The right chemistry

Combinatorial chemistry is a field that has been with us for over a decade. It was very much heralded back in the early to mid 90s as an opportunity to dramatically change pharmaceutical research by producing large numbers of compounds that would provide many options to quickly get to an optimized drug lead. Unfortunately, there was a great deal of hype that went around this particular promise back then, which exceeded what many of us thought was likely to occur from the technology. We recognized it had value, but didn’t think it would single-handedly transform the whole industry.

What happened over the first few years in this field led to some adjustments. In the early days the emphasis of combinatorial chemistry was to prepare enormous collections of compounds. We and others launched into that, but found by the late 90s that these strategies were flawed ­ the kind of compounds that came out of it did not have the kind of properties that drugs typically have. Even though large collections of compounds were being made, they were not that different from each other and the purity was often low. So, as a result, the ability to generate a useful chemical matter out of that initial approach was universally found to be disappointing. However, all technologies evolve and they often fall into areas that are quite useful. This has happened to combinatorial chemistry. As this technology has evolved, I see the field now being better named as high throughput chemistry as opposed to combinatorial chemistry.

With high throughput chemistry, the emphasis now is on preparing smaller collections of compounds ­ not millions, but dozens to hundreds of compounds which are more varied, are pure, and based on basic chemical principles that are consistent with drugs. This has allowed us to focus on producing the kind of molecules that we really need.

What processes take place in your High Throughput Organic Synthesis (HTOS) lab? Have you made any exiting discoveries?

HTOS is really something that we at Abbott pioneered. Many companies, as they tried to take advantage of this high throughput chemistry evolution looked for a way to exploit that within their chemistry organizations. Other companies have trained their entire chemistry staff in high throughput organic synthesis and they have provided large amounts of equipment and created big infrastructures to do that. We did some of that as well. But, the key insight that we had in creating HTOS was to think about creating a central service laboratory. The HTOS lab can prepare almost any library to order that a medicinal chemist working on a drug research programme wants and does that with high reliability, convenience and rapid turnaround time. Rapid turnaround time is particularly important since libraries must be synthesized within the normal timelines of drug research.

The HTOS concept is that a chemist working on a research programme envisions a library that would be appropriate for that programme. The chemist prepares the core part of the molecules that constitutes the library, and then go to the HTOS lab and, using the computer, selects from a wide array of different variations that would be used to decorate that core. There are literally thousands of available reagents that are at our disposal. And then, within two to three weeks the libraries are produced, purified, wholly characterized, analyzed to make sure they are pure and the right compounds, and then distributed to the biology labs for the results. So within a two to three week frame of ordering the library, medicinal chemists get back the biological results from the library they ordered – again with high reliability and efficiency.

The key to our success with HTOS is that we invested over two years working out the chemistry to do this. A lot of the time the key focus of high throughput chemistry has been automation and having the right systems in place. Certainly you need to do that, but if you are going to be able to produce libraries on demand as a service, you can't take the time to work out how to do the reaction every time. That is why we developed standard procedures that are likely to work with a wide variety of different orders that will be placed so that there is very minimal chemistry development time required. This means we can produce almost any library in less than three weeks time. We have spent a lot of effort over the course of the two to three years before the service launched developing a wide array of chemistries that would work reliably with a lot of different libraries and we continue to do so.

What kind of impact has High Throughput Chemistry had on R&D in the industry?

We have measured the kind of impact our HTOS lab is having. There are two real ways of measuring the impact of this kind of approach. One is simply to look at the productivity ­ the raw numbers of compounds that are being produced. About half of the compounds that are being prepared by Abbott chemists are produced by this lab of less than a dozen people. The raw numbers, in terms of output compounds, are substantial.

This is only one measure though and it is probably not the most important measure. A lot of compounds, if they are not the right compounds are not helpful. The group can document and produce the right compounds ­ we can see that in several different ways. One is, if we look at the patents that are produced that come out of Abbott ­ about half of the patents being applied for have HTOS compounds in them.

We can also chart where key decisions are being made as a result of compounds and information that is coming from those compounds. Finally, if you look at candidates that are moving into development, there are pieces of these molecules, which are derived from HTOS compounds. The two metrics basically ­ high efficiency producing a lot of compounds, and more importantly the right compounds.

New synthetic technologies such as HTOS have increased the number of compounds made and tested annually. Because of this existing IT at many companies has been inadequate to accommodate these increases. Has the organization had to overcome this problem? If so, how?

Absolutely. When you are talking about producing thousands of compounds a year keeping track of all that information is really challenging. We had a chemist developer that handled the IT issues. By that I mean we had a trained chemist, but also someone with substantial skills in IT development. We had a person like that who lived in the HTOS group and was responsible for both developing and coordinating the development in IT systems when we needed it. The advantage of that was having a person knowledgeable both in the IT and the chemistry area ensuring there was no need to translate between these two disciplines.

Another advantage is having somebody who is in the lab and able to work as this technology is being developed. All of discovery research (and particularly something like this) is always changing and evolving. In order for IT to keep pace with that, it is important to have people who are imbedded in the research groups who can respond to that. Those are two of the things we've done to try and address the IT challenge.

How do you envisage the future development of High Throughput Chemistry?

It's unrealistic to expect that any HTOS lab or approach is going to completely supersede the traditional manual synthesis of compounds any time soon, so there must continue to be a working role for traditional chemistry. This is true because as projects begin to focus on increasingly narrower types of compounds the chemistries become more difficult to rapidly automate and the availability of specialized reagents becomes less. I want to make it clear that while HTOS has a very important role, I don't see it overcoming that particular role of chemistry.

What I do see it moving more into is automated multi-step libraries. Currently a lot of work is focusing on libraries that involve only one chemical step. The automation in chemistry is now evolving to the point where you can conduct multiple steps on a single platform.

Can you explain what the company has done to automate microwave chemistry? Has this contributed to a significant improvement in productivity?

There are two issues with automating microwave chemistry. The first is the simple fact that if you have a microwave synthesizer running a lot of reactions you need to be able to change the samples. Vendors provide that kind of capability today, so that need is being met.

At Abbott, we focused on a second problem with automating microwave chemistry. The microwave reaction itself is not the rate-limiting step in the process of preparing libraries. We recognized that the real advance would be achieved by also automating the preparation of samples prior to the reaction and the work-up that follows. With this in mind our internal engineering group developed an automated platform we call MAESTRO. It handles reagent preparation, heating, work-up and purification distribution in an automated manner on 100 samples at a time. It can take a three-day process of preparing a library and complete it in one day.

How can microwave technology make a big impact in drug discovery?

I think microwave technology is the biggest technological advancement in chemistry in the last five to ten years. As we get more experience with microwave technology we've definitely learnt that it has applications in many ways that we didn't envision initially. Most people envision microwave technology to be most applicable in cases where for example, a conventional reaction would be heated ­much like you would try heating up your dinner in a conventional over or microwave oven. What we have come to learn is that even chemistry that we did not traditionally think about in terms of heating would be applicable to microwave. It would run much cleaner, with much higher yields and much purer. I think our knowledge is continuing to grow in microwave technology.

What does the future hold for combinatory chemistry?

I think high throughput chemistry, which is the evolution from combinatorial chemistry will continue to be a platform for a large amount of compounds that are made in drug discovery. More and more new chemists come out of school and are familiar with this particular approach, so we will see greater use of that by everybody in this area. The advancements in our knowledge of both automation technology and chemistry that is amenable to automation will allow us to run more and more types of reactions and to produce more compounds in this way. I see it continuing to play a big role in drug discovery research.