Dr. Robert Volkmann, Vice President of Chemistry – Mnemosyne & SystaMedic
“What you’ve done is you’ve designed a product that works for people like me. It not only works great for me but also works great for our biologists and our CRO chemists in India. … I just take it for granted that CDD Vault will work and it will work as planned. So I guess for me the test for the value of something is that you can use it without ever thinking about it.”
Robert Volkmann, Ph.D
Dr. Volkmann is a distinguished medicinal chemist with expertise in neuroscience. His laboratory discovered a series of nNOS inhibitors for the treatment of stroke, novel glutamate antagonists, novel p-type calcium channel antagonists from venoms, novel neurotransmitter release enhancers. In addition, his laboratory prepared Englitazone, a PPAR gamma agonist for the treatment of diabetes and Sulopenem, a potent broad spectrum penem antibacterial drug. Dr. Volkmann has over 60 patents and publications and has given numerous scientific presentations. Dr. Volkmann received his Ph.D. in Organic Chemistry from the University of Pittsburgh and conducted postdoctoral research at Stanford University.
Interviewed by Barry Bunin, CEO Collaborative Drug Discovery, Inc.
Listen to the Full Audio
Edited Interview Transcript
Hello, Bob Volkmann. This is Barry Bunin calling for our CDD Spotlight. I wanted to start with the question of why the NMDA receptors are interesting and why your approach in particular is interesting?
The NMDA receptor has been of intense interest to researchers all over the world including those of us in the pharmaceutical industry. Glutamate is the brain’s primary excitatory neurotransmitter and the NMDA receptor is recognized as a key target for biological intervention in neurodegenerative diseases. A key channel blocker, MK-801, was discovered in the early 80’s. NMDA antagonists originally targeted stroke and traumatic brain injury. Interestingly, these compounds showed some efficacy but caused cognitive disruption and psychosis. These studies are over 20 years old and Pharma left the area without discovering any new medicines. More recently, ketamine, an NMDA antagonist, has been found to be active in the treatment of depression. As a result, people are starting again to think seriously about this target.
We and others are looking at subtype specific modulators of the NMDA receptor in the hopes of finding agents with fewer side effects. We have been looking for allosteric modulators of the NMDA receptor and we think that these subtype selective allosteric modulators could have benefits in a number of diseases, and we’re particularly interested in the NR2B receptor and in fact are looking for positive allosteric modulators (PAMS), which we hope will be effective in the treatment of neuropsychiatric disorders, such as schizophrenia. One can also make a case for negative allosteric modulators of the NR2B receptor for treatment of depression. Others have thought about the NR2A receptor for Rett syndrome, which is a genetic disorder that affects brain development, and it occurs almost exclusively in young women. We at Mnemosyne have a research platform centered on NMDA targets and are looking at subtype specific positive and negative allosteric modulators of the NMDA receptor.
We are a virtual company. We have been fortunate to get financial support for our platform. There are others too that are now doing research in this area. I would say, in concluding, that there has certainly been a resurgence of interest in this target, which everyone knows is so important.
I saw in a recent Current Topics Med Chem Review in January that you mentioned allosteric agents and why allosteric versus orthosteric for this particular target…
Well, Barry, so that’s really important because when you have an orthosteric modulator, you’re interacting with an agonist binding site. From a medicinal chemistry perspective, allosteric sites provide an opportunity of finding agents with better drug-like properties and greater target specificity than orthosteric sites. Furthermore, in our case, allosteric modulators are better suited to maintaining the highly precise temporal and spatial aspects of glutamatergic synaptic transmission. In the case of the NMDA, glutamate and glycine are involved. With allosteric modulators, you are not preventing glutamate or even glycine from binding at their binding sites, but you’re actually modulating the protein in a way that it is increasing or decreasing the open probability of the channel, I think you can make a good argument for allosteric modulation in drug discovery and certainly groups like Jeff Conn’s and Craig Lindsay’s at Vanderbilt make a good case as to why allosteric modulators might be the way to go with this and other receptors. So we believe that allosteric modulation has its advantages for a target as NMDA which is so important for memory and cognition.
Taking a step back from the molecule and targets: the DSM5 (Diagnostic and Statistical Manual of Mental Disorders) is coming out with the latest psychologists’ experimental descriptions of diseases like schizophrenia and bipolar and depression diseases based on what psychologists have seen, talking to people, and yet the National Institute of Mental Health is looking for more mechanistic understandings of these diseases. How do you see that playing out in terms of these diseases being mechanistically-specific versus more blurred and only able to be characterized based on how people talk to other people? Hard question, I know.
Tough question. The DSM5 is for doctors/insurance industry. We are interested in providing new treatments for neurological disorders. Concerning treatment for these disorders, I’m sure you will get a lot of answers. Truth be told, there really aren’t general solutions for treating any of the diseases of the brain. For instance, there are no single drugs/targets that are relevant for all depressed patients, for schizophrenics, for patients with bipolar disease. There are no drugs that currently treat the cognitive aspects of schizophrenia. New descriptions of any of these disorders really has no bearing on our research and I’d be hard pressed to actually support anybody’s contention that any specific target has been unambiguously validated as the “one” for a disorder described by DSM5. For psychiatric disorders like schizophrenia, we have a hypothesis. We’re going to try to find selective drug-like NR2B PAMS and then we will see what these compounds do. But to be honest, do these disease descriptors affect what we do? Not in the least. We continue to see a great therapeutic need for new drugs for schizophrenia, drugs that not only treat psychosis but allow patients to reenter the work force. There have not been new breakthrough treatments for schizophrenia and so there’s still a tremendous need for new medicines. And whether psychosis is part of one disease or another, I mean whether one can separate schizophrenia from psychosis from other similar disorders, I just think that it’s wishful thinking to think that we can match a receptor with a particular disorder described in DSM5. We have a hypothesis and we see what develops from our research. Bear in mind that nine out of ten drugs still fail and that our goal is to obviously provide a new therapy, a new safe therapy for psychiatric disorders.
So I’m going to come back to more of those philosophical questions later. One of the things I was really struck by in preparing for this interview was your background and the breadth and the range of things that you’ve done, and so wanted to walk through some of your experiences in reverse chronological order with the neuronal nitric oxide synthase, glutamate antagonists, calcium channel antagonists from snake venom and all the way to doing process chemistry, so I dug back a bit even to your work with William Johnson and Sam Danishefsky …
Reverse or you want me to start at the beginning with Sam?
Let’s go back in time because that shows where you are now, then to the work you did just previous to the here-and-now, and then work backwards.
OK. Right now I am involved with two companies, SystaMedic and Mnemosyne. Interestingly, both have CDD accounts. How many people have 2 CDD accounts? Prior to working at SystaMedic and Mnemosyne, I was at Pfizer for 30+ years, In the last decade of working at Pfizer, I became convinced that the current way of doing drug discovery was flawed – the idea that by finding a single molecule that interacts with one target will be effective in treating complex diseases. The brain, in particular, is just way too complicated.
In the last few years, I have become more and more involved in thinking about what I call “systems medicine”, and trying to understand the relationship between drugs, genes, proteins and effects. Currently I am fortunate to be able to work with outstanding scientists at two companies, SystaMedic Inc and Mnemosyne Pharma. I co-founded SystaMedic in 2009. I spent the first year at SystaMedic, attempting to capture all the information I could on drugs, diseases, effects and genes, and to basically look for the relationships between all of these; that is, the relationship between drugs and effects and proteins and effects. I had an algorithm developed which enables one to actually go through 22 million abstracts in 10 minutes and basically do searches for lots of interesting information. These searches provide us with invaluable data which allows us to compare drugs and to predict the effects of novel therapeutics.
While I was working just at SystaMedic, I was recruited to join Mnemosyne as the chemist. Mnemosyne interested me a lot because of my interest in diseases of the brain and, in particular, schizophrenia. Truth be told, I love thinking about molecules and I enjoy trying to understand the relationship between structure and function. So working backwards, that’s how I got to where I am today – it’s been a love of molecules throughout my life and thinking about structure-function.
At Pfizer, I was medicinal chemist who worked in the neuroscience area, in diabetes, in inflammation and also worked on antibacterials. It was a lot of fun. One of the great things of working in the industry in the ’80s and ’90s was the fact that you could spend time thinking about problems and testing new hypotheses. Doing basic research became a lot more difficult later on because the industry became so conservative and so goal-oriented that one just couldn’t explore novel avenues that are important but were not in one’s original research plan. If I look over my career in medicinal chemistry, all in all, I was associated with many more failures in research than I’d like to admit. I worked in a number of therapeutic areas on a lot of approaches and almost all of these approaches did not yield drugs – so to me that’s failure – in some cases it was not the approach but the market. The market wasn’t big enough, but in other cases, it was the effects that we saw in animal models or we saw in the clinic that resulted in a compound failing to progress. And actually it was these failures that got me thinking about systems medicine and to looking at targets differently and it also got me into this area of literature data mining.
When I started thinking of this interview, I realized that there was a 20 year break from when I first became interested in glutamate as a target at Pfizer and now at Mnemosyne. Glutamate, as it turns out is also the major neurotransmitter in insects and so in the early 90’s, we developed a collaboration with a spider farm in Utah, because we felt that spiders might provide a source of novel glutamate antagonists. I became for a brief period of time an isolation chemist. In the course of this work, we determined the structures of a number of glutamate antagonists. Along the way, we found a peptide that interacted with a specific type of calcium channel, a p-type calcium channel and we isolated it and we determined the amino acid sequence and disulfide connectivity of this cysteine-rich protein. When we attempted to synthesize it, we could not. It turned out that the compound had a D-amino acid in its peptide chain. This finding was very interesting because the D-amino acid was not at a terminus but rather in the middle of a peptide chain! Luckily for me, this discovery was back in the old days where you could take a few excursions to explore interesting problems. I was very interested in trying to figure out how the D-amino acid got there. Interestingly, we isolated a peptide isomerase and found that this isomerase was able to isomerize amino acids in the middle of specific peptide chains. As it turned out, this was the first time that anybody had isolated such an isomerase, and so we published that in Science. But that was just one of those fun things that we could do back then.
Before that, I was in the diabetes group and I was in the infectious disease groups at Pfizer. Infectious disease research in the late 70’s was a playground for organic chemists who liked to make interesting molecules. That was a lot of fun for a synthetic chemist like me. My antibacterial work led me into the process group for a couple years too and that was a result of my involvement with a broad-spectrum penem anti-bacterial, sulopenem, which was one of the first broad-spectrum penem antibacterials on the market. Since our process group had a tough time making it, I was asked to join the process group, which I did. While in the process group, I also became interested in the activation of imines with Lewis Acids. That interest led me into the development of a synthesis of D-Biotin.
Chronologically, this takes me back to the beginning when I worked with W.S. Johnson at Stanford. That was a fun period for me. It was a great time for chemists interested in organic synthesis because there was an element of uncertainty and urgency in making molecules. It was fun to be involved in synthesis with Bill Johnson. I designed and carried out a synthesis of longifolene in his laboratory. Bill Johnson had a lot of good people in his laboratory at Stanford at that time and most of the people ended up in academia and, in fact, are still scientific leaders in the field of organic chemistry today.
Before that I worked with Sam Danishefsky at the University of Pittsburgh. Sam is a mentor and one of my heroes. I was one of his first students. Being a first student provides one with a unique relationship. When I look back, I appreciate so much about Sam. It was fun. It was challenging. We embarked and were successful in synthesizing camptothecin. It was a great time in my life. In the end, I learned a lot about myself and about life and about family and about science working for Sam.
Interesting, I have a lot of parallels. I was Jonathan Ellman’s first graduate student working on his first benzodiazepine library project, so I was there when the professor was in the lab actually doing experiments. And then at Genentech, I worked on RGD mimics for GP-IIbIIIa which also came from a snake venom initially and that’s been all published from companies like Genentech and SmithKline, so it’s historically recorded older work that seems to parallel what you were doing slightly earlier.
One thing you mentioned about combination therapies; we work a lot with The Gates Foundation on tuberculosis where there are multiple drugs and we have done other collaborations supporting researchers working on malaria where there’s multiple drugs. So I was thinking about that in terms of combinations for complex CNS diseases, where in theory a combination of drugs might be quite useful. Perhaps clinically it’s too expensive to try multiple drugs or the end points are too difficult. I was just wondering, it seems like for the infectious disease area, combination drugs have taken a hold much more than in the CNS arena. So in your opinion is that theoretically ideal or if that’s just due to something more idiosyncratic or arbitrary with the markets rather than something about intrinsically, mechanistically based. How do you think we should treat these CNS diseases?
This is exactly something we’re looking into at SystaMedic, where we have 10 people with different backgrounds and interests. One thing a lot of us believe in is in a systems approach to medicine. I think a big problem with medicine or with drug discovery in the past was the emphasis on one target and one disease. We now realize that for targets in the brain, these approaches won’t work. The brain has developed over millions of years – you would be hard pressed to believe that by shutting down or blocking one pathway, disorders of the brain will be cured. So we believe for drug discovery in some areas – – and certainly for neuroscience, I think that it is better to tweak biological pathways/networks without entirely turning them on or off. So when we’re working with an allosteric modulator on a target in the brain, what we want to do is instead of having a channel being open for 10% of the time, we will try to shoot for it being open for 20% of the time. We don’t want to completely shut it down or we don’t want to completely open it up.
But getting back to your question concerning the brain, it is not likely that drug combinations will be used. However, in cancer and infectious disease, we have pathways/networks of proteins/genes that we want to destroy. In infectious disease, we want to basically kill the organism, right, and so instead of looking for targets on the periphery, you want to look for hub proteins, proteins that are involved right in the center of all the action and when you hit them you want to leave the organism with no options. In diseases of the brain, you can’t target a hub protein. In cancer, however, you’ll see a lot more drug combinations in the future.
So earlier you’d mentioned that you use CDD not for one but for two companies and so I wanted to dig in in terms of how CDD has compared to other informatics companies you’ve worked over your esteemed career on the product side as well as the people side because you know a lot and have seen a lot, so you can talk with some perspective on the relative merits of the CDD Vault and the CDD company.
Well since I have two CDD accounts, you can tell that I do like CDD. I mean on a personal side, I am very pleased with the support I get from CDD. Everybody I’ve talked to at CDD, they have all bent over backwards to help me. And you know when you have small companies, it’s not like there is a great incentive for companies to spend lots of time with you. I currently have accounts with TIBCO and Ingenuity. Both work with Big Pharma and are not really interested in spending time helping small companies. I like the fact that your employees, the face of the company, those are good people and they’re really interested in helping. You just can’t fake that. Everybody that is involved with CDD is special.
We use CDD with our CRO in India. Everybody loves it! So how does it compare with ISIS which we had at Pfizer. Pfizer was always tweaking with systems and trying to develop their own system and I wasn’t really involved in product development. I was just the end user. In Mnemosyne we like CDD because we throw all our data in our vault. You can go into it day and night, go to it quickly. You can order the data easily in terms of activity, in terms of selectivity and you can do structure searches. You can do larger structure searches with compounds and other parameters – and connect the information with other useful databases. There’s just a lot of neat stuff you can do. So for us in Mnemosyne, it’s a great resource since we’re a virtual company. Our biologists can look at the data in Pennsylvania. Our chemists can look at data in Bangalore, India. I can look at it in my home or office. We can look at it all times, and everybody can upload data, so we like it. As for SystaMedic, bear in mind now that our needs are quite different. We’re into structure-function and so we want to throw structures of all drugs in there, not just drugs on the market and nutraceuticals, but we want phase one, phase two and phase three candidates in there. We want to have pro-drugs in there because we want to mine the activity of pro-drugs as well as drugs – we’re interested in designing new pro-drugs and so having a database for SystaMedic works for us because ultimately we tie everything back to structure, and we need that.
Now that brings to mind a challenge we have which we’re working on now and maybe we can help you out in some way that is mutually advantageous. And that is we’re interested in capturing the structures of phase one/phase two compounds and it is a lot harder to easily get information on these compounds and so we’re working on that because we obviously are in the business where we want to predict biological activities.
Getting back to the bottom line, we like CDD for a number of reasons. There are other databases out there, but CDD provides a lot of bang for the buck. If you have limited money and you’re just getting started out, this CDD Vault is a good way to go. It did not make sense for us to buy an ISIS seat. For companies our size that wouldn’t make sense. And not only that, if we did, what if we had problems? We’d be standing in line with the Pfizer’s and the Merck’s of the world and who knows when we would get some help for projects. So we can’t say enough about the product and the CDD folks that we have interacted with. Long and short, we like CDD.
So that’s very nice. I’m sure the folks will appreciate hearing that. I want to focus a little bit on the technology in terms of the CDD Vault. What sort of software features in it do you find most useful? What are the couple most useful things – purely in the product and technology and that can be either in terms of ease of use or in terms of specifically what it does because we put a lot of thought and focus into the design and trying to make things intuitive, yet still able to handle the complexity of the range of biological data people generate. So as an experienced user or from the perspective of some of your colleagues, maybe the biologists use it differently than chemists, et cetera, but I was hoping you could talk about a couple of the technical areas in the product that resonated with you.
Well what’s nice is you throw all this data in to one simple system. We have three programs going and they all benefit from the CDD Vault, so for example, we take NR2B PAM assay and we get all the compounds that have been tested. We can get the ones that were tested last week and we can look at that data right by itself or I can put it all together so that we can order it in terms of activity. We can then take the whole lot and do structure search and see how many compounds have been made with a certain structure and what type of activities they have. Basically I’m not the type of guy who is going to spend hours and hours and hours trying to design a better way of looking at biological data. This just works. For me, I can export everything to Excel, print it out. I have a nice copy of things.
What you’ve done is you’ve designed a product that works for people like me. It not only works great for me but also works great for our biologists and our CRO chemists in India. I like the structure aspect of it too and that helps me in my other company, SystaMedic, where if somebody asks me about a specific compound – if they for example tell me that a particular structure is active or this part of a structure is toxic, I can go through very quickly and look at what other drugs have similar structures and I’ll know a little bit about their activity too. One thing we do have at SystaMedic is we have these enormous databases and we would like to interface these databases with CDD. I can think of times at Pfizer when all of a sudden databases were shut down and we were basically shut out. CDD is always up. Even if it’s going to be shut down for a late night weekend update, we’re notified. I just take it for granted that CDD will work and it will work as planned. So I guess for me the test for the value of something is that you can use it without ever thinking about it. We just assume it’s there and we use it, right, and it’s reasonable. Basically the product is always improving. The product modification improvements are dictated by the customer, and that’s a good thing.
Wonderful note, and really kind words have for the Spotlight interview. I want to end this interview with just a personal invitation for you to return to your roots with W. S. Johnson Group at Stanford in the Bay Area and come back for our ten years of securely hosting collaborative drug discovery data users and community meeting on April 4th, 2014 (and you can already register online). So it’s a ways out next year, but best to plan early if you can be back in the Bay Area for this big meeting down the road at UCSF and either to give a career talk or a panel or something would be quite interesting.
Well that’s cool.
And it’s also a way for me to announce that we’re going to be doing another user and community meeting. We’ll have some of your colleagues from Pfizer we’ve worked with like Chris Lipinski there as well.
All right. That’s good. I’ll put it on my calendar. It’s very nice of you and I’m always excited about going back to the Bay Area. There’s an aspect of me that never left there, so I appreciate that offer. And hopefully by then we’ll have some interesting scientific results so I appreciate the offer, Barry.
Sure, sure. It’s a draw for you and for anyone else who ends up reading the transcript to the bottom. The Bay Area in California is always a pleasure. So thanks for your time, Bob, and unless you have any other questions or comments, bye, Bob. Take care.
All right, so long. Bye. Thanks!
This blog is authored by members of the CDD Vault community. CDD Vault is a hosted drug discovery informatics platform that securely manages both private and external biological and chemical data. It provides core functionality including chemical registration, structure activity relationship, chemical inventory, and electronic lab notebook capabilities!
CDD Vault: Drug Discovery Informatics your whole project team will embrace!