CPSA Digest 2001

Proceedings -Tuesday, October 9, 2001

TuOD1

Mix & Split Libraries to Parallel Synthesis:
A Collaborative High Throughput Chemistry / Analytical Chemistry Transition (1996-2001)

Background:
The progression of analytical chemistry techniques at Glaxo SmithKline is profiled in this presentation. Procedures and hardware in use from 1996 to present day 2001 are reviewed, demonstrating the dramatic impact of today's high throughput chemistry and analysis techniques. Back in 1996-1997, chemistry optimization used open access (OA) flow injection analysis/MS having a throughput rate of 20 samples per hour. LCMS analysis required submitting a request and waiting 1-2 days for turnaround (using 20 min run times). Also about this time, library QC/Hit identification used the one bead/one compound approach. The molecular ion was identified and matched against a list of possible molecular weights; purity was estimated at 214 nm. Various outside pressures emerged to improve throughput needs: higher density HTS platforms (40 million wells screened per year), greater quantities required of synthesized compounds, HT ADME tox assays were introduced, and parallel synthesis methods evolved (requiring LC-based purification techniques). The demand for LCMS services increased and newer strategies to improve throughput, standardization and communication across groups were required.

Premise:
Combinatorial chemistry synthesis techniques have been primarily applied toward the creation of drug discovery leads. Samples can be prepared and assayed as mixtures; active compounds are identified by one of three processes: a deconvolution process, proprietary resin "tagging" schemes and direct MS identification. Hits are confirmed by re-synthesis. Parallel synthesis of discrete compounds is also performed to deliver large masses (5-20 mg) of compounds in support of worldwide HTS, selectivity and ADME tox assays. In this scheme, from 24 to 10,000 compounds are prepared per array; thus, 100,000 - 250,000 compounds per year. Clearly, high throughput chemistry is required to support this accelerated and widespread compound synthesis and analysis need.

Analytical contributions and high throughput chemistry needs today can be classified into four categories:

1. Chemistry optimization and validation
2. Array production and QC
3. Array purification
4. HTS monitoring

Chemistry Optimization and Validation Open access LCMS has proved popular with chemists for many reasons. The technique identifies reaction products (M+H), assesses purity of reagent/product, monitors reaction progress and rates, provides quick answers using convenient microplate formats, and the system is maintained and upgraded by experts. The benefits of standardization are clearly evident, e.g., friendly interface, robust gradient LCMS methods, a common data presentation format and intranet database storage and retrieval of results.

Open access usage trends have increased over the years at GSK, as shown below.

Open access development needs center on several issues. An increase in throughput without loss of selectivity is needed (shorter cycle times, parallel analysis). Scheduling for large batch jobs (overnight) should be available. An on-line universal detector and means for quantitation is needed.

Array Production and QC Using LCMS and HT-NMR LCMS is used for purity assessment of arrays. Standardization for analytical methods and hardware is required. Custom scripts for identification are useful. LCMS data is integrated with in-house decision-making software. Additional LCMS development needs include standard approaches for data archival, intelligent data review and revision systems, and improved throughput (e.g., desire 1-2 hr analysis time for a 500 compound array).

HT-NMR uses a Varian VAST (500 MHz); samples are in microplate format; 6 min/analysis. Development needs for HT-NMR center on reducing the acquisition and analysis time so that it is less than or equal to that of LCMS. Also, reduced compound mass for analysis is desired so that product yield is not compromised when an aliquot is removed for HT-NMR.

Array Purification Preparative LC methods are now in use that are rapid (6 min gradient; 10-100% ACN at 25 mL/min). A 20 x 50 mm column (5 um) is commonly used to purify from 5-100 mg crude material. Generic methods developed are suitable for 90% of the assay requirements. Fractionation is commonly triggered by a UV signal. Topics of debate in array purification include the approach of a "purification specialist group" vs. open access for all chemists, and the benefit to risk of human intervention in the process. Development needs for LC-based methods are: improving sample throughput, improved trigger using both UV and MS, scale-optimized systems and a chemist-friendly and reliable operation.

HTS Monitoring Good decision-making in reviewing HTS data and hit characterization requires accurate, catalogued and accessible data.