ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY PAMELA M. GORMAN A N D H O N G JIANG

Pfizer Global Research & Development Division, Groton, CT 06340

I. INTRODUCTION TO THIN-LAYER CHROMATOGRAPHY (TLC) A. History B. Fundamentals of TLC C. Why UseTLC?

II. TLC APPLICATIONS IN PHARMACEUTICAL INDUSTRY III. TLC METHOD DEVELOPMENT AND VALIDATION

A. Drug Substance Method Development and Validation B. Drug Product Method Development and Validation

IV IMPURITY ISOLATION AND CHARACTERIZATION BY TLC A. TLC-Specified Impurities B. Known Impurities C. Unknown Impurities D Summary of Impurity Isolation and Characterization by TLC REFERENCES

I. INTRODUCTION TO THIN-LAYER CHROMATOGRAPHY (TLC)

Thin-layer chromatography (TLC) is one of the most popular and widely used separation techniques because of its ease of use, cost-effectiveness, high sensitivity, speed of separation, as well as its capacity to analyze multiple samples simultaneously. It has been applied in the disciplines of biochemistry,^'^ toxicology,^'"^ pharmacology,^'^ environmental science,'^ food science,^'^ and chemistry.^^'^^ TLC can be utilized for separation, isolation, identification, and quantification of components in a mixture. It can also be utilized on a preparative scale to isolate a particular component. A large variety of TLC apparatus is commercially available.

203

2 0 4 p. M. GORMAN AND H. JIANG

A. History

Pioneer work in thin-layer chromatography to isolate and analyze medicinal compounds was performed by Izmailov and Shraiber on unbound alumina as early as 1938.^"' However, E. Stahl introduced the term "thin-layer chromatography" in 1956, which was considered the beginning of modern TLC.^^ Since the 1960s, commercialization of precoated TLC plates and automation of sample application and detection have made it accessible to all laboratories. A number of valuable texts have been written about the history of TLC.^^-^^ The most recent one is reviewed by C. F. Poole.

B. Fundamentals of TLC

In TLC, the sample is applied as a small spot or streak to the marked origin of stationary phase supported on a glass, plastic, or metal plate. The sample solvent is allowed to evaporate from the plate that is then placed in a closed chamber containing a shallow pool of mobile phase at the bottom. The mobile phase moves through the stationary phase by capillary forces. The components of the mixture migrate at different rates during movement of the mobile phase through the stationary phase. When the mobile phase has moved an appropriate distance, the plate is removed from the chamber and the solvent front is marked. Mobile phase is evaporated from the plate by drying at room temperature, by forced air flow, or in a heated oven. If the components of the mixture are not naturally colored or fluorescent, a detection reagent is applied to visualize the bands. Sometimes more than one detection technique is used to ensure the detection of all components in the mixture.

The Rf value is a convenient way to express the position of the substance on a developed chromatogram. It is calculated as follows: Rf = distance of component from origin/distance of solvent front from origin.

A variety of sorbents have been used as the stationary phase in TLC, including silica gel, cellulose, alumina, polyamides, ion exchangers, chemically modified silica gel, and mixed layers of two or more materials, coated on a suitable support. Currently in the pharmaceutical industry, commercially precoated high-performance TLC (HPTLC) plates with fine particle layers are commonly used for fast, efficient, and reproducible separations. The choices of mobile phase range from single component solvent systems to multiple-component solvent systems with the latter being most common. The majority of TLC applications are normal phase, which is also a complementary feature to HPLC that uses mostly reverse-phase columns.

The migration of each component in a mixture during TLC is a result of two opposing forces: capillary action of the mobile phase and retardation action of the stationary phase. Both forces contribute to achieve differential migration of each component. Developed TLC plates can be detected by various means, based on the nature of the sample. They could be non­destructive (UV/densitometer), destructive (derivatizing agents), or the combination of both. The results can be documented by photography and saved electronically for archiving and future reference.

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY 2 0 5

C. WhyUseTLC?

Today, while HPLC is widely used for separation and quantification, TLC remains a valuable and commonly used separation technique because of its features complementary to HPLC. The majority of TLC applications use normal-phase methods for separation, whereas reversed-phase methods dominate in HPLC. Some of the most important features of TLC compared to HPLC are briefly discussed here.

1. Open format of stationary phase and evaluation of the whole sample—In TLC separation, a mixture is applied to the stationary phase followed by development. It is an open system from separation to detection. In contrast to TLC, HPLC is a closed-column system in which a mixture is introduced into the mobile phase and solutes are eluted and detected in a time-dependent manner. There are times that TLC reveals new and unexpected information about the sample, while that information is lost in HPLC by retention on the column, because of strongly sorbed impurities, early elution, or detection. In addition, TLC has little or less contamination with a disposable stationary phase while in HPLC the column is used repeatedly.

2. Simple sample preparation—Samples for TLC separation often involve fewer cleanup steps because every sample is separated on fresh stationary phase, without cross-contamination or carry-over. Even strongly absorbed impurities or soHd particles in samples are not of much concern. This would be a disaster for HPLC separation, leading to column buildup, decay, and eventually destroying the performance.

3. High sample throughput—The simultaneous but independent applica­tion and separation of multiple samples in TLC results in higher sample throughput and less time consumption, as well as lower costs. HPLC cannot compete with TLC in terms of the number of samples processed in a given time period.

4. Flexible and versatile dissolving solvent and mobile phase—The choice of the sample solvent is not as critical in TLC as in HPLC because it is removed by evaporation before development of the plate. On the contrary, in HPLC the dissolving solvent chosen must be compatible in terms of composition and strength with the column and mobile phase. The same logic applies to the TLC mobile phase that is completely evaporated before detection. Therefore, the UV-absorbing properties, purity, or acid and base properties of the mobile phase are not as crucial as with HPLC. In addition, there is less solvent waste in TLC than in HPLC.

5. General and specific detection methods^^"^^—In TLC, many different detection methods, including inspection under short- and long-wave UV light and sequential application of a series of compatible reagents, can be employed on a developed TLC plate. Most well-established and routinely used HPLC detection methods still use UV, which cannot always capture all the eluates from column separation.

206 p. M. GORMAN AND H. JIANG

6. One-dimensional multiple development and two-dimensional devel­opment—Multiple development through one or two dimensions can be applied to separate certain components in sequence, with detection at each step. This gives a theoretical increase in the capacity of the spots, so it is ideal for the separation of mixtures with a large number of components. In addition, it is a useful tool to confirm the purity of a given component. Though hyphenated HPLC could serve as a multiple separation technique, TLC takes the lead in this area by its faster separation and choice of different mobile phases and detection methods through each run.

The comparison of TLC and HPLC has been described in a number of publications by a number of authors.^^'^^~~^

II. TLC APPLICATIONS IN PHARMACEUTICAL INDUSTRY

Sherma's biennial review^ on planar chromatography shows that the greatest number of new TLC and TLC-densitometry methods were published for pharmaceutical analysis, including applications identification, purity testing, assay, stability testing, and content uniformity testing of drug products, intermediates, and raw materials, as well as analysis of drugs and their metabolites in biological samples. TLC testing is relied upon because of its simplicity, flexibility, speed of analysis, and unique detection methods on both a qualitative and quantitative basis.

Often synthetic intermediates, process-related impurities, and degradation products do not contain chromophores that can be detected by the UV detector. One can utilize other detection methods in conjunction with HPLC or some other separation techniques to monitor the impurity. But none of them is as simple, fast, and straightforward as TLC. Therefore, these impurities are often specified by TLC, especially when there is a functional group that can be easily detected with certain derivatizing agents in TLC. Sometimes impurities elute at the solvent front in the HPLC method. It can be difficult to monitor and quantitate them, and changes to the HPLC mobile phase or column may not be able to adequately resolve them. On the other hand, sometimes impurities elute very late in the HPLC run, which can leave them undetected. One advantage of TLC is that it is an open system. Samples are evaluated as a whole, whether they remain on the origin or ride with the solvent front. This is not always the case in HPLC.

Thin-layer chromatography plays an essential role in the early stage of drug development when knowledge about the impurities and degradants in drug substance and drug product is limited. It is often used as an orthogonal technique to HPLC to ensure quality of the pharmaceuticals. TLC has been established in almost all pharmacopoeias worldwide. The first stand-alone analytical method by TLC was described in the European Pharmacopoeia of 1974 (Ph. Eur. 1), in which TLC was specified for the identification of 23 drugs.

Figure 1 shows how TLC is involved at different stages of drug development. After a new drug entity has emerged, HPLC methods are

9 ISOLATION METHODS l:THIN-LAYERCHROMATOGRAPHY 207

New Drug Candidate Nominated

Drug Substance Method Development

Drug Substance Release Testing

Drug Substance Stability Testing

Drug Substance Method Validation

Drug Product Method DevelopmentA/^alidation

Drug Product Release Testing

Impurity Isolation and Identification if Needed

Drug Product Stability Testing

Impurity Monitored by TLC or HPLC

m FIGURE I Different stages where TLC is involved in drug development.

needed for potency and purity analysis. At the same time, an orthogonal TLC method is also developed and validated for drug substance and drug product. If there is an unknow^n impurity observed by TLC during release testing, isolation and characterization vs ould be required to identify the impurity. If HPLC is unable to quantify the impurity, TLC v ill remain as the primary testing method to quantitate this particular impurity until an alternative method w ith more accuracy is developed.

In this chapter the applications of TLC in a pharmaceutical environment such as method development and validation for drug substance and drug product, and impurity isolation and characterization by TLC are discussed.

III. TLC METHOD DEVELOPMENT AND VALIDATION

Method development of TLC has been thoroughly discussed in the recent books by Elke Hahn-Deinstrop, and Bernard Fried and Joseph Sherma.^^'^^ The most recent review^ article by Colin Poole and Neil Dias also provided certain guidance on method development in TLC.^^ While the focus of this chapter is not on TLC method development, there are a few points that need to be stressed regarding method development from a regulatory perspective. First and foremost, use of qualified/calibrated equipment is part of the GMP compliance. In other w^ords, all instruments must have updated Instrument Qualification (IQ), Operational QuaHfication (OQ), and Performance Qualification (PQ) according to the company's SOP (Standard Operation

2 0 8 p. M. GORMAN AND H. JIANG

Procedure). The lab environment in which testing is to be conducted should also be GMP compliant. Last but not least, lab personnel must have gone through regular GMP training in order to document their ability to work effectively in a GMP facility.

While densitometric detection is becoming more popular for quantitation in TLC, the work presented in this chapter is based on visual detection that can be routinely used as an acceptable alternative for semiquantitative purposes.

A. Drug Substance Method Development and Validation

Once a new drug candidate has been identified, a TLC method is needed as an orthogonal tool with respect to HPLC methods. Similar to HPLC method development, certain criteria must be met:

1. Demonstration of adequate resolution and selectivity between drug substance and known or unknown impurities.

2. Determination of specified impurities at an appropriate level (for example: LOD (limit of detection) ==0.05%, LOQ (limit of quantitation) = 0.1%).

3. Determination of drug substance at an appropriate level (for example, LOD = 0.05%, LOQ = 0.1%) if it is to be used for quantitation of impurities.

4. Adjustment of R;^~0.50 for drug substance. 5. Demonstration of a stability-indicating assay through purposeful

degradation studies. 6. Demonstration of ruggedness and reproducibility.

The International Conference on Harmonisation (ICH) guideline outlines the following four types of analytical procedures to be vaUdated: (1) identification tests; (2) quantitative tests for impurities content; (3) limit tests for the control of impurities; (4) quantitative tests of the active moiety in samples of drug substance or drug product or other selected component(s) in the drug product. In the pharmaceutical industry, TLC testing falls between categories (2) and (3). The characteristics to be validated are specificity, detection limit, quantitation limit, and low-level linearity. TLC method development and validation can be performed rapidly in accordance with the current guidelines.^^ However, the example of a validation procedure described here is based on semiquantitation by visual examination. All the examples in this chapter utilize HPTLC plates.

I. Specificity

The purpose of this test is to demonstrate that immediate precursors and other known impurities are resolved from the drug substance and do not interfere with any unknown impurities existing in the drug substance bulk. All impurities should also be separated from any TLC-specified impurities. Figure 2 shows a drug substance band separated from all known impurities. The nominal concentration of the compound is determined in the method development phase according to the requirement of LOD and LOQ.

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY 209

• ^" ' . , , - . / ; -^': , ; ; - . i^-j

SEPARATION OF PRECURSORS 254 DETECTION . A|3|»iicatio$i - 6nim, 5ul

L ' mMi& Phase « KtOAc: MeOH: NH40H (35:12:1)

FIGURE 2 Specificity.

The resulting Rf values are recorded for the drug substance and each impurity that is evaluated. It should be emphasized that all impurities show adequate resolution from the drug substance band and TLC-specified impurities.

2. LOD/LOQ/Low-Level Linearity for Drug Substance

This is done to demonstrate the low level linearity of drug substance, which is used to quantitate impurities at certain levels. In most cases, quantitation of an impurity at 0.1% is required. It is imperative that the validation data demonstrate a sufficient LOD and LOQ. An example in Figure 3 shows the LOD at 0.05% and LOQ at 0.1% for a drug substance. Here, a nominal concentration of drug substance at 25.0 mg/ml and a serial dilution of this solution to 0.5, 0.4, 0.3, 0.2, 0.1, and 0.05% relative to nominal concentration are made to demonstrate LOD and LOQ, as well as linearity. It has to be mentioned that sometimes the shape of drug substance band is somewhat compromised in order to fulfill the requirement of LOD and LOQ.

3. LOD/LOQ/Low-Level Linearity for Specified Impurities in Drug Substance

The LOD and LOQ are determined for the specified impurity based on ICH Impurity Guidelines and/or regulatory requirements. It is demonstrated in the same manner as described in the section above. An example of a specified impurity (due to its carcinogenicity) is given in Figure 4.

210 p. M. GORMAN AND H. JIANG

F I G U R E 3 LOD/LOQ/low-level linearity for drug substance.

i Method Validation Staroh-Kl

LOD/LOQ/Linearity for Spedfted \mpurk^

26 0.06% 0.1% 0.2% 0.3% 0.4% 0.5% mg/mt

M& ^tt^ ^Mfe ^iitt' •^"^^^ -~^P^^ ^IgPIR- l lUpi lP' '

^## .-'

F I G U R E 4 LOD/LOQ/low-level linearity for aTLC-specified impurity.

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY 211

4. Stability on Plate/Stability in Solution

The intent of this test is to find any degradation products that arise from sample solution and sample on the TLC plate over a period of time. In our lab, these studies are performed by weighing 5 samples and labeling them as 0, 1, 2, 3, and 4 hours.

• At time zero, solvent is added to the 4-h sample and is immediately applied to the plate for the time on plate studies. It is identified as the 4-h sample, and the solution is set aside to complete the stabiHty-in-solution study later.

• An hour later, the second sample is prepared and applied next to the 4-h sample for the stability-on-plate study, marked as the 3-h sample, and the solution is set aside.

• Two hours later, the third sample is prepared and applied next to the other two samples for the stability-on-plate study, marked as the 2-h sample, and the solution is set aside.

• Three hours later, the fourth sample is prepared and applied next to the other three samples for the stability-on-plate study, marked as the 1-h sample, and the solution is set aside.

• Four hours later, the final sample is prepared and applied next to the other four samples for the stability-on-plate study, marked as the 0-h sample, and the solution is set aside. This plate can now be developed as the stability-on-plate experiment.

• Immediately after the final solution is prepared, the five solutions are applied to a new plate that is developed immediately. This is for the stability-in-solution study.

Once developed, the plates are examined to determine if any degradation has occurred during the time course, either on plate or in solution. Figure 5 presents an example of stability-on-plate validation in which two new bands show up after 1 hour on the plate. If degradation is observed in the first time point (1 hour), a subsequent experiment can be conducted as described above, but with time points as 60, 45, 30, 15, and 0 minutes. Sometimes an existing band in the drug substance will grow in intensity with time, or a new band appears with time. Both are indicative of some type of degradation occurring. If evidence of degradation in either experiment exists, it should be specified in the test procedure.

5. Post-Development DryingTime

This test is conducted to determine the proper amount of time a TLC plate should be dried after development. It is often found that a range of drying time is acceptable. However, a plate that has not dried long enough sometimes shows an adverse effect when sprayed by a derivatizing agent.

If a derivatizing agent is being employed as the detection method, four identical plates are prepared with the following concentrations of drug substance: 0.05, 0.1, 0.2, 0.3, 0.4, 0.5%, and nominal concentration. After development, the first plate is dried for 60min in a room-temperature, air-circulating oven (or equivalent), and a photograph is taken. The remaining

212 p. M. GORMAN AND H. JIANG

F I G U R E 5 Stability on plate.

three plates are dried for 45 minutes, 30 minutes, and 15 minutes respectively. Each plate is photographed to document potential changes as a result of inadequate drying time. The appropriate time range of drying time is determined based on the results.

6. As Is Plate/Prewashed Plate/Prewashed Plate 24 h ago

The intent of this study is to demonstrate the effects of using a TLC plate straight from the box (as is), and then comparing this to one that has been prewashed in methanol, dried, and used the same day, as well as one that was left in a dry, covered TLC tank, and used 24 hours later. Since prewashing plates takes time, if there is no advantage to prewashing the plate, this step can be eliminated once the validation work is complete. By the same token, if several TLC plates were prewashed and dried, then set aside in a dry, sealed TLC tank, and used the next day, this can also be utilized as a time-saving step.

An example is shown in Figure 6. An HPTLC plate is prewashed by adding 10 ml methanol to a 10 x 10 cm developing chamber, inserting a HPTLC plate, and developing to approximately one inch from the top of the plate. After development, allow the plate to dry in a 100°C air-circulating oven (or equivalent) for 20 minutes, then cool to ambient temperature prior to use. This plate is set-aside in a clean dry tank for 24 hours. The same procedure is repeated for a new plate, 24 hours later, such that one will be prewashed and dried for 24 hours, and one will be freshly prewashed and dried. A third plate is taken from the box and untreated (used as is). All three plates are used to develop the following

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY 213

2S«i$W ^S^ 04% a3^» O.^^ 0.1 r. t>,««%

12$*mt 0sm,tk»ts

:^?#st?^^^!^

F I G U R E 6 As Is plate (top)/prewashed plate (bottom left)/prewashed plate 24 hours ago (bottom right).

sample solution of drug substance: 0.05, 0.1, 0.2, 0.3, 0.4, 0.5%, and nominal concentration. Plates are developed and dried, and observations are recorded. The example in Figure 6 shoves the as is plate gives a streak belov^ the solvent front (might interfere v^ith any potential impurities) while the other tv^o prewashed plates provide a clean appearance. Therefore, the prev^ashed plate is preferred.

Plates are usually prewashed and used the same day to avoid any undesirable effects. If densitometric detection is to be utilized, it is a necessity to prewash TLC plates with solvent (usually methanol), to eliminate background noise that is usually acquired from the packing material that TLC plates are stored in.

7. Saturated Chamber/Unsaturated Chamber

The purpose of this study is to determine what effect, if any, is seen when a saturated environment is utilized in comparison to an unsaturated environment.

As in the example (Figure 7), two plates are prepared with the following sample solutions of drug substance: 0.05, 0.1, 0.2, 0.3, 0.4, 0.5%, and nominal concentration. A developing chamber is prepared by placing blotter paper in one trough, covering this blotter paper with 5.0 ml of developing solvent, and adding another 5.0 ml of developing solvent to the other trough, then immediately placing the HPTLC plate to be developed. This would

214 p. M. GORMAN AND H. JIANG

;- '1.%vr'i f«

t

FIGURE 7 Saturated chamber (left)/unsaturated chamber (right).

FIGURE 8 Blotter paper (left)/no blotter paper (right).

constitute the unsaturated chamber. Another developing chamber is prepared in the identical manner except the lid is then replaced on the developing chamber, and the tank is allow^ed to equiUbrate for 10-15 min. This constitutes the saturated chamber. Once the equilibration time has been reached, the HPTLC plate is developed. The example in Figure 7 show^s that the saturated chamber gives a less tailed drug substance band than the unsaturated one.

Generally speaking, the saturated environment allow^s a more consistent environment for TLC plate development. It also shows less development time compared with an unsaturated environment. Therefore, the saturated tank is usually employed in TLC. However, there are also situations where an unsaturated chamber is the preferred choice for plate development.

8. Blotter Paper/No Blotter Paper

The intent of this experiment is similar to the previous experiment involving saturated and unsaturated tanks. TLC tanks are usually lined with appropriate-sized blotter paper to aid in saturation of the TLC tank.

An example is shown in Figure 8. Two plates are prepared with the following sample solutions of drug substance: 0.05, 0.1, 0.2, 0.3, 0.4, 0.5%, and nominal concentration. A developing chamber is prepared by placing blotter paper in one trough, covering this blotter paper with 5.0 ml of

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY 215

developing solvent, and adding another 5.0 ml of developing solvent to the other trough, then the tank is allowed to equilibrate 10-15 min before the HPTLC plate is developed. This would constitute the chamber using blotter paper. Another developing chamber is prepared in the identical manner, except that the developing solvent is added to an empty chamber, containing no blotter paper. The lid is then replaced on the developing chamber and the tank is allowed to equilibrate for 10-15 minutes. This constitutes the chamber using no blotter paper. Once the equilibration time has passed, the HPTLC plate is developed. The example presented in Figure 8 shows that development without blotter paper not only pushes the drug substance band toward the solvent front, but also gives poorer linearity than the one with blotter paper.

In general, blotter paper is usually used in the TLC tank unless there are adverse reasons that appear in the validation studies to utilize an unlined TLC tank. However, there are also situations where a chamber without blotter paper is the preferred choice for plate development.

9. Spraying/Dipping

Traditionally, TLC plates have been sprayed with derivatizing agents, without much consistency because of operator dependence. In addition the technique is difficult to master. Manufacturers of TLC equipment have now developed an immersion device that allows the TLC plate to be dipped into a pool of derivatizing agent, and a timer can be adjusted for the length of exposure, as well as the speed. This has now made spraying plates a less attractive option than dipping plates. However, the dipping chamber cannot be used for some derivatizing agents.

Results are compared and a decision made whether dipping the plate is an acceptable alternative to spraying the plate for the given derivatizing agent. Since the immersion device is more reproducible and there are fewer hazards associated with dipping plates than spraying plates, dipping plates is the preferred choice.

10. Two-Dimenslonal Study

This test is done to rule out any TLC artifacts that can occur from the sample applied to the origin not moving at the same rate as the drug substance upon the first pass in the mobile phase. It is often utilized to rule out TLC procedure-related bands, as described later in this chapter (Impurity Isolation and Characterization by TLC). A plate is spotted, not streaked, with an appropriate amount of sample at a nominal concentration (typically 10-15 )iL). This is done in several applications of about 3-5 units each to prevent diffusion. It is spotted on the lower left corner, at the origin, and this lane is labeled as first pass. The plate is developed in the normal fashion. Once dried, the plate is rotated 90° counterclockwise, another spot is applied in the lower left corner, labeled as second pass, and developed in the same manner. The result shows any TLC artifacts present in the second pass.

Figure 9 shows such an example. A band observed that is initially at the origin in the first development has the same Rf value as the drug substance in the second development. It is labeled as a TLC procedure-related band.

216 p. M. GORMAN AND H. JIANG

4dhod Validatii>n -Dimensional Test Dragcn<Jt>rff s Spray

FIGURE 9 Two-dimensional study.

11. Ruggedness/Robustness

It is considered good practice to have 2 different individuals and, if possible, 2 different laboratories run the method to ensure consistency and reproducibility as part of the validation.

12. Validation Summary

A summary page is generated upon completion of the method validation. This gives the user a quick reference of w^here to find the data and the conclusions of each experiment. It also alerts the analyst to any potential issues, such as degradation. Table 1 is an example of validation summary for drug substance.

B. Drug Product Method Development and Validation

After the drug substance TLC method has been developed and validated, method development for drug product is relatively straightforw^ard unless the dose strength is very lov; (for example, 1.0 mg or less per tablet), and/or the formulation is somewhat unusual (lotion, lyophile, etc.), and/or a suitable extraction solvent cannot be readily found. The drug substance method is a logical starting point for the drug product method development. It is common that the same method can be used by modifying the extraction process, and most of the validation w^ork from drug substance can also be utilized under appropriate justification. As a result, the validation protocol is shortened to reflect the work already conducted for drug substance.

9 ISOLATION METHODS l:THIN-LAYERCHROMATOGRAPHY 21/

T A B L E I Summary of TLC Method Vaiidation for Drug Substance

Experiment

Specificity

LOD/LOQ/low-level linearity of drug substance

LOD/LOQ/low-level linearity of specified Impurity

Stability on plate/stability in solution

Post-development drying time

As is/prewashed/prewashed 24 hours ago plates

Saturated chamber/ unsaturated chamber

Blotter paper/no blotter paper

Spraying/dipping plates

Tw^o-dimensional

Notebook reference

#12345 pp. 11-15

#12345 pp. 16-19

#12345 pp. 20-22

#12345 pp. 23-27

#12345 pp. 28-31

#12345 pp. 31-35

#12345 pp. 36-39

#12345 pp. 40-42

#12345 pp. 43-45

#12345 pp. 46-48

Results

Precursors are resolved from drug substance

Precursor 1

Precursor 2

Drug substance

Impurity 3

LOD = 0.05% LOQ = 0 .1%

Rf= 0.18

Rf= 0.33

Rf= 0.52

Rf =0.7S

LOD = 0.05% LOQ = 0 .1%

Drug substance is stable in solution and on the silica plate up to 1 hour.

Recommended drying time is 0-60 min in a room-temperature oven.

Prewashed plates are recommended.

Plate developed in a saturated chamber is more consistent than the one in an unsaturated chamber.

Plate developed in tank lined with blotter paper is more consistent than the one without blotter paper.

Plate dipped with Dragendorff s spray yields similar result to the one sprayed with the same detection reagent.

The origin band is a TLC procedure-related band.

Any given drug product usually has a number of strengths associated with it. They can be evaluated by bracketing, i.e., studying the highest and the lowest strengths.

A typical validation protocol for a drug product method is discussed in detail next.

I. Separation of Drug Substance from its Excipients

The purpose of this test is to ensure that the excipients are resolved from the drug substance, and do not interfere with any other related known or unknown impurities.

The first thing is to establish a proper extraction solvent and a method to extract the active ingredient from the formulation. It is very Hkely that by now the solubility information of the drug product is available. The process may involve a series of steps to ensure full extraction from the

218 P- M. GORMAN AND H. JIANG

F I G U R E 10 Separation of drug substance from its excipients.

excipients. Here is an example of how to extract the actives from tablets: Wrap a tablet in a piece of weighing paper, crush with a hammer, and quantitatively transfer all contents to a suitable container. To this container, add an appropriate amount of extraction solvent (methanol) and stir for 10 minutes. Sonicate the solution for 10 minutes, transfer to a centrifuge tube, centrifuge at 3000 rpm for 10-20 minutes, and filter the final solution as needed.

To demonstrate that no excipients interfere with the drug substance, a TLC plate is prepared like the one on the left in Figure 10, where the drug substance is extracted from its formulation, compared with that of a placebo (a synthetic blend is prepared according to the formulation order if no placebo is available at the time of method development and validation). The in-going bulk lot of drug substance is weighed at the proper concen­tration, diluted in the extraction solvent, and applied to the same TLC plate for analyses.

Should any excipients in the placebo interfere with the drug substance, another TLC plate will be prepared containing each individual excipient extracted in the same manner and the same concentration as in the placebo. As an example shown in Figure 10 on the right, excipient no. 6 is responsible for the dark band above the drug substance band. Once this has been identified, it will be referred to in the method as a known excipient. Excipients, placebo, tablet, and drug substance should all appear consis­tent with the same intensity since they are all prepared at the same concentration.

2. LOD/LOQ/Low-Level Linearity for Drug Product

As indicated in the validation section for drug substance, the purpose of this test is to demonstrate that low-level linearity can be routinely achieved for drug product evaluation.

If the extraction solvent for the drug product is the same as the dissolving solvent in the drug substance method, there is no need to repeat this validation test. However, if the extraction solvent has changed, this validation test must be repeated to demonstrate desired LOD and LOQ. In general, it is expected that an LOD = 0.05% and an LOQ = 0.1% would

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY 219

'$fk

•u- I -y^f^ -mi^' ,m^'

6 ^ i

F I G U R E 11 LOD/LOQ/low-level linearity for drug product.

be reproducibly demonstrated by the method, as the example shown in Figure 11.

3. LOD/LOQ/Low-Level Linearity for Specified Impurities in Drug Product (Spiking Experiment)

When TLC is used to track a specified impurity in the drug substance, it is Hkely to be apphed to the subsequent drug product method. A plate must be prepared to show that required levels of LOD and LOQ can be achieved for the specified impurity in the presence of excipients. This is known as a spiking experiment. The specified impurity is prepared and spiked into the vial of the drug product at different levels, then applied to the TLC plate to ensure it can be extracted to meet the requirement of LOD and LOQ. If no specified impurities are being tracked by TLC, this validation test can be excluded.

4. Stability on Plate/Stability in Solution

If the same solvent is used for both the drug product method and drug substance method, there is no need to repeat this validation test unless there is a concern that any of the excipients might cause any instability of the drug substance. However, if the extraction solvent has changed from the drug substance dissolving solvent, this needs to be validated in the same manner as that of the drug substance discussed earlier. It is always a good

2 2 0 p. M. GORMAN AND H. JIANG

practice that when all sample preparations are made, the solutions are immediately applied to the TLC plates. This test is conducted to alert the analyst for any potential degradation issues resulting from the extraction solution or the TLC plate.

5. Recovery Plate

The final validation test is done to demonstrate that no drug substance was lost from the extraction process, as this is a purity test. In a typical extraction process, as described above, the sample is crushed, extraction solvent is added, the sample is then stirred, sonicated, centrifuged, and filtered. In some cases, the drug substance will adhere to the filter. In this event, a study of numerous types of filters ensues to find one that yields 100% recovery.

In the example shown in Figure 12, the drug substance is weighed, and solutions of 0.5, 0.3, and 0.1% of nominal concentration are made and applied to one half of the TLC plate as the control solutions. In addition, the same dilutions are made from one extracted tablet and applied to the same plate. Once the plate is developed, the two 0 .1% standards are compared to each other to determine if they have the same visual response in intensity. The same comparison is made for the 0.3% and 0.5% standards. If all three standards appear to have the same intensity as their respective controls, then we state that the tablet has been extracted completely. If the standards are not the same in intensity and it is apparent that some drug substance is being retained by the filter, then another type of filter or possibly centrifugation alone would be examined until full extraction can be achieved.

FIGURE 12 Recovery on plate.

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY

m T A B L E 2 Summary of TLC Method Validation for Drug Product

221

Experiment Notebook reference Results

Separation of drug substance from excipients

LOD/LOQ/low-level linearity of drug substance

#12345 pp. 55-58

#12345 pp. 59-62

No excipients interfere with the drug substance

LOD = 0.05%

LOQ = 0.1%

LOD/LOQ/low-level linearity of specified impurity (spiking experiment)

Stability on plate/stability in solution

Recovery plate

#12345 pp. 63-65

#12345 pp. 66-72

#12345 pp. 73-77

LOD = 0.05%

LOQ = 0.1%

Drug substance is stable in solution and on the plate for 4 hours.

Drug substance is fully recovered from this extraction process

6. Validation Summary

A summary page is generated upon completion of the tablet method validation. This gives the user a quick reference of where to find the data, and the conclusions of each experiment. It also alerts the analyst to any potential degradation. An example of a method validation summary for drug product can be found in Table 2.

lY. IMPURITY ISOLATION AND CHARACTERIZATION BY TLC

A validated TLC method is used to evaluate drug substance or drug product for release or stability purposes. Usually there are three kinds of impurities observed from a TLC analysis: TLC-specified impurities, known impurities, and unknown impurities.

A. TLC-Specified Impurities

As mentioned earlier, some of the synthetic starting materials, intermediates, precursors, and known potential degradants cannot be detected by a UV detector that is routinely used in HPLC analysis because of lack of a chromophore, or it may elute at the solvent front or elute very late in HPLC run. Alternatively, TLC is used to quantify these impurities. An example of these unique situations would be one in which guanidine hydrochloride (Figure 13), a potential degradation product, needs to be monitored in one of the drug candidates.

By looking at the structure of guanidine, it is obvious that regular UV detection would not be suitable to evaluate this impurity. Since it is a potential degradant, TLC was selected because of its simplicity in

222 p. M. GORMAN AND H. JIANG

sample preparation among other reasons. The detection method chosen for this compound was starch-potassium iodide because of its sensitivity to amines. In this case, levels of detection down to 0 .1% are readily achieved (Figure 13).

B. Known Impurities

The knowledge and availability of known impurities (process-related impurities and known degradation products) vary depending on the stage of development of the drug substance/drug product. As part of the method validation, the Rf values of synthetic starting material, intermediates, and precursors, as well as known degradants are also assessed. Impurities observed during the evaluation of a drug substance/drug product are first checked against the validation data to see if any of them match the known impurities. If so, further TLC analysis with both sample and known impurities are run in parallel to determine if the impurity in the sample is indeed the known impurity. In the example shown in Figure 14, an impurity is observed during a stability study that is later confirmed by TLC as a known impurity.

H N = < ' HCI NH.

2

FIGURE 13 Quantitation of guanidine liydrochloride by TLC.

M • • ^ •

5^*1; ^Sft' sW* «%*• V!R

« 'mU «.tt» 0.2% 0.a% ^A\ OS*w

" • wmfm iPijI MHIP ^«^r' *'' ^ ^ ^ ^ ^ ^ ^ ^^^^^^^^^^ .as*™. ,^

mm ^;m mm € P

FIGURE 14 Confirmation of a l<nown impurity by TLC (left—an impurity observed during a stability study; right—confirmed by TLC as a known impurity).

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY 2 2 3

C. Unknown Impurities

Often the impurities observed from TLC analysis are unknown. In this case, identification of the impurities may become necessary, depending on the level of impurities (ICH Impurity Guideline). Questions one should first ask are. Is the impurity real or an artifact? Is it salt-related? If the answers are No to these two questions, isolation and characterization are required to identify the impurity.

1. Is the Impurity Real or an Artifact?

In analyzing samples in a cGMP environment, numerous things should be ruled out in impurity identification, such as equipment malfunction, glassware contamination, etc. A thorough investigation should be conducted to ensure whether the impurity is real or an artifact.

Tailing of the drug substance band in TLC often occurs with acidic and basic compounds. This can be minimized by suitable choice of pH or by the use of inorganic or organic salts during method development. However, during release testing a 2-D TLC experiment can confirm if a band at the origin is drug substance residue or a real impurity. This test is done in the same manner as described in the validation of drug substance section. The result will yield a diagonal line for the first sample applied. If the origin band from first development rises to the same Rf as the drug substance as shown in Figure 9, this indicates the origin band is indeed drug substance, i.e., it is a TLC procedure-related band, or an artifact. If the origin band does not move at the same distance as the drug substance on the second development as the example shown in Figure 15, further investigation is required to identify this impurity.

2. Is the Impurity Salt-Related?

Often drug substance exists as a salt to improve its solubility and bioavailability. The salt form changes sometimes during drug development. It is important to be aware of the impact of the type of salt on the TLC analysis result. An example in Figure 16 shows that Rf of the corresponding salt matches that of the impurity band in a drug substance lot.

In some situations, the salt, when bound to the drug substance, can have a different Rf value than when the salt is appHed to the plate itself. In the example shown in Figure 17, a spike experiment is conducted to show that the salt's Rf value changes when in the presence of the drug substance. Therefore, the impurity band is identified as salt related.

It is also important to mention that the salt does not always remain at the origin, as seen in the example shown in Figure 17. Salts can migrate anywhere on a TLC plate, depending on the salt itself and the mobile phase used as demonstrated in this example.

3. TLC Impurity Isolation and Characterization

Once it has been determined that an impurity is real, i.e., it cannot be called a TLC artifact, the related salt, a known process-related impurity, or a degradation product, then it has to be isolated and correlated to the relevant

224 p. M. GORMAN AND H. JIANG

F I G U R E 15 2-D TLC to con f i rm that the origin band is an impur i ty .

F I G U R E 16 Salt conf i rmat ion by TLC.

HPLC system. There are two approaches to the isolation and correlation of bands seen in TLC assays: (1) Isolate impurity band from TLC -^ reassay on HPLC system; (2) isolate impurity from HPLC -^ reassay on TLC method. The isolation and characterization by TLC is often required at the early stages of drug substance synthesis when a synthetic route is still below optimization.

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY 2 2 5

F I G U R E 17 R^ of salt changes w i t h drug substance.

In the first situation, the band in question that is viewed in the TLC system would be isolated, extracted, reappUed for confirmation, and analyzed by HPLC for correlation and confirmation. The following example gives a detailed description of how this is done in our laboratory.

To isolate the band in question, an 80-mm streak of the drug substance, typically at 25.0mg/ml concentration, is applied to a 10 x 10 cm silica plate, usually at a volume of 200 jil, in an effort to overload the silica plate such that more of the impurity can be isolated. This process is repeated for 10 HPTLC plates or so. The plates are then developed according to the method and the impurity marked with a pencil under UV 254 nm if this is the method of detection. An example of an LP (less polar relative to drug substance) impurity and an MP impurity (more polar relative to drug substance) to be isolated is shown in Figure 18. Once the 10 plates have dried, the bands in question are scraped from the silica plate using a metal spatula and collected in separate containers. Approximately 5-10 ml of extraction solvent (usually the sample solvent in the relevant TLC method) is added to the container. The solution is stirred for about lOmin, sonicated about lOmin, and centrifuged at 2500 rpm for 10 minutes. The resulting clear solution is transferred to another suitable container and concentrated under a stream of nitrogen to approximately 0.5 ml of volume. This concentrated solution is then reappUed to a new TLC plate to confirm the TLC isolation of the impurities. The remainder of the solution is then evaporated to dryness under nitrogen. HPLC sample solvent is used to reconstitute the isolated impurity that is analyzed by a respective HPLC method to determine if the impurity correlates to any impurity previously seen by HPLC analysis. In this case, both LP and MP bands correlate with the impurities in HPLC analysis. Therefore, these two impurities will be monitored by HPLC. In addition, the

226 p. M. GORMAN AND H. JIANG

mm

F I G U R E 18 Scaleup fo rTLC -> HPLC correlat ion (top, scale-up isolation; bo t t om left , scraping of innpurity bands; b o t t o m r ight , conf i rmat ion of isolated impuri t ies).

isolated impurity can be sent for MS and NMR analysis to determine the structure as needed. A blank TLC plate is prepared in the identical manner without drug substance. Silica from this plate is scraped off and extracted by the same procedure as the impurity bands. The blank sample is also analyzed by HPLC to rule out any anomalies from extraction process.

This impurity isolation process (TLC —> HPLC) is commonly utilized in our laboratory. However, the process is tedious and lengthy for extraction from the silica plates, and timing must be coordinated such that the sample is immediately analyzed by the HPLC system to rule out any potential degradation. Sometimes, the band in question is very close to the drug substance or another band, making it difficult to extract for TLC -^ HPLC. An alternative is to correlate an impurity by collecting the impurity from HPLC and reapplying it to the TLC system (HPLC -^ TLC).

It is worth mentioning that whenever it is possible, scale-up is always helpful to collect impurities for identification purposes. This also applies to HPLC -> TLC correlation studies. A preparative column or analytical column can be used to collect these samples with a fraction collector. Such an example is shown in Figure 19. Two impurities were observed during TLC testing. A correlation experiment is needed to confirm if these two impurities

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY

aiOO:

0.090:

0.080:

0.070:

O.O6O;

3 0.050;

0.040i

0.030:

0.020;

0.010;

0.000:

227

4.00 6.00 10.00

l\iiutes 1600 18.00

F I G U R E 19 HPLC -^ TLC correlation (top, HPLC purity profile; bottom,TLC correlation of impurities separated from HPLC).

were identical to the ones observed from HPLC analysis. This time an HPLC -^ TLC correlation is designed since the LP band is right above the drug substance band, which made it difficult to scale up and scrape the impurity on TLC plates. The two impurity peaks were collected from HPLC runs, and subsequent TLC analysis confirmed that they were the same two impurities.

D. Summary of Impurity Isolation and Characterization by TLC

Figure 20 summarizes how impurity isolation and characterization is performed by TLC. If the correlation experiment shows an impurity match between TLC and HPLC, usually HPLC is used to monitor this impurity. On the other hand, if no match is found, TLC will continue to be the primary quantification method for this particular impurity.

In summary, TLC analysis plays a fundamental role in the drug development process. It is involved in release testing and stability study of

228 p. M. GORMAN AND H. JIANG

Impurity found during TLC evaluation

Is it a TLC-specified impurity?

No

Is it a known impurity?

No

Is it a TLC procedure-related impurity?

No

Is it salt related?

i No

Correlate between TLC and HPLC

1 Monitor by TLC or HPLC

Yes

Yes

Yes

Yes

Quantify by TLC

Quantify by appropriate method

Conduct 2-D TLC to confirm

Check with corresponding salt

Identify by MS, NMR if necessary

F I G U R E 20 Summary of TLC impurity isolation and characterization.

identity and purity of active pharmaceutical ingredient (API) in both drug substance and drug product. Especially at early stages of drug development, TLC's complementary properties to HPLC make it an essential part of the quahty control for pharmaceuticals goods.

lEFERENCES

1. Abidi, S. L. Separation procedures for phosphatidylserines. / . Chromatogr. B. 717(1-2): 279-293, 1998.

2. Lawton, L. A. and Edwards, C. Purification of Microcystins.]. Chromatogr. 912(2): 191-209, 2001.

9 ISOLATION METHODS I: THIN-LAYER CHROMATOGRAPHY 2 2 9

3. Bhushan, R. and Martens, J. Separation of amino acids, their derivatives and enantiomers by impregnated TLC. Biomed. Chromatogr. 15(3):155-65, 2001.

4. Reddy, M. V. Methods for testing compounds for DNA adduct formation. Regulatory Toxicology & Pharmacology. 32(3):256-63, 2000.

5. Alemany, G. et al., Thin-layer chromatographic determination of catecholamines, 5-hydroxytryptamine, and their metabolites in biological samples—a review. / . AOAC Int. 82(l):17-24, 1999.

6. Roda, A., Piazza, F. and Baraldini, M. Separation techniques for bile salts analysis. / . Chromatogr. B. 717(l-2):263-78, 1998.

7. Porter, J. K. Analysis of endophyte toxins: fescue and other grasses toxic to livestock. / . Animal Sci. 73(3):871-80, 1995.

8. Oka, H., Ito, Y. and Matsumoto, H. Chromatographic analysis of tetracycline antibiotics in foods. J. Chromatogr. 882(l-2):109-33, 2000.

9. Sherma, J. Thin-layer chromatography in food and agricultural analysis. / . Chromatogr. 880(l-2):129-47, 2000.

10. Muthing, J. High resolution thin-layer chromatography of ganghosides. / . Chromatogr. A. 720(l-2):3-25, 1996.

11. Bereznitski, Y., et al. Thin-layer chromatography—a useful technique for the separation of enantiomers. / . AOAC Int. 84(4):1242-1251, 2001.

12. Poole, C. F. Planar chromatography at the turn of the century. / . Chromatogr. A. 856(1-2): 399-427, 1999.

13. Stahl, E. The historical development of the method. In Thin layer Chrotnatography (E. Stahl, Ed.) pp. 1-6, Springer-Verlag, Berlin, 1969.

14. Pelick, N., Bolliger, H. R. and Mangold, H. K. The Flistory of Thin-layer Chromatography, In Advances in Chromatography (J.C. Giddings and R.A. Keller, Eds.), Marcel Dekker, NY, pp. 85-118, 1966.

15. Stahl, E. Thin-Layer Chromatography., Springer-Verlag, NY, 1969. 16. Heftmann, E. History of Chromatography. In Chromatography—A Laboratory Handbook

of Chromatographic and Electrophoretic Methods (E. Heftmann, Ed.), pp. 1-13. Van Nostrand Reinhold, NY, 1975.

17. Kirchner, J. G. History of TLC. In Thin Layer Chrotnatography—Quantitative Environmental and Clinical Applications (Touchstone, J.C. and Rogers, D., Eds.), pp. 1-6. WileyTnterscience, NY, 1980.

18. Jork, H. and Wimmer, H. Thin-layer Chromatography—History and Introduction. In TLC Report a Collection of Papers. GIT Verlag, Darmstadt 1986.

19. Wintermeyer, U. The Root of Chromatography: Historical Outline of the Beginning to Thin Layer Chromatography, GIT Verlag, Darmstadt, 1989.

20. Berezkin, V. The discovery of thin layer chromatography. / . Planar Chromatography.-Med. TLC. 8:401-405, 1995.

21. Jork, H., et al. Thin-Layer Chromatography—Reagents and Detection Methods. Vol. la: Physical and Chemical Detection Methods: Fundamentals, Reagents /, Weinheim: VCH, pp. 464, 1990.

22. Jork, H., et al. Thin-Layer Chromatography—Reagents and Detection Methods. Vol. lb: Physical and Chemical Detection Methods: Activation Reactions, Reagent Sequences, Reagents II, Weinheim, VCH, pp. 496, 1994.

23. Borman, S.A. HPTLC: taking off. Anal. Chem. 54:790A-4A, 1982. 24. Coddens, M. E., et al. Quantitation in high performance TLC. LC Magazine, Liquid

Chromatography and HPLC. 1:282-289, 1983. 25. Costanzo, S. J. High performance thin-layer chromatography. / . Chetn. Educ. 61:

1015-1018, 1984. 26. Fenimore, D. C. and Davis, C. M. High performance thin-layer chromatography. Anal.

Chem. 53(252A-266A), 1981. 27. Geiss, F. Fundamentals of Thin Layer Chromatography, Alfred Huthig Verlag, Heidelberg,

pp. 5-8, 1987. 28. Maugh, T.H. IL TLC: the overlooked alternative. Science 216:161-163, 1982. 29. Sherma, J. Comparison of thin layer chromatography with liquid chromatography. / . Assoc.

Off. Anal. Chem. 74:435-437, 1991.

230 p. M. GORMAN AND H. JIANG

30. Fried, B. and Sherma, J. Thin-Layer Chromatography. In: Chromatographic Science Series, Vol. 81 (J. Cazes, Ed.), Marcel Dekker, NY, Inc. pp. 499, 1999.

31. Hahn-Deinstrop, E. Applied Thin-Layer Chromatography. Wiley-VCH, Weinheim, pp. 304, 2000.

32. Poole, C. F. and Dias, N. C. Practitioner's guide to method development in thin-layer chromatography. / . Chromatogr. A. 892(1-2):123-142, 2000.

33. Renger, B. Contemporary thin-layer chromatography in pharmaceutical quality control. / . AOAC Int., 81(2):333-339, 1998.