Conversations with Lipid Leaders: Dr. Richard Schwarz
The Conversations with Lipid Leaders Interview Series brings forward insights and experience from the communities most influential and upcoming scientists to stay connected to current colleagues and inspire the next generation of Lipid Leaders! Keep reading for our interview with Dr. Richard Schwarz!
Tell us a little bit about yourself (current role, background, family, etc.)
I have been a basic research scientist for 40+ years focusing on how tendon cells control collagen production and cell proliferation. The two parameters critical for tendon morphogenesis: basically, a collagen rope that links a muscle to a bone. At every stage of development from embryo to adult, the tendon has to be the right size. If the tendon is too long, there is slack and fine muscle control over the bone is lost. If too short, the range of motion is diminished. Nature is the world’s best chemist and nature had millions of years to perfect its art. Deciphering its methods takes time.
Currently, I am transitioning from being a basic scientist to a founder of a small company, SNZR LLC. This transition is required because granting institutions do not consider animal and human clinical trials basic research. This is a major change requiring help with business decisions. support from both human and veterinary orthopedic surgeons, and a new site in order to do the animal trials in chickens (preferably at a veterinary school).
My college education was at the University of California, Berkeley in biophysical chemistry. My Ph.D. was from Harvard in biochemistry and molecular biology (my advisor was Paul Doty) and this was where my tendon research began. I did my postdoc with Mina Bissell at the Lawrence Berkeley National Lab (LBNL). I spent a year in London at the Imperial Cancer Research Fund as a visiting scientist and then, three years at the Jackson Laboratory (Bar Harbor, Maine) as an associate scientist. Finally, returning to LBNL as a research scientist. SNZR LLC was started in 2019 and has one issued patent.
I am married with three adult children living in Oakland, CA.
What do you consider the largest breakthrough in lipid research in recent years?
I realize that every scientist thinks that their research project is important. So, when I nominate the unique lipid that I discovered as a game changer, I expect a fair number of sceptics in the audience. The importance of this lipid comes from its ability to inform the cell of the number of cells and type of cells in its vicinity. This signaling guides the cell to proliferate, differentiate, or apoptose. In tendon morphogenesis, this allows the collagen fibrils to form by means of a growth plate. where cells are growing on the front facing the muscle. While behind them cells have grown to a high cell density, and this triggers them to stop growing and to start making high levels of procollagen. The cells behind them have been at high cell density for a long time and they apoptose. This gives the impression of a cellular machine (growth plate) laying down an even layer of collagen as it extends the collagen rope. Moreover, we postulate with some evidence that muscle connective tissue surrounding muscle bundles are secreting their on cell density signal with another unique lipid, and this signal binding to tendon cells also affect tendon cell growth. If the muscle signal is low, it could enhance proliferation and if high halt tendon proliferation. This would eliminate any potential for slack in the collagen rope. The membrane of the tendon cell would be acting like the input of an analog computer in order guide the cell through the complexity of development.
Did you always envision yourself becoming a scientist? If not, what did you want to be when you grew up? Who influenced you to become a scientist?
I think I have always been curious about how things work. I like to take broken things apart, fix them, and then try to put them back together again. In school I did much better in math and science and this was rewarded with better grades. In college, I had some of the best and a few of the worst teachers. The influence of the good ones outweighed the bad. A good teacher, who loves his field and has been successful at it, is infectious to their students. One of those teachers, Gunther Stent, had been a molecular biologist from the start of the field and had been a part of the rapid burst of knowledge that came from studying bacteria and gene regulation. He taught his class like a reporter at the scene of the event. How one experimental approach worked well while his and others failed for reasons that only became clear years later. At the end of his class, he suggested three fields that he thought would be exciting to study. One was the nervous system and the brain. The second was differentiation. And the third, I think was immunology. I picked differentiation and at Harvard at that time there were only two professors working with higher organisms. One of them ran a large lab studying silkworms but his lab was full. The other had two projects: one was studying early development using freshly fertilized sea urchin eggs and the other was studying collagen expression. Sea urchins were always an unpredictable source of biological material and the scientist working with them were always working late into the night because of a delayed shipment or were frustrated by a canceled shipment. So, I chose to study collagen differentiation. A paper by Dehm and Prockop in 1972 described isolating embryonic tendons by pulling on the toes and then dissociating the cells with bacterial collagenase and off I went.
What caught your interest and motivated you to get involved in the study of lipids and the role they play in human biology?
My first interest focused on a common observation that when cells were taken out of the organism and put in culture, they lost their ability to be highly differentiated. Chicken embryonic tendon cells were no exception. When put into culture they would rapidly lose their ability to make high levels of procollagen (60% of their total protein production). I had to develop a simple assay for procollagen using a purified bacterial collagenase and then use it over and over again to try to figure out what was wrong with the cell culture environment. With what I know now, it is amazing that I succeeded. Sometimes ignorance is a blessing. For instance, if you knew that you were looking for a diffusible growth factor who would put 10 mls of medium in your flask and dilute it. Luckily, the factor was part lipid and had a good affinity for the cell layer. The usual cell culture medium with 10% serum that is good for cell proliferation inhibited procollagen production. By lowering the serum concentration to 0.2% we could achieve good growth, but procollagen production stayed at 12%. Ignorance came in handy again. Chickens make their own ascorbate so one would think that chick tendon cells would not require it. The ignorant scientist gives their cells ascorbate and to his joy as they grow to high cell density, they make 60% procollagen. It turns out that ascorbate is made in the liver, or the kidney (depending on species) and tendon cells take up ascorbate from the blood. Why this complication is not understood? But for me this was a breakthrough for two reasons. One, I no longer had to focus on optimizing the cell culture medium. Two, I had a small molecule, ascorbate, whose only known function is that of a reducing agent, and ascorbate at moderate to high cell density could induce procollagen synthesis by 6-fold to be one of the most highly differentiated cells in the body. A tremendous amount was learned using ascorbate induction. Procollagen secretion went up 6-fold in less than an hour; procollagen translation went up after a 3h lag and reached a 6-fold increase at 48h; procollagen transcription after a lag of 12h rose 3-fold and procollagen degradation decreased 2-fold but mRNA levels took 72h to reach a 6-fold increase. This all makes sense once you realize that there is a strong limitation on how much mRNA can be transcribed from a single copy gene. To get around this constriction point, tendon cells accumulate large amounts of procollagen mRNA and then regulates the pathway dependent on whether the procollagen can form a triple helical conformation, this long skinny structure can only form if a post-translational enzyme, prolyl 4-hydroxylase is active and hydroxylates ~45% of the prolines in the collagen molecule. Its activity depends on two things. One, it is composed of two parts and the alpha subunit is cell density regulated and it goes up approximately 6-fold with cell density. Two, prolyl 4-hydroxylase requires ascorbate to keep its ferrous ion in the reduced state because the hydroxylation step uses oxygen. With this and other information we can postulate a reasonable mechanism for how prolyl 4-hydroxylase activity can link the rate of procollagen secretion and procollagen translation but what is missing is how the cell detects cell density. It is in this phase of the work that I became interested in lipids. I will discuss both in answering the next question.
How did you identify tendon repair as an area of interest for your research efforts?
Success in science requires that your experiments work and that your grant applications based on those experiments get funded. As a consequence, if your findings about how nature works agree with the current fad, then a good score is more likely. If your findings about nature go against the current fad, then a grant application is not the place to convince a study section panel to change their minds. My renewal of my NIH grant first missed by 1 point and was renewed for 1 year. The new members of the study section were rooted in molecular biology, and this came with the firm belief that all protein production was regulated at transcription and that all that was needed was finding the right transcription factor for the procollagen gene.
So, when the wind is not blowing in your direction you tack. My approach was to shift and study cell density signaling. This has been an important problem that has been studied by many labs with limited success. Almost all the previous attempts had focused on cell proliferation. Since so many things affect the ability of the cells to divide, there was always difficulty in distinguishing cell density signaling from other growth factors and inhibitors. Because we could look at a 6-fold procollagen stimulation that is much more specific, we had a real advantage. So, I developed a model where I could study cells at low density and at high density at the same time. Freshly isolated embryonic tendon cells were plated inside a 6 mm cloning ring, they attached quickly to the plastic, and then the cloning ring was removed. As a result, one had an island of cells in the middle of the dish. The cells became confluent in the middle of the island, and they were at low density at the edges of the island. We did in situ hybridization for procollagen mRNA, and we could see high levels in the confluent areas but as we scanned towards to edges the procollagen levels dropped dramatically. This is a short-range signal and a lonely cell at the edge remains unaware of a high cell density region 1 mm away. Another experiment was done where we put dishes with an island of cells on a gentle rocker in the incubator and analyzed them after 24 and 48h. We wanted to see if shaking could alter the ability of the cells to sense cell density and it did. The cultures lost their confluent appearance by 24h and had doubled in number by 48h. Procollagen mRNA levels had dropped to a third of the unshaken control at 24h. If you shake a culture, and you think you have removed something from the cell layer, and this is causing cells to grow then that factor would be called an inhibitor. So, we tested this by shaking confluent cultures cells and taking the conditioned medium and putting it back on cells grown as an island. The conditioned medium did not inhibit growth it stimulated growth. So, we had a factor that if you added it or took it away, the cells were stimulated to grow. A conundrum whose only logical solution is a two-factor model where one factor influences the actions of the other. One is diffusible and the other is not. But we now had an assay for a cell density signal – add a test sample to the medium over an island of cells and see if it stimulates proliferation. You give higher scores if it can stimulate both low density and high density areas.
We optimistically submitted another proposal to the study section and failed. My program manager was at the study section meeting and she said that they would never fund my proposal. Not the words you want to hear from your program manager. But she went on to say that it was not my fault. The study section had funded two other projects where the investigators were purifying factors and they had turned out to be things available in the Sigma catalog. Their take home message was that they were not going to waste precious grant money on this approach again.
So how was I supposed to tack my way out of this shipwreck? Failure gives you time to think. I did a lot thinking and began to realize that I had not exploited all my advantages. I had been thinking like an academic and advancing basic knowledge but if you had a factor that could stimulate tendon cell proliferation and procollagen production then you had a factor that could be very useful in tendon repair – speeding the healing and the strength of the repair. I spent a year cold calling companies, sending letters to about 30 companies, giving seminars to several companies, and networking through people I knew and through friends of friends. One of these yielded a phone call to a former graduate student at Berkeley who had decided to go to Amgen instead of being a postdoc. She was one of the first 10 employees. She gave my letter and some of my publications to the head of protein chemistry and they invited me to give a seminar which I did. After the seminar, as I was waiting in the LA airport, I was wondering why I was doing this. Why was a company going to fund such early-stage work like purifying a protein? I had lots of questions but no real answers. Two weeks later I got a letter from Amgen, they wanted to collaborate on purifying my factor. Miracles do happen. I later learned that in that two week period Amgen had learned that the FDA had approved its first drug and their revenue was going to be in the billions. Also, the head of protein chemistry, wanted a side project that was more challenging than just purifying recombinant proteins. Somehow the stars had all aligned in my direction.
After some administrative difficulties were resolved, we started working. We made 1 liter of condition medium and sent it to Amgen. They purified it over a column, and they sent blind samples back to us for assay. The head of protein chemistry was a wise man and a good scientist. He told me that “all the easy factors had already been purified. The ones that were left would be difficult.” He was right. After 4 years we sent Amgen 400 liters of conditioned medium. They purified it over 4 columns and sent us many blind samples to assay. In the final column, the biological assay tracked with a band on a SDS gel at 16 kD. We had 1 ug of that protein and we sent it for chemical sequencing. We got a 14 amino acid sequence and as expected for this type of sequencing only the early sequences were accurate. In this case, the first eight but this is more than sufficient for finding the gene. However, at the time the only sequenced genome was human, and the human protein varied from the chicken protein at the 5th amino acid making it impossible for us to see that they were 86% homologous over the rest of the gene (this is highly conserved since human and chicken had a common ancestor estimated at 315 million years ago). We had to wait two long years for the chicken genome to be sequenced before this was clarified. The 8 amino acid sequence gave an exact match to the chicken gene, The proposed gene in the database was very large – 324 amino acids. The 8 amino acid sequence does not start with a methionine, and it is located near the carboxy terminus. It was a cleavage product that would release a 94 amino acid protein with a 10.8 kD molecular weight. However, we purified a molecule with a 16 kD apparent molecular weight. At this point, I have to summarize and condense our experiments. We were able to track the protein using an antibody to it and Western blotting. If we changed salt concentrations followed by ultrafiltration with a slightly hydrophobic membrane, we could show that the protein could be released from the cofactor by a shift back to a 11 kD MW. This led us to test various glycosylases and lipid degradative enzymes to see if they would alter the mobility of the factor. Only a general lipase and some specific lipid cleaving enzyme would alter the mobility of the factor. Lipase would release the free protein. The question now became how do we characterize an unknown lipid. Mass spectrometry is the only technique with the sensitivity to analyze a lipid at low concentrations. But at these low concentrations every reagent, every container, and every enzyme has contaminants that can confuse the data. We came up with the following protocol. We would separate the protein from the lipid using ultrafilters and checking with Western blotting that the free protein was released. The free lipid in the flow through was divided into several aliquots: an untreated, treated with lipase, treated with the first specific lipid enzyme, etc. Then, the untreated and treated samples were bound to small C18 hydrophobic spin columns that are mostly used to bind proteins for desalting. But we used them to bind the hydrophobic lipids. Three solvents were used to release attached molecules – methanol, acetonitrile, and chloroform. The mass spectrometer would scan from 100 to 1500 MW. The software would stack scans from all the methanol extractions, and we could easily see if a band in the untreated was missing in the treated samples. None were missing. The same was true for the acetonitrile extractions. But the chloroform extractions did show two missing bands in the lipase and phospholipase C samples. The missing bands were separated by 28 MW a distinctive indicator of lipids because they are lengthened by two CH2 additions at a time and are known to generate these variations. This gave us the MW and ms/ms fractionation showed what looked like two fatty acids. There was a phosphate because of the specificity of the enzyme. Mass spectrometry gave us a lot of information, but it has trouble distinguishing between rearrangements of groups with the same mass. More about this in the last section.
What are your hobbies? What do you like to do outside of the lab?
I like working with my hands and I like figuring out puzzles but outside the lab I like to work on and do thing that I can touch, feel, taste, smell, and see. I like wood working and I built a strip cedar/fibroglass canoe. I was the electrician, plumber, and finish carpenter on the second story addition to our house. I have taken my kids and the dog backpacking all over the Sierras in Northern California. Similarly, I like to cook, garden and be entertained with movies, music, books, and sports.
What was your favorite and least favorite course in school? What was the hardest course for you while you were in school?
I never had a good ear for foreign languages. Luckily, I had a Spanish teacher in college who had trouble getting up for a noon class. She tolerated my poor pronunciation and grammar so that I was able to move on to other subjects. As much as I like the concept of being multilingual, I have retreated to improving my speaking and writing abilities in English.
My hardest courses were in advanced math. Math gets more and more abstract. I still have some of those final exams and I can no longer even understand the questions. On the other hand, when you know that cell density can stimulate a 6-fold increase in secretion, in translation, and in mRNA levels and that 2 factors are involved in cell density regulation, what does this tell you about tendon morphogenesis? Alan Turing, the same person that the movie, “Imitation Games,” was based on, wrote a paper in 1952 explaining how 2-factor models could explain certain types of morphogenesis in biology (he was brilliant, and math came very easily to him). My research assistant at the time was married to a physicist. He thought that tendon morphogenesis was a simple problem in calculus. You write three partial differential equations, and you integrate them, and rate of change turns into cell density and collagen deposition.
We wrote these out in mathematical terms and over lunch he wrote a Fortran program to solve the problem on the computer and generate a picture of tendon morphogenesis. It took a while to work out the bugs, but it proved to me that math can be extremely powerful and beautiful. A problem I had contemplated for hours finally revealed itself through math.
The Avanti team recently completed a custom synthesis project for your lab. Could you describe what it was like working with the Dr. Li and the research chemistry team to identify and synthesize your custom lipid?
There is a short back story to this event. I started a company in order to get a small business grant from NIH or NSF. These grants come in two parts. Phase I is a small one-year award that is designed to advance the project. I thought that this would be perfect for having Avanti make the lipid component and I would make lots of the small protein by genetically manipulating E. Coli. Then combining them, I would have lots of factor to progress to the next step. In Phase II I would do a small animal trial in chickens the first year and then in the second year take this knowledge to do animal trials in the horse. With the ultimate aim of doing clinical trials in humans. The horse has a market in the tens of millions of dollars and the human has a market in the tens of billions of dollars. I thought this was a good plan, but the study section did not. Their main objection was that they would only fund a Phase I project that showed that Phase II would work. In other words, they wanted me to make the factor and test it and then they would fund a small animal trial in Phase 1 that would tell the study section whether the larger trial in Phase II would work. They had other objections that a business of one did not look very viable and that I need advisors so that the transition to animal trials would go smoothly and business advisors who would help me with growing my company. But most importantly they wanted less risk. And they put a lot of risk on me and my small company. Mass spec data, as discussed earlier, does not give the precise structure of the molecule and biological systems are very sensitive to small changes. Without a grant to fall back on, I had to make the right guess on the first try. With that knowledge, I reached out to Dr. Li about whether his group was willing to take on this project of making a unique lipid. He thought it was feasible and he sent me an estimate and I sent him a check and we were underway. I think it was about two weeks later, he wrote to me about the original structure that had two fatty acids, one 20 carbons long and the other 18, bound directly to the phosphate with anhydride bond. This is a high energy bond that can be unstable. He was worried that the intermediate form with only one fatty acid attached would be unstable and he might not be able to get the second one attached. They could make a molecule with two of same length fatty acids but not one of each. I thought that if they added both fatty acids at the same time, they would get a mixture where half would be the right form. However, I expressed my concern that the molecule that I had worked with was very stable and I was wondering if that would be true of this molecule with two high energy bonds. Dr. Li said that he would meet with the experts in mass spec in his group and they would discuss these concerns. A week or so later I received an email from Dr. Li with a suggestion from someone at that meeting who thought that moving an oxygen atom one carbon down would allow the acid to an alcohol that could form an ester bond that is much more stable and yet this new form would have shown the same banding pattern in mass spec. Which form did I think that they should make? This was a good question with two risky answers. As usual in academics, my experts in mass spec had left for other jobs. After a lot of thought, I went with the stable form because it was a trait of the natural product. Dr. Li hedged his bet and made a small amount of both. The predicted unstable form turned out to be unstable and the predicted stable form was stable. Avanti sent me 0.4 g of the stable form. It was an exciting day to open my package and actually see the white powder. For 20 years of research this lipid was part of a predicted factor, missing after enzyme treatment, or a band on a piece of graph paper. I put 0.1mg in 50 ml of ethanol and watched it slowly go back into hiding. It shows optimal growth effects at .000000000025g/ml. Of course, it is bound to a much heavier, small protein but it is the lipid that is active. You cannot add too much of the whole factor to the medium or you will make the cells think they are at a high cell density and they will stop growing and then apoptose. The tendon cells make this factor but they only need an additional boost in concentration when cells are seeded in a culture flask at relatively low numbers. Our addition of factor to the medium makes the tendon cells think that they are at a moderate cell density that favors proliferation despite the 10 mls of medium in the flask. We would also predict that this is why we had to add 0.2% serum to accomplish the same task. In this case serum would be stimulating the overall metabolism of the cell and they would produce more factor by themselves. We are learning more and more about this factor and we will see if this abundance of factor will open the gateway for funding for animal trials. Finally, I want to thank Dr. Li and his team for being lipid experts, giving tremendous advice, and making 20 years of research transformed into an active white powder.
We would like to thank Dr. Richard Schwarz for taking time to catch up with us!
To learn more about Avanti's custom synthesis services, email the team at customsynthesis@avantilipids.com