Robert Huber (2016) - Protein Structures in Translational Medicine and Business Development, My Experience

Having seen so many beautiful molecular structures, I thought I should begin with a historical slide. It is a slide showing the people who are to be regarded as the fathers of structural chemistry. The date of birth of structural chemistry, that is the discovery of a method to see atoms in crystals and molecules, can be precisely dated to the observation of X-ray diffraction by crystals, by Max von Laue, of course. He used X-rays and his work required X-rays which had been discovered by Röntgen. And then proper use had to be made by X-ray diffraction and that is due to father and son Bragg who worked in Cambridge, England. And last then the discoverer, the father of protein crystallography, who also worked with the Braggs in Cambridge, England. What I would like to point out to the young people is that these are the 2 original instruments that Röntgen used for their experiments and Laue. They are on display in Munich at the Deutsches Museum. So if you have time then it’s worthwhile to see these instruments that revolutionised science and technology. Now, when Perutz and his associate Kendrew determined the first protein crystal structures, that was in the late ‘50s, early ‘60s, there was slow growth for several decades as you can see from this graph. And then from about 1990 on we see exponential growth. So what was the reason for this progress? It was the technological advances. And it was the recognition that understanding biology at the molecular level required to see these molecules and the applications in medicine. This latter point I would like to focus on in my lecture today. Well it is, I think, more than just a happenstance that the progress or the development of my own institute, which was the Max Planck Institute of Biochemistry, which was inaugurated in 1972, somewhat resembles the growth rate of protein crystallography. It was an institute in the forest, as you can see, in the outskirts of Munich. And some called it Martinsruh. It was really a quiet place. This is how the Campus Martinsried looks today. This is the original institute, the second Max Planck Institute had been added and a number of university institutes have been built on the campus. And here, number 3, is the Innovation and Founding Centre, an incubator for biotech companies. And I will come to that later. What use can we make of protein structures for medicine? Of course, I could have chosen examples from many different protein classes, perhaps except the photosynthetic proteins which, as Hans has said, have probably no medical application. Now, I choose proteolytic enzymes. These are enzymes that catalyse the very simple reaction of degrading by proteolysis, proteins, a simple chemistry. Yet we do have, as you can see here, a large number of different proteases in humans, in higher organisms, even in bacteria, more than 600. So why? For such a simple digestive reaction? Now, the reason is again easy to see because proteolysis, in particular limited proteolysis, is a major regulatory mechanism. So we started to work on proteases early on, in the early ‘70s, and then focused on the aspect of the regulation of proteolytic activity. Proteases are in fact quite dangerous because if left uncontrolled, they would digest cells, even organs and whole organisms. So they need careful control and, of course, this is also of great relevance as you will see for medical application. Now, this cartoon shows the various kinds of protease regulation that we found during our more than 3 decades long work in that field. I have no time to discuss these various mechanisms but just begin with the simplest one. That is we do have an enzyme, a protease, and we do have a natural inhibitor which binds very tightly and specifically to its enzyme and, of course, then prevents binding of substrate. This was the first structure in the early ‘70s. A simple digestive enzyme trypsin and the basic pancreatic trypsin inhibitor, tightly fitting, no structural change when we compared the complex structure with the structure of the individual components. It was the first high-resolution protein-protein interaction that was seen here. BPTI itself would deserve to have an hour's lecture because of its role that it played, not only in this system and in crystallography and explaining protease regulation but also it was the model compound for developing protein NMR. And it was the model compound for developing molecular dynamics. I hope that Martin Karplus will say a few words on that aspect. I would like to stay with these simple proteases which are called 'serine proteases', to which the digestive enzyme trypsin belongs, and now focus on a quite complex physiological cascade, namely the blood coagulation cascade which leads from a vessel injury to a thrombus. It is a cascade of limited proteolysis reactions which are carefully controlled. And we do have inhibitory components for most of the steps. But, as I said, essentially the main process is limited proteolysis by an upstream protease which cleaves and activates a downstream protease. I don’t have to say that the blood coagulation is of great medical importance, with the focus right now on the terminal protease which is called thrombin. So we did work on this enzyme. And this already shows the basic principle of structure-based drug design. So starting with a Hid peptide and seeing how it fits and then changing the chemistry such that there is a better binding. In fact this was successful by simply exchanging a phenylalanine residue by a naphthyl group. Now, we can also have a look at nature’s solution to this problem. These blood-sucking animals are actually bags full of coagulation inhibitors, mostly thrombin inhibitors, and we looked at their structures. So this, the blue object, is thrombin and the coloured objects are these natural inhibitors. Most prominent is hirudin from the leech which actually is a clinically used drug. And also a variant, a semi-synthetic variant, has been made by making use of this C-terminal tail of hirudin and linking it or derivatising it with a small chemical that attacks the active site residue. So we can look at nature and derive from it new design principles for inhibitors. Thrombin has another phase, too. Thrombin, of course, is the major coagulant. It cleaves fibrinogen, as I said, when it is activated but it is also an anti-coagulant. And it is an anti-coagulant when there is another protein around to which it binds which is called thrombomodulin. This is how thrombomodulin looks like. This is the binary complex of thrombin and thrombomodulin. Now what happens? Why does thrombin become an anti-coagulant? It becomes an anti-coagulant because now this binary complex is able to bind protein C which is also a blood coagulation cascade component. And only in the presence of thrombomodulin protein C is stably bound. And when it is bound and it’s processed by thrombin the proteolytic activity of thrombin, becomes active and then acts as a degradative enzyme in the coagulation cascade and so stopping coagulation. Well, there is another aspect which is quite interesting. Thrombin and all other natural serine proteases are activated by a limited proteolysis step themselves. They are synthesised as inactive precursors which is, of course, absolutely required, otherwise they would digest the organ, the pancreas where they are made. So they are made as inactive precursors, as zymogens and then activated, a limited proteolysis step. The mechanism, the basic mechanism, is such that the new N-terminus, which is generated by this limited proteolysis step, then folds into the centre of the molecule and causes a dramatic structural change which activates the enzyme. Now, nature is clever and pathogenic bacteria, staphylococci, misuse this mechanism by making a virulence factor that you see here which embraces the inactive prothrombin in a way that it sticks its N-terminus into the hole that is actually used to activate by limited proteolysis thrombin. So we had discovered this mechanism many years before we had the structure available. We called it mechanism of molecular sexuality. And we were really surprised to see it perfectly realised in nature by the staphylococci. Staphylococcal infections are very dangerous if not treated, they cause disseminated coagulation and actually the death of the patient. This is another development, structure-based development of a protein in the coagulation cascade. This is factor 10a, again we had determined the structure, it’s a membrane associated protease. And the company Bayer has developed a small molecule on that basis, which has become a blockbuster and a billion euro business right now. I think I will skip this. Just telling you that the principle of lock-and-key inhibition of the serine proteases, as exemplified with thrombin and trypsin, is also seen in the other major classes of proteinases with cysteine-like proteinases, the metalloproteinases, though there are natural inhibitors that prevent and inhibit these proteinases in a lock and key and in a substrate-like manner. A few words about proteases in physiology and in pharma development. This is the dipeptidyl peptidase, a protease port and an entry port where the substrate enters. Now why is this interesting? It’s interesting because it inhibits insulin secretion, or insulin synthesis actually, which is induced by incretins, these are gut hormones which are then in the pancreas inducing insulin synthesis. Dipeptidyl peptidase inhibits or cleaves the N-terminus from these 2 incretins, makes them inactive. So this offered a novel way to treat diabetes, and you see from the list of molecules and of big pharma that are working in this field. Of course, all on the basis of the crystal structures that we had determined years ago. The same is true with another protease, that is membrane-associated and it acts both intracellular and extracellular and has an important physiological function in membrane transport. But also it is misused, again, by pathogenic bacteria and viruses which use furin as a receptor and also they require a proteolytic cleavage step by furin. Now this is how furin looks like, the protease part and an associated domain. It has a rather specific specificity, namely it cleaves after polybasic stretches of sequence. The fact that it is required by pathogens then, of course, suggested to develop this as an antibiotic, furin as an antibiotic target. In fact, these mouse experiments clearly show that you can rescue mice from serum 1 as exotoxin intoxication. They die after 2 days and when you give a specific furin inhibitor they survive. Still, there was no development after these first mouse experiments, probably or simply because of toxicity effects. I told you about the problems and the role that furin plays in human physiology. As I said I skip all of these different regulatory mechanisms that we had found in this almost 30 years of work and jump to the last one, which was a novel regulatory mechanism not seen before. It is a large intercellular protease, called the proteasome, which is made out of 28 subunits. It has the active sites inside the cylindrical molecule and substrate cannot enter because the entry gates are closed. So regulation there is by opening and closing the entry gate. The role of the proteasome is manifold. Of course, it plays a most prominent role as the executioner of the ubiquitin conjugation pathway. It is also the waste cleaner. It is essential in immunology because if the proteins it degrades in cells are from viral origin or foreign, then the peptides generated are antigenic and transported to the cell surface to trigger a T-cell immune response. These were the first structures that we determined, first the simple one from archaea. It has a conserved architecture, as you can see, 4 rings, 7-fold symmetry. But the archaeal enzyme is chemically much simpler. You saw the example Hans has described of the archaeal reaction centre compared to photosystem 2. The same we see here again when we compare the archaeal proteasome with the yeast, the eukaryotic and yeast and mammalian proteasome. Now, in that all the Alphas are different and all the Betas are different but still the architecture is maintained. The main finding concerning medical application was by serendipity in a small American company. They found that the simple boronate peptides offer a new strategy against haematological disorders, multiple myeloma. Now, we found it was quite clear what it does: it inhibits the proteasome by forming an ester bond between this boronic acid and the active site N-terminal threonine which is, as I said, located inside the molecule. This spurred then enormous activity worldwide. People searched for novel chemical entities, also as proteasome inhibitors. We were approached for many years by people who had discovered novel chemical entities and they wanted to know how they bind. Now, this is a very small part of that story of this collection. What you see here in yellow were new chemicals found in natural compound libraries. It’s quite surprising that natural compound libraries are extremely rich in proteasome inhibitors. Now, they all look different from a distance and they are different, except that they all do have a head group, 'a warhead' it’s called, which chemically reacts with the N-terminal threonine. This was the example where we were directly involved, a collaboration with a plant biologist who had discovered that this plant pathogen which kills bean plants needs a virulence factor now of this formula. What we found is that it inhibits the plant proteasome, leads to apoptosis and finally the killing of the bean plant. It was again a novel chemical entity synthesised and modified and developed further. The problem with the proteasome as a drug target is that, as said, it offers a novel way to treat haematological disorders but it shows disastrous side effects, in particular against nerve cells. It was known that hematopoietic cells do have a slightly different proteasome which is called immunoproteasome - slightly different only but somewhat more active and somewhat different specificity. So we thought if we are able to specifically target the immunoproteasome then this might help to avoid the toxicity of inhibiting the constitutive proteasome and even it would offer a possibility and a new strategy against autoimmune disorders where one would like to reduce the immune response. Of course, this is already work that is beyond what a at that time small academic group, I was already retired, can do. This cartoon shows the problem, the basic problem of the translation of academical results into industry, the translation, the financing gap. Fortunately we do have in the Max Planck Society a daughter that was set up in order to advance ideas or materials from the academic Max Planck Groups to a further stage such that it becomes interesting for industry. So I suggested to them the proteasome and in particular the immunoproteasome and they accepted it. We still continue to work with these limited resources on basic sciences, basic science of the immunoproteasome. We found out, for instance, that when we reanalysed these 50 or more different proteasome complexes and compare them with the complexes from the immunoproteasome, we found out by principle component analysis assisted by molecular dynamic simulations, that the constitutive proteasome upon peptide ligand binding undergoes a small structural change, which, of course, is energy-costly while the immunoproteasome does not require this conformational change and immediately explains the higher activity of the immunoproteasome. So that directed, of course, the process of finding specific immunoproteasome inhibitors to exclusively follow the peptidic model. Of course, also the LDC can’t do without the partner and they joined a strategic partnership with big pharma. And there is substantial progress now with animal models, rheumatoid arthritis which is an autoimmune disease and systemic lupus. Now, there is little time left, but in the last few minutes I would like to mention my direct involvement in the foundation of business. That goes back to 1999 when a former doctoral student, a postdoc and myself founded a company. It was just a few people at that time, 3 or 4 people. Now it has grown to about 70 members. They have their own building in the Munich area and what they do is they provide services, enabling technology services and lead discovery services to big pharma. This is the work flow and they have some special instruments, namely to improve crystalline quality. I mean, this has not really been said so far but, most of the crystals that we get are weakly or badly disordered, show disorder diffraction. But there is a way where we can improve the crystal quality by changing in a very controlled way the humidity. And more recently also finding out that we can simplify this by using an infrared laser. What happens then is that there is a slight shrinkage of the crystal which in a number of cases then improves the crystal quality from 3.5 angstrom resolution to 1.7 angstrom resolution. So that’s quite easy to apply. Now, this is the list of the customers this service company has, which is in principle quite surprising because all of these pharma companies do have their own structural biology groups. Yet they approach service groups like Proteros and buy these services because they are cheaper in a sense and perhaps faster and better. So there is room for these service groups, even in this pharma environment. Now, the most interesting part – well, I co-founded with again 2 doctoral students and postdocs in 2003 a company called SuppreMol which works on therapeutic proteins for autoimmune diseases. And what the title here says, that the recent outcome, which was 1.5 years ago, made investors very happy, made lawyers and brokers happy and founders also happy for a certain reason. So the work actually goes back to work that we did 40 years ago, namely antibody structures. So this is a complete antibody with its Fab-arms, with which it recognises the antigens. And this is the stem part, so to say the business end of the molecule. What antibodies do is they either trigger the cellular immune response when they are bound to an antigen or the humoral immune response. We were interested in the cellular immune response. So this is a cartoon that shows what is happening. There is the antigen opsonised, so with a large number of antibodies bound. And this opsonised antigen reacts with receptors on immune cells and triggers either activation or inhibition. So this system requires an extremely careful balance of activation and inhibition, therefore you have 2 different kinds of receptors which, by the way, look very similar. Now, what we did was looking at the structures of the receptor and of the Fc part, then having seen that we thought we could make use of it. In order to modulate immune response we thought we’d make use of a soluble Fc receptor. The cartoon here shows what the business idea was, namely there is an antigen opsonised. The soluble receptor is around which then binds to the stem part, to the business end, and this, of course, is no longer able to bind to the cellular receptor. So we inhibit the immune response. This is what the idea was. We had another approach by using specific antibodies. I have no time to say that, of course. The group, the SuppreMol group, had already grown to about 20 people, it needed investment. At the end it had an investment of about €40 million after these 10 years. The investors stayed on board. They, at the end, started to become impatient but, I mean, the animal pre-clinical experiments and also the first phase clinical experiments were very promising. You see it on this curve but then, fortunately, it happened and a big American company bought the SuppreMol for a large sum of money. As I said it made investors very happy. One of the investors was the Max Planck Society even. And it made the founders happy, not because of money, but because of the fact that now the project continues under the umbrella of big pharma. Even the team stayed together, even in the premises they are. Well, this is the last slide I would like to show and hope that I have shown in particular with these 2 examples of spin-offs, there is satisfying life outside academia. And you can do work there for the benefit of mankind also. This is my home town, Munich, this is the old part of the university. And this is where Röntgen, who was Professor of experimental physics, and Laue worked at the time of Laue’s key discovery. Thank you.

Robert Huber (2016)

Protein Structures in Translational Medicine and Business Development, My Experience

Robert Huber (2016)

Protein Structures in Translational Medicine and Business Development, My Experience

Abstract

My lecture will start out with very brief remarks on the history of protein crystallography and continue with our studies since 1970 on proteolytic enzymes and their control. Proteolytic enzymes catalyse a very simple chemical reaction, the hydrolytic cleavage of a peptide bond. Nevertheless they constitute a most diverse and numerous lineage of proteins. The reason lies in their role as components of many regulatory physiological cascades in all organisms. To serve this purpose and to avoid unwanted destructive action, proteolytic activity must be strictly controlled. Control is based on different mechanisms some of which I will discuss and illustrate with examples of systems and structures determined in my laboratory: a) by specific inhibition with natural and synthetic inhibitors, b) by enzymatic specificity, c) by activation from inactive precursors accompanied or not by allosteric changes,d) by co-localization of enzyme and substrate, e) by cofactor binding accompanied or not by allosteric changes, f) by controlled access to the active site and g) by substrate-induced allosteric changes of the active site associated with oligomerization.The regulatory principles unveiled by structural studies offer new opportunities of intervention for therapeutic purposes as illustrated with the essential intracellular protease, the proteasome.

I then will let you share my experience with the foundation and development of two biotech companies with different business models, but both based on basic academic research in structural biology: Proteros (www.proteros.com) offers enabling technology services for Pharma- and Crop science companies imbedding all steps of the workflow molecular and structural biology can provide and commands and uses its platform for the generation of leads from identified targets to in vivo Proof of Concept (PoC). Suppremol (www.suppremol.com) specializes in the development of novel immune-regulatory therapeutics for the treatment of autoimmune diseases on the basis of a recombinant, soluble, non-glycosylated version of the human Fcy receptor IIB and of receptor binding antibodies. Suppremol was recently acquired by Baxter International Inc. (NYSE:BAX) offering an ideal setting for its therapeutic projects.

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT


Related Content