Great things start small: an insight into fragment based drug discovery

Have you ever noticed that great things almost always start small? Take for example Richard Branson, he started the Virgin brand with a student magazine!.

We want ‘big’ things, we have ‘big’ aspirations, we want to jump ‘big’ too soon without thinking that maybe starting from small is the best way to go.

Nowadays, drug development is adopting this thought as well: start small. And this can be achieved by screening for small chemical fragments that bind only weakly to a biological target and only then growing them or combining them to produce a compound with higher affinity. This method is known as fragment-based drug discovery (FBDD) and the main topic of this post.

Think small, think FBDD

Developing a new drug is a complex process. Once the target has been chosen, the process of lead discovery begins. A lead is a compound from a series of related compounds that has some of a desired biological activity. A variety of methods to identify hit molecules exist, but the most commonly used is high-throughput screening (HTS), in which libraries with up to millions of compounds with molecular weights of around 500 daltons are screened and nM binding affinities to target are sought. But could we think ‘small’? Smaller molecules, smaller affinities? Definitely YES.

In contrast with HTS, in FBDD typically about 1,000 fragments are screened using biophysical techniques. The fragments are between 150-250 daltons and mM affinities can be considered useful. After biophysical analysis, candidate fragments are grown to form new interactions using structure-based drug design. Although initial fragment hits have low potency due to their small size, they form high-quality interactions and can be readily optimized into potent lead molecules.

Does size matters then? YES, because the size, complexity and physical properties of small molecules can be more easily controlled than when starting from higher affinity HTS hit.

How can fragment binding be screened?

Biophysical methods to screen fragment binding to its target include NMR spectroscopy, X-ray diffraction, isothermal titration calorimetry (ITC), surface plasmon resonance (SPR) and mass spectrometry. Which one to use? Each method has advantages and disadvantages in terms of sample consumption, degree of automation and assay complexity, but the general rule is that whatever method you choose, it has to be FAST, EFICIENT, PRECISE and to REDUCE THE NUMBER OF FALSE POSITIVES AND FALSE NEGATIVES. Remember that the number of fragments that are frequently analyzed can go from hundreds to thousands, so it is very important to choose well.

A new effective method for fragment screening

An emerging technique has appeared very recently for fragment screening¹. Microscale thermophoresis (MST) has been broadly applied to investigate biomolecular interaction of a variety of drug targets, thus it seemed perfect for trying it in FBDD… and it is.

In a recent fragment screening made by Sanofi, MST was used in an automated manner to screen a library of fragments targeted against the kinase MEK1 from the MAPK signalling cascade, thus with looks into developing anticancer therapies. MST identified multiple hits that were confirmed by X-ray crystallography but not detected by orthogonal methods.

But what I found great about this technology and what marks a difference is related to an important aspect of fragment screening campaigns: The EXCLUSION OF FALSE POSITIVES AND FALSE NEGATIVES. Binding of fragments and small molecules can either stabilize or destabilize target proteins. Destabilization or denaturation of proteins is often accompanied by protein aggregation and MST provides a direct feedback on ligand-induced aggregation and other secondary effects. Thus the information provided by MST prevents false-positive hits from entering later stages of hit expansion, as well as rescuing fragments classified as false negatives.

All in all, this speed technology, used in conjunction with other biophysical techniques has the potential to significantly improve FBDD workflows.

In the early days of FBDD there was some scepticism that it would ever work, for example, “How can such low-affinity compounds be identified?” or “Drug discovery is hard enough starting with a high affinity hit so why go to smaller and lower affinity compounds?”. Today, the consensus has changed. At least nine FBLD projects have led to compounds in phase II or III studies, including one approved drug. The fragment approach has been adopted throughout academia and large pharma and biotech companies, and there are a number of commercially available fragment libraries.

So think about it……… maybe it’s time for you to start small.

“Start small, think big. Don’t worry about too many things at once. Take a handful of simple things to begin with, and then progress to more complex ones. Think about not just tomorrow, but the future. Put a ding in the universe”. Steve Jobs.

1. Linke, P., Amaning, K., Maschberger, M., Vallee, F., Steier, V., Baaske, P., Duhr, S., Breitsprecher, D., Rak, A. An Automated Microscale Thermophoresis Screening Approach for Fragment-Based Lead Discovery. 2015. Journal of Biomolecular Screening. doi:10.1177/1087057115618347

Biosimilars are the next money makers of pharma industry

It has happened to everyone: got sick, went to the doctor, got diagnosed and received that desired white piece of paper named “drug prescription”. We run to the drugstore and just ask straight ahead for what has been prescribed. Until recently not many options regarding price of the drug or effect were available, but now some alternatives are there for us.

GENERIC DRUGS are one alternative. They are copies of brand-name drugs, and by ‘copies’ I mean in dosage, safety, strength, route of administration, quality, performance, characteristics and intended use. They have the exact same ‘active component’ of the brand name drug, which is the part that has the therapeutic function, while the ‘inactive components” which in general do not have any pharmacological effect can be different, and they can include dyes, preservatives, flavouring agents, etc.

Generic drugs are important options that allow greater access to health care… why?? BECAUSE THEY ARE WAY CHEAPER THAN BRAND NAME DRUGS. The latter are developed under a patent, which protect a drug company’s investment in developing the drug. This gives the company EXCLUSIVE RIGHTS to sell the drug while the patent is in effect for up to 17 years. After the patent expires, other drug companies can produce the drug. But generic drugs are able to be sold for lower prices because they are not required to repeat the costly clinical trials of new drugs and generally do not pay for costly advertising and marketing. In addition, multiple generic companies are often approved to market a single product; this creates competition in the market place often resulting in lower prices. Good for us consumers. (Story not told, is that 70 to 80% of all generic drugs are produced by the same companies who make brand name drugs… obviously).

BIOSIMILARS, the new kid on the block, are another alternative. They are biological products (not chemicals such as brand name and generic drugs) highly similar to brand name drugs with no clinically meaningful differences in terms of safety and effectiveness from the reference product. Only minor differences in clinically inactive components are allowable in biosimilar products. They also differ from ‘generics’ in that THEY ARE NOT EXACT COPIES of brand name drugs.

Producing generic small-molecule drugs is relatively simple–it’s like following a recipe with standard ingredients. Biosimilars are much more challenging because living cells (where they are produced) are highly sensitive to their environments, and manufacturers have to create their own, unique process to make these cells to produce an identical outcome to an existing treatment.

Biosimilars have made drug approvals challenging. Generics are approved based on matching chemical structure, but that doesn’t work for biosimilars. Each new biosimilar has to run clinical trials to prove the outcome matches that of the biologic it is imitating, even though it looks structurally different, according to recently announced guidelines from the Food and Drug Administration.

Let’s talk about money

When talking about Pharma business there is a lot of money involved and the opportunities for biosimilars are huge for both manufacturers and consumers. Many leading biologic medicines worth more than $81 billion global annual sales will lose their patent protections by 2020… only 4 more years, that means the war between pharma companies has begun.

Much like generics, biosimilars can help cut drug costs, though the savings are smaller because of their complexity as well as regulatory challenges of getting FDA approvals. Biosimilars cost about $75 million to $250 million to reach the approval stage, versus around $2 million to $3 million for a generic small-molecule medicine.

So far, an inflexion point in the pharmaceutical industry has been seen with the approval of two biosimilars: the Hospira’s biosimilar version of infliximab (a monoclonal antibody used to treat autoimmune diseases) in Europe and the subsequent approval in the U.S. of its first biosimilar in history, Novartis’ Zarxio (which targets Amgen’s Neupogen).

In September 2015, Pfizer  shelled out big bucks–$17 billion to be exact–to buy the much smaller drugmaker Hospira. A big reason for the rather expensive acquisition is to gain access to the company’s biosimilar portfolio, an insight on how important these new drugs are becoming for pharma industry.

What can be expected then? Well, that some of the big firms — including Abbvie, Roche and Pfizer — will be highly impacted by how this biosimilar market shapes up in near future. There are around 11 biosimilars under development to compete with Abbvie’s Humira (infliximab competitor) alone, which loses its patent exclusivity in the U.S. in 2016.

Biosimilars are extending to the rest of the world as well. Last month, Brazil registered the first latin-american biosimilar (filgrastim), one of the 20 existing biosimilars to date.

What’s great about these drugs is that many people will definitely benefit as biosimilars take off. Significant savings for medical programs will be seen, since the worldwide trends is that governments are pushing theirselves for granting marketing authorizations for biosimilars…. I hope less money spend in health means more money spent for good purposes for us all tax-payers.

The Ocean holds secrets about genome editing

It’s being told that the ocean holds many secrets, of lovers and murders, of treasures and ships untouched, that it makes me wonder what other secrets could it possibly hold. We have to thank the ocean for giving us biochemists such special gifts as our beloved GFP (Green Fluorescent Protein), but what other cool proteins can come to us from the ocean?

Everyone has a favorite protein, what’s yours?

I was once asked by a biochemistry teacher: “everybody has a favorite protein, what’s yours?” I think I was way too young to have an answer back then. It’s been over a decade since I graduated and right now I don’t really think I have one favorite protein, but I have a group of favorites and among them are the TALEs. That’s why I think they deserve a post dedicated just to them.

Is it possible to edit the genome?

According to the great Wikipedia, editing is the process of selecting and preparing written, visual and audible media used to convey information. But editing can be applied in other aspects of life as well, even at the gene level. Genome editing is a type of genetic engineering in which the DNA is inserted, replaced or removed from a genome using artificial nucleases or “molecular scissors” or in other words is about re-writing the DNA code.

Genetic engineering emerged in the lab of Paul Berg in 1972 in the form of a recombinant DNA technology, when scientists combined the E. coli genome with the genes of a bacteriophage and the SV40 virus. Since then, this science has achieved tremendous success; many methods to manipulate DNA have been developed as well as vector systems and methods for their delivery to the cell.

From 1990 to 2003, the DNA sequence of human, E. coli, mouse and others was deciphered, retrieving information about nucleotide sequence only. In 2003, the U.S. National Human Genome Research Institute launched a new international project, ENCODE (Encyclopedia Of DNA Elements), which aim was to obtain a complete list of the functional elements of the human genome. Unfortunately, the methods used back then are not only expensive but also quite labor-intensive and they do not allow one to introduce precise changes into a strictly defined genome locus. Currently, researchers have several tools that allow them to solve the problems of precise plant’s, animal’s, and human’s genome editing.

A tale about TALEs

Bacteria of the genus Xanthomonas are pathogens of crop plants, such as rice, pepper and tomato that cause significant economic damage to agriculture. Cool thing about these bacteria is that they were found to inject proteins into the host plants where they mimic eukaryotic transcription factors thus hijacking the host’s transcriptional machinery to control gene expression. These proteins are known as TALEs for Transcription Activator-Like Effectors. But why would TALEs do that? Well, to increase the plant susceptibility to the pathogen, and they do it pretty much like a virus.

One of the most interesting things about TALEs is what I call their “outfit”. It turns out that TALEs bind to DNA in a specific manner and this specificity is given by its DNA-binding domain “outfit” which consists of tandem repeat arrays of amino acids, with each repeat binding a single DNA base. Moreover, The residues located at positions 12 and 13 in each repeat are highly variable and are called the BSR (base specifying residues), which are responsible for the recognition of a specific nucleotide. This TALE code is degenerate in the way that some BSRs can bind to several nucleotides with different efficiencies.

After TALEs were described, two other plant disease associated bacteria were found to encode for TALE-like proteins as well. These proteins are RipTALs from Ralstonia solanacaerum and Bats from endofungal bacterium Burkholderia rhizoxinica. TALE-likes seems to be united only by possession of DNA binding repeats with a conserved code, in fact Bats lack the domains necessary to function as eukaryotic transcription factors.

What can TALEs be used for?

So lets imagine: we have a cool protein, wearing a nice tandem repeat outfit, which holds a special code that allows to bind to DNA in a specific manner. So, what if we knew the code? Well, then one could predict the DNA binding element for any given TALE and to design them to match any DNA sequence of interest. Now THAT is interesting. But it gets even better! Genetic engineering in not called that just for nothing, so the idea is to make these proteins even cooler. Now, what could be done? Specific DNA-binding in one thing, but the function is another thing. So a great idea is to couple TALEs to a functional domain of choice and those chimeras are invaluable tools for precision manipulation of genome.

But we can go beyond: non-BSR polymorphisms might also be useful to tune DNA-binding properties and further expand the diversity of TALE-DNA interactions. One could then create libraries of designed TALEs with a range of binding strengths for the same DNA element, useful for the regulation of synthetic genetic circuits.

The birth of a new pair of genetic scissors: TALENs

After deciphering the code of DNA recognition by TALE proteins, which attracted the attention of researchers across the world due to its simplicity (one monomer – one nucleotide), the first chimeras of TALEs and nucleases were created: the sequence encoding the DNA-binding domain of TALE was inserted into a plasmid vector previously used for creating Zinc Finger Nucleases. This resulted in the generation of genetic constructs expressing artificial chimeric nucleases (TALENs) and this was so amazing that in 2011 Nature Methods named the methods of precise genome editing, including the TALEN system, method of the year.

How to ‘pimp’ your TALE

One approach to TALE repeat engineering is random mutagenesis and screening. Alternatively, mutations could be introduced in a more targeted fashion, but this requires information on the impact of different types of polymorphisms at different positions in the TALE repeat. So, where do we get information on what sites should be good targets?....in mother nature. Natural variation would provide useful information on what residues can or cannot be tolerated at which positions and with what effect. Unfortunately, TALEs sequence diversity is very low and residues clustered around the BSR are largely invariant across all currently known TALEs, RipTALs and Bats. Good news is RipTALs and Bats are only near 40% identical to TALEs, making them useful for repeat engineering.

TALEs from the Ocean

The ocean is quite an interesting place to do research. We often look for answers or work on the model of terrestrial organisms, but the ocean holds many secrets.

On the Gulf of Mexico/Yucatan Channel, during the Global Ocean Sampling (GOS) expedition, biological samples were taken and used for DNA sequencing. Among them, a particular sample was analyzed, likely from bacterial origin based on size filtering of the biological material that was used. During sequence analysis two predicted proteins came into the light: MOrTL1 and MOrTL2 (Marine Organism TALE-likes, names that reflect the limited information scientist had regarding their provenance).

These sequences have been previously suggested to encode modular DNA binding repeats, but no functional analysis had been reported. And guess what, both proteins are tandem repeat arrays … does it ring a bell?

This MOrTLs repeats differ at more than 60% of positions from each other and from all other TALE-likes. Although MOrTL sequences are incomplete and likely to be fragments of larger, incompletely sequenced genes, they were synthesized and MOrTL1 was further expressed and purified from E. coli. As analyzed by electrophoretic mobility shift assay (EMSA), MOrTL1 showed a weak DNA binding, inconsistent with TALE-likes. This would likely reflect the incomplete sequence of MOrTL1, yielding a incomplete functional protein.

Then, how to study a predicted protein, based on an incomplete sequence, if even when expressed it’s not fully functional? Well….  CHIMERAS!

MOrTL repeats were embedded within the repeat domain of a Bat named Bat1. Now there were ready to be studied.

Can Bat1-MOrTL chimeras bind to DNA?

Yes they can! And this was confirmed by mixing Bat1-MOrTL chimeras and their cognate TALE-code predicted DNA binding element and analyzed by two approaches: a qualitative classical EMSA and the quantitative MicroScale Thermophoresis (MST) which can quantify the affinity of the binding and calculate binding constants (K_D) of any intermolecular interactions with high precision. In fact, Bat1-MOrTL chimeras bind with similar strength to the wild type Bat1 protein.

Is this binding specific?

When you prove that a TALE-like protein binds to its predicted on-target sequence, does it prove adherence to the TALE code? NOT NECESSARILY!

How to prove it then? Well, specificity needs to be tested and by this I mean proving the binding to the worst predicted match DNA sequence based on the TALE code. How can this experiment be done? By competition experiments. So when analyzing probe-protein interactions (DNA-binding elements-MOrTLs interactions) with on and off-target competitors it was confirmed that both MOrTL1 and MOrTL2 TALE-code consistence base preference. Additionally, when quantifying the protein-DNA interactions with the off-target probe by MST a very interesting result came into light: the Bat1-MOrTL1 chimera had an affinity 19 times lower than the on-target interaction, while Bat1-MOrTL2 was only two times lower, thus they differ in the discriminating power. The higher discriminatory power of MOrTL1 repeats thus make them better for integration into TALE-like repeat arrays for biotechnological applications. A very nice piece of information.

Could protein stability account for the differences observed?

Functional differences among proteins could be due to stability issues. Protein stability can be analyzed by inducing thermal denaturation and then monitoring the fluorescence of SYPRO Orange, which binds non-specifically to hydrophobic surfaces. So, when the protein unfolds, the exposed hydrophobic surfaces bind the dye, resulting in an increase in fluorescence. This method is called DSF (Differential Scanning Fluorimetry), also known as Thermofluor. But as the music band ‘Outkast’ asks “what’s cooler than being cool?”, well the answer is not “Ice Cold”, it is "nanoDSF". This technology allows to do the same as DSF but taking advantage of the Tryptophan’s intrinsic fluorescence and by using this pretty cool technology, it was shown that the functional difference between MOrTL1 and MOrTL2 chimeras are not due to differences in protein stability.

By demonstrating that MOrTL repeats mediate DNA binding behavior analogous to that of other TALE-likes repeats, scientists have gain insights into the nature of the whole TALE-like family… hopefully this amazing discovered secrets from the ocean will enable further research into the distribution and functions of these fascinating DNA binding proteins.

1.    

11. DNA-binding proteins from marine bacteria expand the known sequence diversity of TALE-like repeats. Nucl. Acids Res. first published online October 19, 2015 doi:10.1093/nar/gkv1053