|Home | About | Journals | Submit | Contact Us | Français|
Since 2010, Dr. Thomas Lundbäck has been a senior scientist at Karolinska Institutet (KI), working with assay development and screening at the KI node of Chemical Biology Consortium Sweden (CBCS). He received his PhD in biophysical chemistry at Karolinska Institutet under the supervision of Professor Torleif Härd in 1996, followed by two years of post-doctoral studies in the laboratory of Prof. John Ladbury at University College London. In 1999 he joined the Pharmacia & Upjohn organization in Stockholm as a biophysical chemist, but quickly moved to the area of assay development for high throughput screening purposes. Dr. Lundbäck stayed within the pharmaceutical industry for 10 years before he joined GE Healthcare in Uppsala in 2009, working with the development of new applications on their instruments for biomolecular interaction studies. His research interest is primarily focused on understanding the driving forces behind molecular interactions, with particular focus on the actions of small molecules, and when the CBCS initiative was started in 2010 he was the first person to join the founders. Besides working with collaborative projects he is responsible for the project prioritization process at CBCS. Dr. Lundbäck is also part of the Europe Education Advisory Committee for the Society for Laboratory Automation and Screening (SLAS) organization and he is an Associate Editor at ASSAY and Drug Development Technologies. Photo (from left to right): Helena Almqvist, Thomas Lundbäck, Hanna Axelsson, and Brinton Seashore-Ludlow. Photo by Lars Hammarström.
When did you first develop an interest in science? Were there certain individuals who had a particularly strong influence on your chosen career path?
I have had a special interest in medical sciences for as long as I can remember, but thinking back, it is difficult to identify a specific triggering event. A key period, however, was the inspiring environment created by my high school chemistry teachers, who motivated me to participate in the qualifications for the International Chemistry Olympiad and attend summer research school at Karolinska Institutet. During a 6-week period, we attended lectures and completed our own project work. I was lucky because I was allocated to a lab at Huddinge Hospital where, for the first time, I came in contact with patient material, taking part in isolating and characterizing to what extent leukocyte populations from chronic myeloid leukemia patients responded to various mitogenic stimuli. The hypothesis was that this could be used for prognosis of treatment response. Although at the time I was too young to understand what a privilege it was for me to work alongside practicing clinicians on their research project, I later realized the importance of these opportunities for young students. Last year, the summer research school celebrated its 30th anniversary, so I must have attended one of the first.1
Initially, the very positive experience at Karolinska Institutet made me regret my choice of university studies because I had decided on a master's degree in chemical engineering, whereas after the summer research school, I was keen on doing medical studies. However, given my later career choices, I am now very glad that I received a rigorous education in chemistry and an understanding of the fundamental aspects of the discipline. I believe this is vital for any biochemist who wants to work with small molecules and drugs. My interest in medical sciences remained, and I completed my university degree with diploma work at Karolinska Institutet in the lab of Torleif Härd. Torleif has a very strong structural and biophysical chemistry background, and I stayed in his lab to complete a thesis in biophysical chemistry, specializing in fluorescence spectroscopy and isothermal titration calorimetry. This background led me to a post-doc position in the lab of John Ladbury at University College London, who sent me for a short sabbatical to the lab of late Prof. Emeritus Julian Sturtevant. In both of these environments, my interest in the driving forces behind biomolecular interactions grew stronger.
You have worked at Pharmacia, Biovitrum, GE Healthcare, and now you are back at Karolinska Institutet, where you obtained your PhD. How do the scientific environments differ between academia and pharmaceutical and instrumentation companies? Did you need to make adjustments to move between these organizations?
When I had finished my post-doc, I was keen to learn more about the early drug discovery process and to find out whether there was a possibility to apply my knowledge in biomolecular interactions. This led me to join the Pharmacia organization in 1999 as a biophysical chemist. Working in this field for the next 10 years really taught me to pay attention to details as to the action, and often unwanted action and promiscuous behavior, of small molecules. However, this time was also plagued by a large number of reorganizations, mergers, and spin-offs, and I therefore left drug discovery fairly frustrated and somewhat discouraged. My interest in the action of small molecules and drugs was not attenuated though, so when the Chemical Biology Consortium Sweden initiative was funded in 2010, I was among the first to join the founders.
In terms of differences in scientific cultures, I would first just like to say that it has been very valuable to see and work in several different environments, and I think it would be beneficial if there was a way to improve the flow between these careers. Today, this is quite a challenge, as we value them so differently. With that said, one of the things I enjoy the most in academia is that the most important goal is to educate, meaning that a higher purpose than the projects themselves is to create opportunities to learn and teach. I will try to exemplify what I mean with this, as I think it is important and in fact the biggest difference between these environments. This does not mean to say that you are not constantly learning while working in industry; I certainly had ample opportunities. However, in my experience, the education was always directed toward what we needed for the projects, while in the academic environment, the projects are equally important as a means to learn, meaning things are not filtered in the same way. Although this is not always effective from a project perspective, it can well be for the purpose of learning, and it does allow for serendipitous findings.
When I first came back to academia, it took a while to understand how some of the academic investigators prioritized compounds in their collaborations with us. This was because I was so used to always filtering for the best candidates. Instead, we often encountered situations in which the key driver was to select for compounds or unexpected behavior that had not been studied before. It took some adjustment to leave behind the compounds that my colleagues with an industry background and I saw as the most promising. Although the search for novelty can be influenced by an urge to identify potential intellectual property, the main driver from an academic viewpoint is that new findings are more interesting and easier to publish. Now, we are all quite stubborn, so it is not uncommon that we end up with two parallel lists of prioritized compounds; I think this is a great argument for mixing people with different backgrounds. At a minimum, this creates opportunities for surprises that will not show up when things are run strictly according to a single set of criteria that must be fulfilled to get a lead or a candidate drug, regardless of how well-founded they are in each specific case.
You have been a key member of Chemical Biology Consortium Sweden (CBCS) since it was established in 2010. What led to the founding of CBCS? Have there been notable successes or challenges? To what extent is the work of CBCS coordinated or integrated with that of the Europe-wide network EU-OPENSCREEN?
The starting point for CBCS was the shutdown of small molecule research at Biovitrum, which overnight made a significant number of medicinal chemists redundant. This naturally led to many different initiatives focused on the most promising projects. Inspired by the NIH Molecular Libraries effort in the United States, one idea was to build on the Biovitrum know-how and assets and make those available to academics in Sweden. Five medicinal chemists (Annika Jenmalm Jensen, Anna-Lena Gustavsson, Lars Hammarström, Lars Johansson, and Martin Haraldsson) teamed up with the well-established academic screening site at Umeå University under the management of Mikael Elofsson and the Uppsala Drug Optimization and Pharmaceutical Profiling unit led by Per Artursson. They joined forces to enable two critical achievements: one was to obtain long-term financing from the Swedish Research Council in order to establish a national infrastructure for chemical biology, and the other was to convince Biovitrum to donate valuable small molecule libraries and equipment. This donation was a fantastically generous and constructive way to make the most out of the downsizing process at Biovitrum.
One of the things we are most proud of at CBCS is that we have managed to establish a project model that operates in a truly objective and collaborative manner with research groups from all Swedish universities. We realized very early on that our work would span a large area of biology and that projects would extend beyond the scope of our own expertise. It therefore became essential to retain every principal investigator (PI) and their students and post-docs as an integral part of each project. In practice, we achieve this by working alongside the groups in their labs or in our labs, with our minds focused on transferring knowledge in both directions. In this way, we fulfil one of our most important criteria for success, which is to educate in the field of chemical biology. We also believe this close collaboration helps us to stay on track within the intent of the original project proposal. The operative model has been very successful. Along the way, we have also become part of the Science for Life Laboratories. SciLifeLab serves as a national resource for life sciences research in Sweden, with a particular focus on health and environment. CBCS is now 1 of 9 platforms at SciLifeLab that collectively provide services and access to technology not generally available in academic environments.
CBCS has been involved in the preparatory phase for EU-OPENSCREEN, particularly our node at Umeå University, which has been responsible for the education work package. During the preparatory phase, our internal project activities have not been coordinated with EU-OPENSCREEN, but they will be in the future for the participating countries. However, a key challenge at present for CBCS is to secure continued long-term funding. As in many other countries, the national research council is forced to take tough decisions in times of limited resources, and unfortunately this has resulted in a decline of our funding for 2016 and 2017. In practice, this also means that Sweden has yet to decide formally on membership in the operative EU-OPENSCREEN organization.
Following 6 years of operation, CBCS has published more than 80 papers, been involved in the filing of more than 10 patent applications, and has taken part in 5 projects that attracted external funding for clinical studies. Over time, we have realized that it can take as much as 4 years for our collaborations to result in publication, which means that there are many interesting projects that are yet to deliver. We foresee a good year in 2016, including one example of a phenotypic screen in embryonic stem cells. The screen and follow-up work have led to the identification of low nM compounds and elucidation of the molecular mechanism of action. This project is very important for our internal discussions and prioritizations, as we see a lot of phenotypic-based proposals and know that it is difficult to truly deliver on them.
On a more personal note, I would like to highlight our recent work on the screen formatting of the cellular thermal shift assay, CETSA. We recently published data from the first screen done in 384-well format on thymidylate synthase as a model system, and were very pleased with both the performance and outcome of this new method, which confirmed known drugs and also identified several new inhibitors. This project naturally appeals to my background as a biophysical chemist, and I was fortunate to be asked to do a pre-review of the original 2013 Science paper from the group of Prof. Pär Nordlund. One thing led to another, and in the end, this has been a very fruitful collaboration. At present, we have several new collaborations working on microplate formatting of different CETSA variants for screen purposes.
What about your role in the Laboratories for Chemical Biology Karolinska Institutet (LCBKI)? Do you perform assay development and screening as a service to academic “clients,” or are your projects more collaborative in nature? Do you also conduct your own research?
I have several additional roles at CBCS besides working as a senior scientist in assay development and screening. One of my great pleasures is that I still get to take part in the laboratory work, including running some of the screening campaigns, although I wish I had time for more of this. As I mentioned earlier, all projects are collaborative in nature. In practice, the first steps in assay development are nearly always completed by the collaborating group. We engage with final assay validation, and this could possibly also include transfer of the assay to a higher density microplate format before we run the screens and perform the hit confirmation. At this stage, the collaborating group is re-engaged, and the hit confirmation results are compared with those observed in the PIs' lab. When it comes to later-stage programs that involve chemistry optimization, they are nearly always run in both labs, with either us doing chemistry and the PIs' lab running the assay or, in some rare cases, the other way around. The actual setup of each project very much depends on the capabilities and experience of the collaborating group.
Another key role that I very much enjoy is taking part in all the first meetings with PIs at LCBKI. In these meetings, I present our working model, and we start a discussion on feasibility and relevance for a potential collaboration. These meetings are generally extremely rewarding, and there is usually significant common excitement. This work also fits well with one of my most important job duties—responsibility for the Project Review Committee (PRC) at CBCS. All major activities, such as a large-scale screening campaign or structure activity relationship (SAR)-driving chemistry, must be evaluated and prioritized by independent peer review. This means that we do not decide ourselves which projects to run; instead, this is done by the PRC, which is composed mainly of representatives from universities not affiliated with CBCS. We see this as a great way to make sure that prioritizations are objective, and we also believe that this is critical in making CBCS a national rather than a local resource.
You have spoken and written about the importance of assessing cellular target engagement to validate small molecule probes. Tell us why this is so important, and the work that you and others have done in this area. How has your training and expertise in biophysical chemistry helped to guide you?
I think that the major reason for my interest in measuring binding of small molecules to the intended target protein within cells comes from numerous disappointments downstream of screening campaigns. When I came to the Pharmacia organization as a biophysical chemist, I was naively unaware of the phenomenon of false positives, because prior to this, I had only studied well-validated interactions in academia. Facing the outcome of many screening campaigns therefore became quite a hard learning experience. For a while, I felt that we discovered one artifact after another, for example auto-fluorescent compounds binding to antibodies in fluorescence polarization assays followed by compounds that induced aggregation of the target protein, or compounds that in the presence of reducing agents oxidized the target protein. At the time, data were emerging from Brian Shoichet's lab on promiscuous aggregators, and this helped tremendously. Still, we had instances where the wrong hit compounds were prioritized just because they gave an annoyingly good but misleading effect in a follow-up phenotypic assay.
It is likely that most drug discovery labs had their fair share of this, and we often argued for the urgent need for “close to target” assays alongside the phenotypic assays that were used as an early predictor of potential clinical outcome, which in our case could be, for example, fatty acid oxidation or glucose uptake assays, as we worked in the field of metabolic diseases. At the time, we did not even consider the possibility of doing direct binding assays with library molecules in live cells, so when Pär Nordlund first showed me their protein denaturation studies in cells, I immediately felt that this could be an effective way to clean the hit lists. After all, the only thing we want to do is to focus on the right compounds rather than waste time on things that unfortunately can look attractive, even in some of the phenotypic follow-up assays. Working with CETSA has also prompted me to look at other ways of studying target engagement in live cells, and now there are several ways to do this in microtiter plate format, including tagged proteins with, for example, a Nanoluc tag or beta-galactosidase fragments to simplify the detection.
In my opinion, the true benefit of having these assays in a microplate format is that it allows for studies of target engagement for all members of a compound series with full concentration–response curves. Rather than proving target engagement for a single compound, this helps us to compare the SARs with those observed in both the biochemical assay on the isolated target and the cell-based assay with desirable phenotypic readout. This is important because even if target engagement is demonstrated for a few compounds, there is a significant risk that the phenotypic response is still affected by potential off-target effects. However, in cases where the SARs match up, it becomes more likely that the effects are mediated primarily by pharmacological modulation of the primary target.
Are there other exciting future directions in your own research or in the drug discovery field generally that you would like to highlight?
One of the areas in which we as an academic screening lab have invested significant resources is the field of phenotypic assays in primary cells, which in some cases come from the nearby hospital. These activities come with a joint responsibility with the collaborating academic group to understand the mechanism of action of hit compounds, focused primarily on those for which we manage to develop a reasonable understanding of the SAR. I already mentioned above a very successful example using embryonic stem cells that we hope to publish this year. In this case, the target deconvolution activities were largely governed by the structures of the hit compounds, which pointed to a certain target class. We have also been engaged in collaboration with AstraZeneca, in which we share non-competitive knowledge in the design and synthesis of generic reagents that can be used to derivatize hit compounds into baits that we apply for affinity-based target fishing. For anyone interested in these approaches, there is a recent review published on chemical proteomics from Giulio Superti-Furga's lab.2
I would also like to highlight a recent example of a successful hit identification published in Nature Chemical Biology.3 In this study, a group led by Matthew Meyerson investigated the mechanism of a hit compound from a phenotypic screen, and did so by exploring different sensitivities to the compound in a broad range of genetically characterized cell lines. This led to the identification of phosphodiesterase 3A (PDE3A) as the putative target, which was confirmed by several different approaches. The interesting twist to this story is that other potent PDE3A inhibitors were also tried, and not all of them created the same cytotoxic phenotype; in fact, they were instead shown to rescue the cells from the action of the hit compound. Substantial follow-up work showed that the distinguishing factor between the different PDE3A inhibitors was their differential effect on the proteins recruited to the multiprotein complexes involving PDE3A. This study serves as an excellent example of the need to apply a broad range of different technologies to identify not only the target protein of phenotypic hits, but also their mechanism of action.