, a gradient elution method is employed in which the percentage concentration of at least some of the components of the mobile phase, e.g. ammonium acetate solution and/or acetonitrile, are varied during the chromatographic separation process. Preferably, when a mobile phase consisting of ammonium acetate solution and acetonitrile is used, the percentage concentration of acetonitrile in the mobile phase is increased from around 10 vol. % to around 100 vol. % during the chromatographic separation with the balance of the mobile phase being made up with an appropriate amount of ammonium acetate solution. Thus, in this case, as the percentage concentration of acetonitrile is increased from around 10 vol. % to around 100 vol. % the percentage concentration of ammonium acetate solution in the mobile phase is decreased from around 90 vol. % to around 0 vol. %.

The rate at which the percentage concentration of a particular component of the mobile phase is varied during the chromatographic separation depends on the composition of the mobile phase and the period of time over which the chromatographic separation is carried out. The rate is preferably constant but, in certain circumstances, it may be more preferable to employ a rate which varies during the separation. It is preferred that the rate at which the percentage concentration of acetonitrile in the mobile phase is increased during the chromatographic separation is less than 10 % per minute, more preferably the rate lies in the range 2 to 7 % per minute, and most preferably the rate is about 4.5 % per minute.

The chromatographic separation may be carried out over any suitable period of time provided it is sufficiently long to ensure adequate separation and/or determination of Trilostane and any other compounds of interest which may be present. While the appropriate period of time will be influenced by many different factors such as the physical and chemical nature of the stationary and mobile phases, the flow rate and the temperature, it is preferred that the separation is carried out over a period of less than approximately 1 hour, more preferably less than around 30 minutes and most preferably around 20 minutes.

Where it is desired to reuse a chromatographic stationary phase following a chromatographic separation, a cleaning and re-equilibration cycle may be employed. For example, once all of the compounds of interest have been eluted from the stationary phase, the mobile phase may be maintained at its final composition for a further period of time (e.g. a few minutes) to ensure any remaining unwanted compounds are eluted from the column. Once the stationary phase has been cleaned in this way, the composition of the mobile phase running over the stationary phase may be adjusted to its original starting composition to re-equilibrate the stationary phase ready for further use.

It is envisaged that the method of the present invention is especially suited to the determination of Trilostane and structurally-related impurities formed during the preparation of Trilostane, and/or for the separation of Trilostane from such impurities. While the present invention is well suited for use in HPLC methods, it is anticipated that the concept of using a stationary phase comprising a cross-linked resin incorporating aromatic groups may be applied to any suitable chromatographic technique. Moreover, it will be evident to the skilled person that the present invention may be applicable to both analytical and preparative chromatographic techniques.

The invention is illustrated with reference to the following non-limiting Examples and accompanying figures, in which:

Figure 1 is chromatogram illustrating the constituents of a first Trilostane containing mixture which has been subject to a method in accordance with the present invention employing an aromatic cross-linked resin stationary phase;

Figure 2 is a chromatogram illustrating the constituents of a second Trilostane containing mixture which has been subject to a chromatographic separation employing a silicia-based stationary phase; Figure 3 is a chromatogram illustrating the constituents of the second Trilostane containing mixture which has been subject to a method in accordance with the present invention employing an aromatic cross-linked resin stationary phase; and

Figure 4 is chromatogram illustrating the constituents of the second Trilostane containing mixture which has been subject to the same method as that employed to produce the results depicted in Figure 3 to test the reproducibility of the results obtained employing the method of the present invention. REFERENCE EXAMPLES

The following experiments were carried out to investigate the stability of Trilostane during chromatographic analysis on a silica-based stationary phase.

Two samples of Trilostane were analysed by silica-based TLC employing a chloroform : methanol : glacial acetic acid (94 : 4 : 2 % v/v) mobile phase. The expected spots due to Trilostane and known related impurities (due to the method of preparation of Trilostane employed) were observed on each TLC plate, however, a number of other spots were observed on each plate indicating the presence of various unknown impurities. When each sample was reanalysed a few days later, additional unidentified spots were observed indicating that Trilostane and/or other compounds on the silica-based stationary phase had degraded whilst resident on the silica-based stationary phase. The results of these experiments illustrate the unsuitability of silica-based TLC methods for the accurate and reliable determination of Trilostane and related impurities.

hi order to determine whether Trilostane degrades whilst in contact with a silica-based TLC stationary phase, a two dimensional TLC experiment was performed. A first dimension TLC experiment was carried out using a chloroform : methanol : glacial acetic acid (94 : 4 : 2 % v/v) mobile phase. The expected pattern of spots was obtained including those relating to Trilostane, known impurities and unknown impurities. The TLC plate was then rotated by 90 ° and a second dimension TLC experiment carried out using the same conditions as the first dimension experiment. If Trilostane was not degrading on the silica- based stationary phase, the second dimension experiment should have produced a single Trilostane spot because all related impurities should already have been separated from Trilostane in the first dimension experiment. However, this was not what was observed. Following the second dimension experiment, a series of new spots were observed which strongly indicated that Trilostane had degraded on the silica-based stationary phase. EXAMPLE 1

In view of the problems described in the Reference Example in relation to the degradation of Trilostane on a silica-based stationary phase, an improved HPLC system, described below, was developed for the determination of Trilostane.

Preparation of Test Mixture Containing Trilostane and Compounds X and Z

100 mg of Trilostane was accurately weighed into a 100 ml volumetric flask and 80 ml of methanol added to form an initial solution of Trilostane in methanol. 25 mg of a structurally-related compound X and 25 mg of a structurally-related compound Z were accurately weighed into separate 25 ml volumetric flasks, dissolved in and made up to volume (i.e. up to 25 ml) with methanol. 2 ml of the solutions containing compounds X and Z was pipetted into the 100 ml flask containing the initial Trilostane solution in methanol and made to volume with further methanol to provide a Test Mixture containing 1 mg/ml Trilostane spiked with 0.02 mg/ml compound X and 0.02 mg/ml compound Z.

Preparation of Mobile Phase

Eluant A - 0.770 g of ammonium acetate was accurately weighed into a 1 litre mobile phase jar along with 1 litre of deionised water. The pH of the solution was adjusted to pH 8.3 using ammonia solution (0.88 Sp. Gr) then sonicated prior to use. This provided a 10 mM solution of ammonium acetate.

Eluant B - I litre of HPLC grade acetonitrile (1 M), which was sonicated prior to use.

HPLC Conditions

Stationary Phase: styrene-divinylbenzene copolymer having a bead size of 3 μm Mobile phase: ammonium acetate / acetonitrile, gradient as in Table 1 Flow rate: 0.7 ml/min Temperature: 28 0C Injection volume: 10 μl DAD* wavelength: 280 nm for Trilostane and compound X, 238 ran for compound Z *Diode Array Detection. Mobile Phase Gradient

Table 1

Between 0.0 and 20.0 minutes the concentration of Eluant A was decreased from 90.0 % to 0.0 % and the concentration of Eluant B was increased from 10.0 % to 100.0 % at a controlled rate of 4.5 % per minute. Between 20.0 and 25.0 minutes the column was washed with a mobile phase consisting of 100.0 % Eluant B. Between 25.1 and 30.0 minutes the stationary phase was equilibrated with the use of a mobile consisting of 90.0 % Eluant A and 10.0 % Eluant B.

HPLC analysis of the test mixture containing Trilostane and compounds X and Z was carried out using the above HPLC conditions. The DAD results obtained were used to produce a chromatogram, shown in Figure 1, and to calculate the results presented in Table 2.

Table 2

The results presented in Figure 1 and Table 2 illustrate that the above HPLC system employing an aromatic cross-linked resin stationary phase does not cause degradation of Trilostane (or compounds X and Z) and facilitates good separation of Trilostane from structurally-related compounds X and Z. The method of the present invention therefore provides a robust, accurate and reliable method for the determination of Trilostane in a mixture containing Trilostane and at least one further compound. EXAMPLE 2

The following experiments were carried out to compare two different HPLC systems for the analysis of a standard mixture containing Trilostane, the two structurally-related X and Z employed in Example 1 and a further structurally-related compound Y. A first system employed a silica-based stationary phase and a second system employed a stationary phase comprised of a cross-linked resin incorporating aromatic groups in accordance with the present invention. Two experiments were carried out using the second system to determine the reliability of the inventive method.

Preparation of Test Mixture Containing Trilostane and Compounds X, Y and Z

The test mixture used in Example 2 was prepared in accordance with the procedure described in Example 1 for the preparation of the test mixture containing Trilostane and compounds X and Z save for appropriate modification such that the test mixture was spiked with 0.02 mg/ml of compounds X, Y and Z.

Silica-Based Stationary Phase

HPLC Conditions

Stationary Phase: octadecyl substituted silica having a bead size of 5 μm Mobile Phase: water : methanol : acetonitrile : phosphoric acid (45:23:31:1 % v/v), isocratic Flow rate: 2.0 ml/min Temperature: 400C Injection volume: 10 μl DAD wavelength: 254 nm

HPLC analysis of the standard mixture containing Trilostane and compounds X, Y and Z was carried out using the above HPLC conditions. The DAD results obtained were used to produce a chromatogram, shown in Figure 2, and to provide the results presented in Table 3. Table 3

The results presented in Figure 2 and Table 3 illustrate that a unique peak for compound X cannot be resolved from the predominant Trilostane peak. This demonstrates that the HPLC system employing a silica-based stationary phase does not provide adequate separation of compound X from Trilostane. This method is therefore not suitable for the determination of Trilostane and the structurally-related compounds X, Y and Z in a mixture of the four compounds.

Aromatic Cross-linked Resin Stationary Phase

Preparation of Mobile Phase

An ammonium acetate (10 mM, pH 8.3) / acetonitrile mobile phase was prepared in accordance with the procedure set out in Example 1.

HPLC Conditions

Stationary phase: styrene-divinylbenzene copolymer having a bead size of 3 μm Mobile phase: ammonium acetate / acetonitrile, gradient as in Example 1, Table 1 Flow rate: 0.7 ml/min Temperature: 28 0C Injection volume: 10 μl DAD wavelength: 280 nm for Trilostane and compound X, 244 for compound Y and 238 nm for compound Z Two separate HPLC analyses of the test mixture containing Trilostane and compounds X, Y and Z were carried out using the above HPLC conditions. The DAD results obtained were used to produce two chromatograms (Run 1: Figure 3 and Run 2: Figure 4) and to provide the results presented in Table 4.

Table 4

The results presented in Figures 3 and 4, and Table 4 show that the HPLC system employing an aromatic cross-linked resin stationary phase does not cause degradation of Trilostane (or compounds X, Y and Z) and facilitates good separation of Trilostane from structurally-related compounds X, Y and Z. Thus, the HPLC method of the present invention provides an accurate and reliable method for the determination of Trilostane and represents a significant advance over existing methods, which have now been shown to be unreliable and potentially inaccurate. REFERENCES

1. Mori Y., Makoto T. and Makoto S., Chem. Pharm. Bull, 1981, 29(9), 2478-84. 2. Powles P., Robinson D.T., Andrews R.S. and Robinson P.R., J. Chromatogr., 1984, 311(2), 434-42. 3. Robinson D.T., Earnshaw RJ., Mitchell R., Powles P., Andrews R.S. and Robertson W.R., J. Steroid Biochem., 1984, 21(5), 601-5. 4. Brown, Stroshane and Benziger, J. Chromatogr., 1985, 339, 440-44. 5. McGee, Palin and Shaw, J. Chromatogr., 1991, 567, 282-7.