Strategic Considerations for Screening Natural Products

 

- medicinal plants

 

Strategic Considerations for Screening Natural Products

Dr. Matthew A. Sills 

Natural product screening has been a component of lead finding for the better part of 15-20 years. Novel pharmaceuticals, such as taxol, have been discovered through screening of extracts from plants, microorganisms and marine organisms. However, a number of technological advances during the past few years have significantly impacted upon earlier strategies for screening natural products. In particular, combinatorial chemistry has provided an alternative means to increase sample diversity, and advances in automation and instrumentation have enabled in-house compound libraries to be screened in a fraction of the time required only a few years ago. In this article, I will discuss how these advances in biomolecular screening have impacted upon natural product screening strategies.

In the pre-Renaissance era of biomolecular screening, programs were significantly smaller than today's programs. During the late 1980s to early 1990s, screening 10,000 samples per year per assay was considered reasonable. The duration of an assay within the program might be 2-3 years to evaluate a "significant" number of samples in each assay. The goal of the natural product screening program was to discover a potential lead compound within this time frame.

During these early days at Ciba, we evaluated 10,000 compounds per year in 15 assays, and the typical life span of an assay was three years. Each month, approximately 900 samples were evaluated in each assay. In 1989, a five-year collaboration was established with Harbor Branch Oceanographic Institute (http://www.hboi.edu), a non-profit research organization located in Ft. Pierce, FL, in order to identify novel compounds isolated from marine organisms. When a number of factors were considered in establishing a strategy to identify pure compounds from these natural product extracts, the primary limiting factor was the resources to purify natural products; i.e. the defined number of natural product chemists could work on only a limited number of extracts. The strategy ultimately implemented involved screening a set number of extracts per month, selection of extracts for purification, and efforts to purify and identify the active component. In the program, extracts were prepared by Harbor Branch and evaluated each month in assays at Ciba. From extracts of interest, aqueous and organic partitions were prepared at Harbor Branch and then were assayed at Ciba. Samples were selected for purification based on potency, preliminary structural information and the priority of the target assay. Purification and identification of the active compound occurred at Harbor Branch, with evaluation of the purification fractions at Ciba. Purification of approximately six extracts was carried out concurrently with efforts being focused on one or perhaps two extracts for a particular target at any given time. The time to purify a compound and identify its structure was typically six months but could range from 1-12 months. The overall process, from when the extract was first screened to when the compound was identified, could take from 1-2 years. Since assays remained in the screening program for several years, this was considered an acceptable timeline.

The collaboration with Harbor Branch resulted in the identification of more than 40 novel compounds. The structures of several of these compounds are shown in Fig. 1 and the time frame for their isolation is summarized in Table 1.

 

Figure 1: Three structures isolated from the collaboration between Ciba Pharmaceuticals and Harbor Branch Oceanographic Institute. The values shown represent the average K_i or IC₅₀ value in the respective binding assay.

The sollasins, which were identified as angiotensin AT-1 receptor antagonists, were isolated from the organism, Poecillastra sollasi ⁽¹⁾. As indicated in Table 1, the extracts were initially screened in June, 1989. Purification was initiated in June, 1990 and the compounds were isolated and identified in May, 1991. This time frame was considered acceptable then since it fell within the life span of the assay in the screening program.

Table 1
Time Frame for Identification of Compounds
from Marine Extracts

Compound	Assay	Screened	Begin Purification	Compound Isolated
Sollasin	Angiotensin AT-1	June 1989	June 1990	May 1991
Strongylin	Endothelin	Dec. 1991	Jan. 1993	Aug. 1993
Nordercitin	C5a	April 1992	Aug. 1993	Sept. 1993

Another group of compounds isolated from a marine organism were the strongylins. These compounds, purified from Strongylophora sp., were identified as endothelin-A receptor antagonists. These compounds, which were initially screened as extracts in December 1991, were purified in 8 months.

During the course of the collaboration, the need to identify compounds from natural product extracts more rapidly was identified. This issue developed as active compounds were identified from the Ciba pure compound library and medicinal chemistry efforts were allocated to these leads. Compounds isolated from natural product screening six months or more after internal compounds had demonstrated activity had difficulty competing for medicinal chemistry resources. In order to more effectively identify potential lead compounds, the natural product screening strategy was revised to place emphasis on purifying fewer organisms more rapidly. An example of this effort was nordercitin ⁽²⁾, which was isolated from a Stelleta sp., and identified as a C5a receptor antagonist. This compound was purified in less than two months.

During the past few years, the need to rapidly identify leads has intensified to the point that the strategy of screening a defined number of extracts per month over a period of 2-3 years would no longer be successful. The main problem with this strategy is that the most interesting or potent extract might not be identified until the second half or latter part of the assay's life span in the screening program. This also leads to the problem that purification work might not commence on a particular extract for several months, possibly extending to a year, because resources are being utilized to purify extracts for other projects or to purify other extracts for the same assay. As shown in Table 1, it took one year or longer for purification efforts to be initiated on each of the three examples once the extract had been initially screened. Although it took less than two months to purify nordercitin, it still took almost 16 months before purification efforts were initiated.

In considering an alternative strategy, it is important to realize that not all purification efforts result in the identification of a potential lead compound. In some instances, these efforts result in purification being terminated due to loss of activity or the isolation of a structure that is not of interest. Unfortunately, it is not clear beforehand which organism will yield the most interesting compound.

Today, with advances in instrumentation and automation, screening programs have the ability to evaluate hundreds of thousands of samples within a period of 6-12 months. With this increase in capacity, the ability of natural product compounds to "compete" with compounds identified from the in-house library as potential lead compounds has become increasingly difficult. In order to compete more effectively, an alternative strategy to screen natural products is required. one strategy to address this issue would be to screen all of the natural product extracts at the beginning of the assay, such as during the first 1-3 months. This would allow the most interesting extracts to be selected for purification toward the beginning of the assay's life span in the screening program, and should enable compound isolation and identification to be completed by the end of the assay's life span. In addition, this strategy would minimize the time between identification of extract activity and initiation of purification, and would allow the most active samples to be purified with the highest priority.

An alternative strategy would be to screen natural products only if no leads were identified from the in-house compound library. The main advantage of this approach is that natural products would no longer have to compete with active in-house compounds. Even when natural products are screened at the beginning of an assay and compound isolation occurs rapidly, lead optimization may not occur until all of the internal compounds are evaluated to determine if a simpler structure or a more potent compound was identified. If no potential leads are identified from the internal library, compounds isolated from natural products would receive full consideration and be the focus of any medicinal chemistry effort. A disadvantage of this approach is that, if competitive pressures limit the window of opportunity for the project to identify a lead, there may not be sufficient time to screen natural products after the pure compound library has been completed.

In summary, natural products remain an excellent avenue to identify novel structures for therapeutic targets. However, advances in biomolecular screening have enabled HTS programs to evaluate internal libraries within a period of 6-12 months or less. Older strategies of screening natural products that resulted in compound isolation in 1-2 years must be revised in order for these compounds to be seriously considered as potential leads.

 

Acknowledgment

I would like to gratefully acknowledge Dr. Amy Wright (Wright@hboi.edu) for her helpful comments and discussion in the preparation of this article.

 

References

1 Killday, K.B., Longley, R., McCarthy, P.J., Pomponi, S.A., Wright, A.E., Neale, R.F. and Sills, M.A., "Sesquiterpene derived metabolites from the deep water marine sponge poecillastra sollasi ", J. Nat. Products, 56: 500, 1993.

2 Gunawardana, G.P., Kohmoto, S., Burres, N.S., "New cytotoxic acridine alkaloids from two deep water marine sponges of the family pachastrellidae", Tetrahedron Lett., 30: 4359, 1989.