WO2009061326A1

WO2009061326A1 - Evaluation of chromatographic materials

Info

Publication number: WO2009061326A1
Application number: PCT/US2007/084344
Authority: WO
Inventors: Zheng Yang
Original assignee: Wyeth
Priority date: 2007-11-09
Filing date: 2007-11-09
Publication date: 2009-05-14

Abstract

Quick, non-destructive NIR methods and reliable spectral libraries have been developed for the identification of chromatographic materials, such as sepharose resins.

Description

EVALUATION OF CHROMATOGRAPHIC MATERIALS

CROSS-REFERENCE TO RELATED APPLICATION

This application claims priority from United States Provisional Patent

Application No. , filed November 9, 2007 [Zheng Yang,

"EVALUATION OF CHROMATOGRAPHIC MATERIALS", Attorney Docket No.: 16158-125P01, Client Ref. No. AM103012], the entire contents of which is hereby incorporated by reference herein.

TECHNICAL FIELD

This invention relates to evaluation, e.g., identification, of chromatographic materials, such as polymeric resins, e.g., sepharose resins.

BACKGROUND

Chromatographic materials are used in pharmaceutical manufacturing processes for the purification of pharmaceuticals. For example, one family of materials is the sepharose family of resins, which are functionalized agarose resin beads. The sepharose family is exemplified by the members Butyl-sepharose, CM- sepharose, Phenyl-sepharose, Q-sepharose, S-sepharose and SP-sepharose. FIG. IA shows a schematic representation of the sepharose family, while FIG. IB shows a schematic representation of a sub-unit of agarose on which the family of resins is based.

Ensuring the identity and quality of chromatographic resins, such as sepharose resins, is important for maximizing their usefulness in complex biopharmaceutical processes. Many of the wet chemistry methods used to identify chromatographic resins are based on buffer equilibration and/or color indicators. Typically, the wet chemistry methods require experience-based subjective judgment. In some instances, e.g., in the case of S-sepharose and SP-sepharose, the chromatographic members of a family are particularly difficult to distinguish because of their chemical similarity (see FIG. IA).

Therefore, there is a need for new analytical methods that are non-subjective, quick and reliable that can replace or that can be used in addition the currently available wet chemistry methods. SUMMARY

In one aspect, the invention features methods of evaluating, e.g., identifying, chromatographic materials, e.g., a sepharose resin, that include providing, e.g., selecting or developing, a library of reference near infrared spectroscopic signatures for each of a plurality of different reference chromatographic materials; providing, e.g., collecting, a test near infrared spectroscopic signature on a sample of a test chromatographic material; and comparing the test near infrared spectroscopic signature with the library.

Embodiments may include any one or more of the following features. The test chromatographic material is a material whose near infrared spectroscopic signature is provided in the library. The comparison includes comparing the test near infrared spectroscopic signature with a plurality of reference near infrared spectroscopic signatures from the library. The reference chromatographic materials include at least two having the same polymeric backbone but are differently functionalized. Any method further includes selecting the test chromatographic material and using it in a method of preparing a sample, e.g., purifying, fractionating or separating a sample. Any method further includes selecting the test chromatographic material and disposing the material in a chromatographic device, e.g., for purification or separation of a substance, such as a biopharmaceutical. Any method can further include classifying, selecting, accepting or discarding, releasing or withholding, processing into a commercial product, shipping, formulating, moving to a second location, labeling, packaging, releasing into commerce, selling the preparation based on the result of the comparison. The test material and/or the library is/are provided by a first party, e.g., a vendor, of said test material. The test material is provided as a kit which includes the test material and said library and, optionally, a sample of a second chromatographic material represented is the library. The library is accessed over the internet. Any method may further include, based on the comparison, performing one of the following: selecting the test material for use in a manufacturing process, or rejecting the test material for use in a manufacturing process.

Embodiments may also include one or more of the following features. The library is provided by: selecting a family of chromatographic materials on which to build the library; preparing individual family members for analysis; and collecting near infrared spectroscopic signatures on individual family members to provide the library. The preparation of individual family members for analysis has been optimized by: preparing individual family members for analysis by more than a single technique; collecting near infrared spectroscopic signatures on the individual family members; and selecting the preparation technique that maximizes near infrared differentiation among individual family members. The selection process is aided by application of principal component analysis. The techniques are selected from washing, drying, grinding, and mixtures thereof. Drying includes subjecting the sample to reduced pressure, such as in a centrifugal evaporator. The collection process is optimized by: collecting near infrared spectroscopic signatures on the individual family members by more than a single technique; and selecting the technique that maximizes near infrared differentiation among individual family members. The selection process is aided by application of principal component analysis. The technique includes utilizing a continuous range of wavelengths. The range is between about 1,100 nm and 2,500 nm. The technique includes utilizing discrete bands of wavelengths, such as those bands that substantially exclude water absorbances. The discrete bands include a first band having a range between about 1,122 nm and 1,250 nm, a second band having a range between about 1,550 nm and about 1 ,780 nm and a third band having a range between about 2,050 nm and about 2,450 nm. The only the second and third bands are utilized. The family of chromatographic materials includes a polymeric resin, such as a polymeric resin in bead form. Each member of the family of chromatographic materials includes a sepharose resin, such as Butyl-sepharose, CM-sepharose, Phenyl-sepharose, Q- sepharose, S-sepharose and SP-sepharose. The family of chromatographic materials includes a ceramic material, such as a hydroxyapatite, such as type I or type II hydroxyapatite. Any near infrared spectroscopic signature includes intensity and wavelength data. Any near infrared spectroscopic signature includes a derivative of intensity, such as a first or second derivative of intensity, and wavelength data. A noise reduction function, a peak smoothing function, a function that reduces scattering effects, such as multiplicative scattering correction, or mixtures these functions is applied to any near infrared spectroscopic signature. Comparison of the near infrared spectroscopic signature collected on the sample to the library is performed utilizing a technique selected from the group consisting of Mahalanobis Distance on Principle Component Scores (MD), Residual Variance (RV), Wavelength Correlation (WC) and Wavelength Distance (WD).

In another aspect, the invention features methods of performing quality control on making any test material described herein. The methods include evaluating a parameter in a process for manufacturing the test material and, optionally, altering the parameter in the manufacturing based on the evaluation.

In another aspect, the invention features methods of performing quality control on a chromatographic material that include providing, e.g., selecting or developing, a library of reference near infrared spectroscopic signatures for each of a plurality of different reference chromatographic materials; providing, e.g., collecting, a test near infrared spectroscopic signature on a sample of a test chromatographic material; and comparing the test near infrared spectroscopic signature with the library.

In another aspect, the invention features methods of making any library described herein. The methods include providing reference signatures for a plurality of reference chromatographic materials and memorializing said signatures, e.g., writing them onto an optical or an electronic medium.

In another aspect, the invention features methods of a first party, e.g., a manufacturer or vendor, assisting a second party in evaluating a test chromatographic material. Such methods include the first party provides any library described herein to the second party; the second party collects a test near infrared spectroscopic signature on a sample of a test chromatographic material; and then compares the test near infrared spectroscopic signature with the library.

In another aspect, the invention features methods of instructing an end-user on the evaluation of a test chromatographic material that include instructing said end- user to obtain a library of reference near infrared spectroscopic signatures for each of a plurality of different reference chromatographic materials; instructing the end-user to collect a near infrared spectroscopic signature on a sample of a test chromatographic material; and then instructing the end-user to compare the near infrared spectroscopic signature of the test chromatographic material with the library.

In another aspect, near infrared libraries are provided that contain near infrared spectroscopic signatures on a family of chromatographic materials.

In another aspect, the invention features systems for identifying chromatographic materials that include a near infrared spectrophotometer, a near infrared library containing near infrared spectroscopic signatures on a family of chromatographic materials, a computer communicating with the library and the near infrared spectrophotometer and a sample comprising a member of the family of chromatographic materials.

In still another aspect, the invention features methods of purifying therapeutic agents, such as biopharmaceuticals, that include identifying a chromatographic material using a near infrared library containing near infrared spectroscopic signatures on a family of chromatographic materials; and passing a material, such as a reaction mixture, that includes a therapeutic agent through a bed of the chromatographic material.

Any embodiment or aspect described herein can have one or more of the following advantages. The disclosed methods are simple, fast, non-destructive, robust and non-subjective. The methods are less labor intensive than wet chemical methods. The methods generally have small of analyst-to-analyst variation. The methods can simplify and streamline quality control. The methods can be configured to suppress the interference of water on the analysis. The methods are more environmentally- friendly since they do not use wet chemical processes and generate little waste. Identification is generally enhanced relative to wet chemistry methods because the disclosed NIR methods identify a material based on its entire structure, both physical and chemical. The methods are generally applicable to chromatographic materials, including polymeric resins and ceramic materials.

All publications, patent applications, patents, and other references mentioned herein are incorporated by reference herein in their entirety for all that they contain. Other features and advantages will be apparent from the following detailed description, and from the claims.

DESCRIPTION OF DRAWINGS

FIG. IA is a schematic representation of the sepharose family of chromatographic resins, while FIG. IB is a schematic representation of a sub-unit of agarose on which the family is based.

FIG. 2 is a method flow diagram, illustrating development of a near-IR (NIR) library, validation of the NIR library, and then validation of the overall NIR method. FIG. 3 shows overlaid second derivative NIR spectra for Butyl-sepharose samples prepared by Wash+Vaccum Dry+Grind (blue), Wash+SpeedVac (green) and SpeedVac (only) (red).

FIG. 4 is a 3-D principal components analysis (PCA) plot that illustrates NIR differentiation ability for sepharose samples prepared by Wash+Vaccum Dry+Grind (WDG), Wash+SpeedVac (WSV) and SpeedVac alone (SV).

FIG. 5 shows overlaid NIR spectra for CM-sepharose samples after a predetermined time in a SpeedVac (pink-1.5 hours, blue-3 hours, brown-6 hours, pearl-9 hours, purple-12 hours, green-15 hours, yellow-24 hours and red-30 hours). FIG. 6 shows overlaid second derivative NIR spectra for CM-sepharose samples after a predetermined time in a SpeedVac (pearl- 12 hours, green-15 hours, blue-24 hours and red-30 hours).

FIG. 7 shows overlaid full range (1,100-2,500 nm) NIR spectra for six types of sepharose resins (blue is Butyl-, pearl is CM-, yellow is Phenyl-, brown is Q-, red is S- and green is SP-sepharose resin).

FIG. 8 shows overlaid second derivative NIR spectra for the six types of sepharose resins of FIG. 7 utilizing the same identifying color scheme as FIG. 7 (the three selected regions are in red frames).

FIG. 9 is a 3-D PCA plot that illustrates NIR differentiation ability for sepharose samples using full range NIR spectra (1,100-2,500 nm).

FIG. 10 is a 3-D PCA plot that illustrates NIR differentiation ability for sepharose samples using combined three wavelength segments (1,122 nm- 1,250 nm; l,550nm- 1,780 nm; and 2,050nm-2,450 nm), which excluded moisture bands for improved differentiation. FIG. 11 is a 3-D PCA plot that illustrates NIR differentiation ability for sepharose samples using two regions (1^st overtone and combination bands) of the NIR spectral results.

FIG. 12 is a 3-D PCA plot for library specificity tests.

FIG. 13 is a graph of matching value versus months, which illustrates the long-term reproducibility of the method.

FIG. 14 is a graph of matching value versus sepharose resin analytes for three different analysts. FIG 15. shows representative overlaid second derivative NIR spectra for SP- sepharose resin in vials having various inner diameters (green -1.0 cm, blue-1.3 cm and red-2.0 cm).

FIG. 16 is a 3-D PCA plot that illustrates the effect of different sample vial size on NIR differentiation ability for sepharose samples.

DETAILED DESCRIPTION

Quick, non-destructive near infrared (NIR) methods and reliable NIR spectral libraries have been developed for the identification of chromatographic materials, such as sepharose resins. Generally, the disclosed methods of identifying chromatographic materials include providing (e.g., developing or selecting) a library containing identifying NIR spectroscopic information, e.g., signatures, on a family of chromatographic materials. NIR spectroscopic information is then collected on a sample that includes a member of the family of the chromatographic materials. Finally, the NIR spectroscopic information, e.g., signature, collected on the sample is compared to the library that contains the identifying NIR information to identify the member. Such methods described herein are simple, fast, non-destructive, robust, non-subjective, and generally less labor intensive in comparison to the conventional wet chemical methods. In some embodiments, the library is provided by selecting the family of chromatographic materials on which to build the library, preparing individual family members for analysis, and collecting NIR spectroscopic information on individual family members to provide the library of identifying NIR spectroscopic information. For example, in preferred embodiments, the preparation of individual family members for analysis is optimized by preparing individual family members for analysis by more than a single technique, collecting near infrared spectroscopic information on the individual family members, and selecting the preparation technique that maximizes NIR differentiation among individual family members. For example, the selection process can be aided by the application of principal component analysis (PCA). If desired, such analysis can utilize three dimensional PCA plots to help visualize differentiation.

For example, the preparation techniques can be one or more of washing the materials, drying the materials (e.g., removing water or entrained solvent from the materials) and grinding the materials. A preferred technique includes drying the material by subjecting the sample to reduced pressure, such as in a centrifugal evaporator, such a SCl 10 SpeedVac concentrator available from Savant Instrument of Holbrook, NY. In preferred embodiments, the collection process is optimized by collecting

NIR spectroscopic information on the individual family members by more than a single technique and selecting the technique that maximizes near infrared differentiation among individual family members. For example, this optimization process can also be aided by PCA. For example, the technique can include utilizing a continuous range of wavelengths, such as a range between about 1,100 nm and about 2,500 nm, or between about 1,500 nm and about 2,100 nm. In preferred embodiments, the technique includes utilizing discrete bands of wavelengths, such as those bands that substantially exclude interference of water absorbances. For example, the discrete bands can include a first band having a range between about 1,122 nm and about 1,250 nm, a second band having a range between about 1,550 nm and about 1,780 nm and a third band having a range between about 2,050 nm and about 2,450 nm. In preferred embodiments, only the second and third bands are utilized. In still other embodiments, only one of the three bands is utilized.

FIG. 2 shows a preferred process. Such a process begins with selection of samples to include in the library, followed by optimization of the sample preparation to maximize differentiation. Next, the NIR conditions are optimized, preferably with the aid of PCA analysis. Using the optimized conditions, the library is built upon the selected family. Generally, the more family members used in the library and the more samples of a particular member used, the better the results. Next, the library is validated using a number positive and negative controls. Finally, the method itself is validated by analyzing parameters such as long-term stability, analyst-to-analyst variability and other parameters, such as the effect of the sample holder (e.g., glass vial diameter) on the results. Once the method itself is validated, the method can be utilized for unknown samples. NIR methods are discussed in "Guidelines for the Development and Validation of Near Infrared (NIR) Spectroscopic Methods," October 29, 2001, Pharmaceutical Analytical Sciences Group.

Generally, in preferred embodiments, to perform PCA data processing, the generated NIR spectra are exported as tables to a spreadsheet program, such as Microsoft Excel, and here the spectra are converted to their second derivative form. Differentiation (separation) of the data can be visualized utilizing PCA plots with the help of a chemometric software, such as SIMCA, which is available from Umetrics AB of Kinnelon, NJ. PCA analysis is discussed in Shlens, Jonathon, "A tutorial on Principal Component Analysis," December 10, 2005 (version 2).

Generally, in preferred embodiments, to perform an unknown identification analysis, a unknown sample is selected and subjected to SpeedVac drying, followed by NIR measurement. The library is then searched to identify the unknown sample.

In some embodiments, the family of chromatographic materials includes a polymeric resin, such as a polymeric resin in bead form. For example, each member of the family of chromatographic materials can be or can include a sepharose resin, such as Butyl-sepharose, CM-sepharose, Phenyl-sepharose, Q-sepharose, S-sepharose and SP-sepharose. Other polymeric resins include ion exchange resins, e.g., mixed, cationic or anionic exchange resins, cellulosic resins, such as cellulosic gel filtration media, protein or antibody purification media, such as those available from Millipore.

In other embodiments, each member of the family of chromatographic materials includes a ceramic material, such as a hydroxyapatite (e.g., type I or type II), a silica, or an alumina.

In still other embodiments, each member of the family of chromatographic materials includes those configured to exclude by size, such as size exclusion chromatographic resins.

Other embodiments utilize mixed mode chromatographic materials, such as those that exhibit more than one type of interaction, such as ionic, hydrophobic, affinity and size exclusion. In still other embodiments, combinations of any chromatographic materials can be utilized, e.g., ceramics and resins.

In any embodiment described herein, the collected near infrared spectroscopic information on the sample can include intensity and wavelength data, or a derivative of intensity, such as a first or second derivative of intensity, and wavelength data. In any embodiment described herein, a noise reduction function, a peak smoothing function, a function that reduces scattering effects, such as multiplicative scattering correction, or mixtures these functions can be applied to the collected near infrared spectroscopic information. When such functions are utilized, they may be applied before or after any derivative is taken.

In any embodiment described herein, comparison of the near infrared spectroscopic information collected on the sample to the library containing the identifying near infrared spectroscopic information to identify the member can be performed utilizing the Mahalanobis Distance on Principle Component Scores (MD) technique, the Residual Variance (RV) technique, the Wavelength Correlation (WC) technique or the Wavelength Distance (WD) technique.

EXAMPLES

The disclosure is further described in the following examples, which do not limit its scope.

1. INSTRUMENT. APPARATUS AND SOFTWARE

Near Infrared spectra were acquired on a Foss 5000 near infrared (NIR) spectrophotometer with Rapid Content Analyzer (RCA) and Vision 2.5.1 operational software (FOSS NIRSystem, Laurel, MD).

Samples were dried using a SCl 10 SpeedVac concentrator (Savant Instrument, Holbrook, NY).

Chemometric analysis, such as principal component analysis (PCA), was preformed using SIMCA-P+ 11 software (Umetrics AB, Kinnelon, NJ).

2. MATERIALS

Aminobutyl-, Butyl-, Blue-, CM-, DEAE-, Octyl-, Phenyl-, Q-, S-, and SP- sepharose resins were obtained from GE Healthcare, Uppsala, Sweden. These sepharose resin samples were obtained as slurries in 20% ethanol aqueous solutions. 0.2M sodium acetate was also present as preservative in S- and SP-sepharose slurries.

Cellufϊne Butyl, Cellufϊne Phenyl, Cellufme Sulfate and Cellulose Gel Filtration Media were obtained from Chisso, Tokyo, Japan, ceramic hydroxyapatite (types I and II) and UNOsphere Q were obtained from Bio-Rad, Hercules, CA, Poros 50HE was obtained from Applied Biosystems, Foster City, CA, and Toyopearl QAE 550C and Toyopearl DEAE 650C was obtained from Tosoh, Tokyo, Japan. A. GENERAL METHODOLOGY 1. Sample Preparation

Sepharose resins were prepared utilizing one of three methods, which are summarized below in Table 1.

Table 1. Sepharose resin sample preparation procedures

Eight SpeedVac drying time intervals (1.5, 3, 6, 9, 12, 15, 24 and 30 hours) were tested.

2. Data Collection and Analysis

NIR spectra were acquired from vials filled with sample powder placed on the sampling window of the RCA of the calibrated NIR instrument.

Data pre-treatment included computing the second derivative of spectra, and then three different combinations of spectral regions were evaluated by PCA. The wavelength combinations included: (a) full NIR range (1,100 nm-2,500 nm), (b) three spectral regions (1,122 nm-1,250 nm for C-H 2^nd overtone, 1,550 nm-1,780 nm for C-H 1^st overtone, and 2,050 nm-2,450 nm for combination bands of N-H, N-H + C-H, and C-H + C-H) and (c) two regions (1,550 nm-1,780 nm and 2,050 nm-2,450 nm). 3. Library Building

Four library search methods (Mahalanobis Distance on Principle Component scores, Residual Variance after PC analysis, Wavelength Correlation and Wavelength Distance) were compared for identification methods, and associated parameters were optimized.

Library validation tests, such as internal specificity, correlation for library sepharose analytes and external specificity tests were executed.

The three groups of challenge samples for external specificity tests were: (1) positive (+) controls, a new lot for each library-contained sepharose resin was searched as "unknown" for its identity; (2) two groups of negative (-) challenge resin samples, sepharose type and non-sepharose type, were checked for any false (+) library search results; and (3) Several sepharose resin analytes were deliberately mixed to test mismatch tendency for the library.

4. Method Validation

For long-term repeatability, one lot from each of the six-sepharose types was run monthly for six months. Inter-analyst repeatability was performed by three analysts using same lot of each sepharose analyte but on different days.

For method robustness tests, the effect of vial bottom inner diameter (or change of NIR reflectance sampling area) was verified. The sample glass vials in three bottom-inner-diameter sizes (1.0 cm, 1.3 cm and 2.0 cm) were tested (each size in triplicate determinations) by using same ground sepharose resin samples.

B. SAMPLE PREPARATION 1. Comparison of Sample Preparation Treatments

Second derivative NIR spectra were utilized to enhance spectral features and to decrease unwanted sources of variation, such as sizes of sample particles and NIR reflectance sampling areas. Overlaid second derivative spectra from a representative sepharose, Butyl-sepharose, treated as outlined in Table 1 are shown in FIG. 3. In FIG. 3, blue color is the spectra obtained by the Wash+Vaccum Dry+Grind treatment, green is that obtained from the Wash+SpeedVac treatment and red is the SpeedVac (only) treatment. FIG. 3 show that the curve intensities varied from large to small in order from Wash+Dry+Grind method, Wash+SpeedVac to SpeedVac, respectively. Without wishing to be bound by any particular theory, this variation might be caused by washing away sodium acetate and other preservatives from the sepharose resins in the washing steps, and/or by morphological effects of drying.

FIG. 4 is a 3-D principal components analysis (PCA) that illustrates NIR differentiation ability for sepharose samples prepared by Wash+Vaccum Dry+Grind (WDG), Wash+SpeedVac (WSV) and SpeedVac alone (SV). As shown in FIG. 4, generally all of the sample preparation techniques show good differentiation. Generally, SpeedVac sample preparation is preferred since it is simple and less labor- intensive, and provides acceptable NIR discriminative results.

2. SpeedVac Optimization

During SpeedVac procedure optimization, moisture band intensities around 1,900 nm and 1,400 nm decreased as SpeedVac drying hours lengthened. FIG. 5 shows overlaid NIR spectra for CM-sepharose samples after a predetermined time under a SpeedVac (pink- 1.5 hours, blue-3 hours, brown-6 hours, pearl-9 hours, purple-12 hours, green-15 hours, yellow-24 hours and red-30 hours). FIG. 6 shows overlaid second derivative NIR spectra for CM-sepharose samples after a predetermined time under a SpeedVac (pearl- 12 hours, green-15 hours, blue-24 hours and red-30 hours). There was little difference in 2^nd derivative NIR spectra when SpeedVac drying times ranged between 12 hours and 30 hours.

C. EXEMPLARYNIR SPECTRA FOR THE SEPHAROSE ANALYTES

FIG. 7 shows overlaid full range (1,100 nm-2,500 nm) NIR spectra for six types of sepharose resins (blue is Butyl-, pearl is CM-, yellow is Phenyl-, brown is Q-, red is S- and green is SP-sepharose resin), while FIG. 8 shows overlaid full range second derivative NIR spectra for the six types of sepharose resins of FIG. 7, utilizing the same identifying color scheme as FIG. 7. The three boxes in red indicate the three spectral regions discussed herein.

Substantially identical agarose backbones (see FIG. 1) diminishes spectral difference among these sepharose analyte analogs increasing the difficulty to distinguish them. Thus, 2^nd derivative treated spectra can employed to modulate spectral differences. For instance, the typical 2^nd derivative spectrum of Q-sepharose (see brown color curve in FIG. 8) has large rather sharp peaks in the region (2,050 nm-2,450 nm) of N-H + C-H and C-H + C-H combination bands.

D. LIBRARYBUILDINGAND OPTIMIZATION FIG. 9 is a 3-D PCA plot that illustrates NIR differentiation ability for sepharose samples using full range NIR spectra (1,100 nm-2,500 nm). Analysis based on full range (1,100 nm-2,500 nm) spectra is subject to interference by moisture peaks and other variations.

As shown in FIG. 10, which is a 3-D PCA plot that illustrates NIR differentiation ability for sepharose samples using combined three wavelength segments (1,122 nm- 1,250 nm; 1,550 nm-1,780 nm; and 2,050 nm-2,450 nm), excluding water bands improved separation among the sepharose resin species.

FIG. 11 is a 3-D PCA plot that illustrates NIR differentiation ability for sepharose samples using the two high intensity regions (1^st overtone and combination bands). FIG. 11 shows results of excluding the weak 2^nd overtone region.

Separation power for each technique is summarized in Table 2. The three- region method was selected for NIR library building because is generally gave the most favorable differentiation.

Table 2. Spectral region comparison for the PCA plots

E. LIBRAR Y SEARCH OPTIMIZA TION

Four library search methods (Mahalanobis Distance on Principle Component scores, Residual Variance after PC analysis, Wavelength Correlation and Wavelength Distance) were compared for identification methods, and associated parameters were optimized, the results of which are summarized in Table 3. Combining the proper ID method, followed by a qualification method at right threshold insures correct identification of the library analytes, which is especially useful for distinguishing chemically similar compounds. Table 3. Threshold tuning for the library methods ^l

Approximate 200 NIR spectra were screened against each library method tested.

²MV = Matching Value, and PL = Probability Level.

³Optimum thresholds may depend on the number of spectra for each library product and on (-) challenge samples.

F. LIBRARY VALIDATION

1. Library Internal Specificity

Library internal specificity for six sepharose analytes in the library is summarized below in Table 4. Table 4. Internal specificity for six sepharose analytes in the library

'Using WC method with MV threshold = 0.900 for Identification

2. Correlation for the Sepharose Analytes NIR spectral similarity among six sepharose analytes in the library was evaluated, and are summarized in Table 5. High correlations between two pairs of sepharose resins were 72.9% for Butyl-sepharose vs. CM-sepharose and 69.5% for S- sepharose vs. SP-sepharose. This agrees with the PCA plot shown in FIG. 10.

Table 5. Correlation (%) for the sepharose resin analytes in the library ^l

Comparing the mean values (n = 15)

3. Library External Specificity

Control samples for library specificity tests are summarized in Table 6, and visualized in FIG. 12. Closely correlated negative resin species to certain sepharose analytes are summarized in Table 7. Table 6. Specificity test results

Table 7. Closely correlated negative resins to the sepharose analytes (in % Matching Value)

^-sepharose and SP-sepharose had no (-) resin samples with a correlation > 45%. G. METHOD VALIDATION

1. Intermediate Precision

FIG. 13 is a graph of matching value versus months, which illustrates the long-term reproducibility of the method over 6 months (ID threshold, MV = 0.900), while FIG. 14 is a graph of matching value versus sepharose resin analytes for three analysts, illustrating that analyst-to-analyst variation is minor.

2. Method Robustness Tests

FIG 15. shows representative overlaid second derivative NIR spectra for SP- sepharose resin in vials having various inner diameters (green -1.0 cm, blue-1.3 cm and red-2.0 cm). The absorbance of sepharose in 1.0 cm-diameter vials showed the highest values among three parallel curves, while sampling with 2.0 cm-diameter vials shows the lowest. However, the absorbance differences were offset by 2^nd derivative spectra. Some intensity variation near 1,500 nm is believed to be caused by absorbance of glass materials. This variation is not within the selected three wavelength regions so it generally does not interfere with analysis. Insignificant vial size effect was also confirmed by a PCA plot, which is shown in FIG. 16.

OTHER EMBODIMENTS

A number of embodiments have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the invention. Other embodiments are within the scope of the following claims.

Claims

WHAT IS CLAIMED IS:

1. A method of evaluating a chromatographic material, comprising: providing a library of reference near infrared spectroscopic signatures for each of a plurality of different reference chromatographic materials; providing a test near infrared spectroscopic signature on a sample of a test chromatographic material; and comparing the test near infrared spectroscopic signature with the library.

2. The method of claim 1, wherein said test chromatographic material is a material whose near infrared spectroscopic signature is provided in said library.

3. The method of claim 1, wherein said comparison comprises comparing said test near infrared spectroscopic signature with a plurality of reference near infrared spectroscopic signatures from said library.

4. The method of any one of claims 1-3, wherein the reference chromatographic materials include at least two having the same polymeric backbone but are differently functionalized.

5. The method of any one of the above claims, further comprising selecting said test chromatographic material and using it in a method of preparing a sample.

6. The method of any one of claims 1-4, further comprising selecting said test chromatographic material and disposing said material in a chromatographic device.

7. The method of any one of claims 1-4, further comprising classifying, selecting, accepting or discarding, releasing or withholding, processing into a commercial product, shipping, formulating, moving to a second location, labeling, packaging, releasing into commerce, selling the preparation based on the result of the comparison.

8. The method of any one of the above claims, wherein said test material and/or said library is/are provided by a first party of said test material.

9. The method of any one of claims 1-4, wherein said test material is provided as a kit which includes said test material and said library and, optionally, a sample of a second chromatographic material represented is said library.

10. The method of any one of the above claims, wherein said library is accessed over the internet.

11. The method of claim 1 , further comprising, based on said comparison, performing one of the following: selecting said test material for use in a manufacturing process; or rejecting said test material for use in a manufacturing process.

12. A method of performing quality control on making the test material of claim 1 , comprising: evaluating a parameter in a process for manufacturing said test material and, optionally, altering the parameter in the manufacturing based on the evaluation.

13. A method of performing quality control on a chromatographic material, comprising: providing a library of reference near infrared spectroscopic signatures for each of a plurality of different reference chromatographic materials; providing a test near infrared spectroscopic signature on a sample of a test chromatographic material; and comparing the test near infrared spectroscopic signature with the library.

14. A method of making the library of claim 1, comprising providing reference signatures for a plurality of reference chromatographic materials and memorializing said signatures.

15. A method of a first party assisting a second party in evaluating a test chromatographic material, comprising: said first party provides said library of claim 1 to said second party; said second party collects a test near infrared spectroscopic signature on a sample of a test chromatographic material; and then compares the test near infrared spectroscopic signature with the library.

16. A method of instructing an end-user on the evaluation of a test chromatographic material, comprising: instructing said end-user to obtain a library of reference near infrared spectroscopic signatures for each of a plurality of different reference chromatographic materials; instructing said end-user to collect a near infrared spectroscopic signature on a sample of a test chromatographic material; and then instructing said end-user to compare the near infrared spectroscopic signature of the test chromatographic material with the library.

17. The method of claim 1 , wherein the library is provided by: selecting a family of chromatographic materials on which to build the library; preparing individual family members for analysis; and collecting near infrared spectroscopic signatures on individual family members to provide the library.

18. The method of claim 17, wherein the preparation of individual family members for analysis has been optimized by: preparing individual family members for analysis by more than a single technique; collecting near infrared spectroscopic signatures on the individual family members; and selecting the preparation technique that maximizes near infrared differentiation among individual family members.

19. The method of claim 18, wherein the selection process is aided by application of principal component analysis.

20. The method of claim 18, wherein the techniques are selected from washing, drying, grinding, and mixtures thereof.

21. The method of claim 20, wherein drying comprises subjecting the sample to reduced pressure.

22. The method of claim 18, wherein the collection process is optimized by: collecting near infrared spectroscopic signatures on the individual family members by more than a single technique; and selecting the technique that maximizes near infrared differentiation among individual family members.

23. The method of claim 22, wherein the selection process is aided by application of principal component analysis.

24. The method of claim 22, wherein the technique comprises utilizing a continuous range of wavelengths.

25. The method of claim 24, wherein the range is between about 1,100 nm and 2,500 nm.

26. The method of claim 18, wherein the technique comprises utilizing discrete bands of wavelengths.

27. The method of claim 26, wherein the discrete bands comprise a first band having a range between about 1,122 nm and 1,250 nm, a second band having a range between about 1,550 nm and about 1,780 nm and a third band having a range between about 2,050 nm and about 2,450 nm.

28. The method of claim 27, wherein only the second and third bands are utilized.

29. The method of any one of claims 17-28, wherein the family of chromatographic materials comprises a polymeric resin.

30. The method of any one of claims 17-28, wherein each member of the family of chromatographic materials comprises a sepharose resin.

31. The method of claim 30, wherein the sepharose resin is selected from the group consisting of Butyl-sepharose, CM-sepharose, Phenyl-sepharose, Q-sepharose, S- sepharose and SP-sepharose.

32. The method of any one of claims 17-28, wherein each member of the family of chromatographic materials comprises a ceramic material.

33. The method of any one of the above claims, wherein any near infrared spectroscopic signature comprises intensity and wavelength data.

34. The method of any one of claims 1-33, wherein any near infrared spectroscopic signature comprises a derivative of intensity and wavelength data.

35. The method of any one of the above claims, wherein a noise reduction function, a peak smoothing function, a function that reduces scattering effects, or mixtures these functions is applied to any near infrared spectroscopic signature.

36. The method of any one of the above claims, wherein comparison of the near infrared spectroscopic signature collected on the sample to the library is performed utilizing a technique selected from the group consisting of Mahalanobis Distance on Principle Component Scores (MD), Residual Variance (RV), Wavelength Correlation (WC) and Wavelength Distance (WD).

37. A near infrared library containing near infrared spectroscopic signatures on a family of chromatographic materials.

38. The near infrared library of claim 37, wherein each member of the family comprises a polymeric resin.

39. The near infrared library of claim 37, wherein each member of the family comprises a ceramic material.

40. A system for identifying a chromatographic material, comprising: a near infrared spectrophotometer; a near infrared library containing near infrared spectroscopic signatures on a family of chromatographic materials; a computer communicating with the library and the near infrared spectrophotometer; and a sample comprising a member of the family of chromatographic materials.

41. A method of purifying a therapeutic agent, comprising: identifying a chromatographic material using a near infrared library containing near infrared spectroscopic signatures on a family of chromatographic materials; and passing a material comprising a therapeutic agent through a bed of the chromatographic material.