A unique method for identifying bioactive chemical compounds produces a potential new drug molecule

31.01.2019

A new inhibitor (magenta) covalently bound to MKK7 (gray), as validated by X-Ray crystallography

Our cells receive more messages than a teenager’s phone and, just as with those phones, we sometimes need a way to screen harmful messages – preferably before they are viewed. Dr. Nir London, his group in the Weizmann Institute of Science’s Organic Chemistry Department and colleagues used new computational methods to identify a molecule that enabled them to interrupt a harmful message in the cell – before it could be “read” in the cell’s nucleus, and without interfering with other messages needed for the cell to function. The findings of this research were recently published in Cell – Chemical Biology.

Once a signal is taken in through the cell’s outer surface and transferred inside, the information is relayed through a series of proteins down one of the cell’s signaling pathways until it reaches the nucleus. The signals are mostly those that direct bodily functions, but they also play numerous roles in disease – from inflammation, to autoimmune and metabolic disease, to the erroneous growth signals in cancer; so these pathways are attractive drug targets. Among other things, the pathway London and his team investigated, known by the initials JNK, mediates the cell death that follows the buildup of plaque in the brain in Alzheimer’s disease, as well as playing a role in inflammation.

Goldilocks links

The signaling molecules in this pathway – proteins called kinases – pass their messages through chemical interactions in which each attaches a phosphorous group to the next kinase in the chain. The “upstream” end of the chain may include a number of kinases, collectively called MAP3Ks. But the next link down consists of just two kinases – MKK4 and MKK7 – and these pass the signal to the final link – the protein JNK that gives its name to the pathway – which passes the message on to the transcription factors that turn genes on and off in the nucleus. “This is the way that many signaling pathways are designed, narrowing down to two kinases and then a bottleneck in the last step,” says London. “We used this feature to home in on the best place to interfere with the signal. With between five and ten MAP3Ks at the upper end, where there is redundancy, you would have to block them all. And that JNK at the downstream end performs a number of functions in all the body’s cells, so targeting this kinase may create serious side effects. The MKKs, we think, are the ‘Goldilocks’ links in the chain. Targeting just one could reduce the harmful messages without completely shutting down the pathway.”

The researchers focused on one of the two – MKK7 – a kinase that has no known inhibitors and which is not entirely understood, in part, because there has not been a way, until now, to untangle its actions from that of its sister protein, MKK4. To their delight, they noted an amino acid called cysteine sitting in a location on MKK7 where cysteines are rarely found in kinases. Cysteine, they knew, can form chemical bonds that are strong and irreversible (in chemical terms, they are covalent). Thus a molecule that binds to this specific site and forms such a special bond with this cysteine could be highly selective, producing few side effects.

(l-r) Dr. Nir London and Amit Shraga

London had previously developed a method to discover biologically active compounds for protein inhibition – on a computer. That is, he can reproduce today’s high-throughput methods – done with large robotic machinery and actual compounds – in silico, screening more than 100,000 molecules at a time. The method even enables its users to evaluate molecules that don’t yet exist.

In the study, research student Amit Shraga and other members of London’s group worked together with the group of Dr. Ziv Shulman of the Institute’s Immunology Department; Dr. Bruce Lefker and Dr. Chakrapani Subramanyam, visiting scientists from Pfizer Research and Development, USA, and their colleagues in the Nancy and Stephen Grand Israel National Center for Personalized Medicine on the Weizmann campus; Prof. Takayoshi of the Osaka Prefecture University, Japan; and Dr. Robert Hudkins of Teva Pharmaceuticals, Inc.

The method even enables its users to evaluate molecules that don’t yet exist

After the researchers had identified around ten promising compounds on the computer, the next step was to test them in lab test tubes. Of these ten, three worked well in the lab – but one of them bound to the kinase with extraordinary efficiency. Next, the group examined a “docking model” produced by the software, to understand how this efficient compound binds to the cysteine region; they then created variations of the compound to see if it was possible to improve it further.

When the compound molecule was prevented from forming the covalent bond in lab experiments, there was no inhibition – proving that this binding was, indeed, responsible for interrupting the signal. Next, the researchers inserted this molecule into immune cells taken from mice that were then exposed to a substance known to trigger the JNK pathway. Here too, they found the molecule to be highly efficient at preventing the inflammatory response of the cells.

Atomic resolution

Finally, in collaboration with the Japanese lab, the group crystallized the molecule together with its target kinase and determined their structure – a process akin to taking a photo of the interaction, only at atomic resolution. They found that the binding structure was exactly as their model had predicted. When the compound was later tested in lab tests against some 76 different kinases, only two or three reacted in any significant way. This was further evidence that the side effects of a drug based on this molecule could be relatively light or nonexistent.

The molecule London and his team have identified could conceivably have therapeutic applications in preventing unwanted cell death in such diseases as Alzheimer’s, or in reducing inflammation in such disorders as irritable bowel disease or diabetes. “Because the molecule binds covalently to the kinase, it continues to inhibit that protein for its entire lifespan. This means that if the molecule proves itself in clinical trials – which are still a long way off – it could be given in small doses or infrequently,” says London. YEDA Research and Development, the technology transfer arm of the Weizmann Institute of Science, has applied for a patent on the molecule.

“Now that we have an inhibitor for MKK7, we can begin to tease apart the separate functions of the two intermediate messaging proteins in this pathway,” says Shraga. “This will ultimately help us figure out how to let signals travel down this pathway properly, without passing on any unwanted messages.”

Dr. Nir London’s research is supported by the Helen and Martin Kimmel Center for Molecular Design; the Rising Tide Foundation; the Joel and Mady Dukler Fund for Cancer Research; and Virgin JustGiving. Dr. London is the incumbent of the Alan and Laraine Fischer Career Development Chair.