Professor Pardis Sabeti was able to apply findings from her research on Ebola to now develop a test for detecting COVID-19.
Published March 9, 2021
By Brittany Aguilar, PhD
Pardis Sabeti, MD, DPhil, MSc
This isn’t the first time that Pardis Sabeti, MD, DPhil, MSc, a professor of organismic and evolutionary biology at Harvard University, and newly elected member of the National Academy of Medicine, has worn the hat of viral genome detective in the earliest days of a deadly outbreak or viral disease. Sabeti and her team began sequencing Ebola samples just days after the virus was first detected in Sierra Leone during the 2013-2016 West African outbreak. Since January 2020, she has been working on diagnostics for COVID-19, developing models to predict the most sensitive and accurate assay design candidates for the rapid detection of SARS-CoV-2, including an assay that harnesses the powerful accuracy of CRISPR technology.
Describe the innovative, rapid COVID-19 test that you helped create—how does it work, and why is it an improvement on current testing methods?
Over the last several years, my lab, colleagues, and I have been developing an assortment of technologies for genomic surveillance of pathogens. In particular, we have been deeply invested in CRISPR technologies. CRISPR was first discovered within bacterial immune systems, where it is used to protect the bacteria from invading pathogens by rapidly identifying and targeting a genomic sequence with very high fidelity. Thus, it is immensely powerful as a diagnostic tool, since it can be designed to detect any sequence of genetic material with impressive accuracy.
It is an incredibly exciting technology: it is highly accurate, it would be able to rapidly detect pathogens using little equipment and a simple, paper-strip read-out, and it could be developed in a matter of days to detect newly discovered pathogens or new variants of known pathogens. Crucially, the test is also inexpensive to manufacture, which means it could be easily scaled and distributed as pathogens—or novel variants of pathogens—emerge.
Throughout the duration of the COVID-19 pandemic, some have suggested that testing is optional, unnecessary or unreliable—can you describe why the creation of rapid, reliable tests is so important? Does that change depending on where we are in the infection curve?
Testing is extremely critical to fighting the spread of any infectious disease, and this has been demonstrated through history. However, testing technology has been achievable but not prioritized—if we had invested in this space after the SARS-CoV epidemic [the SARS outbreak in 2003], I believe we could have been poised to respond to SARS-CoV-2 before it spread throughout the world.
The need for diagnostics is critical everywhere, from pre-empting a pandemic, to response and recovery. To be as useful as possible, diagnostics must also be affordable and accessible to all—this is not just in infectious disease but throughout all medicine. The sooner individuals and communities have information, the better they can respond, enabling better outcomes.
You wrote a book last year entitled “Outbreak Culture.” Are there any key learnings from that book that can be applied to COVID or future pandemics?
In this book we argue that a dysfunctional “outbreak culture”—the collective mindset that develops among responders and communities that emerges in the chaos and crucible that is disease outbreaks—poses a great threat to our ability to curb outbreaks and save lives, and that we must continually watch for and dismantle toxic response systems where possible. This includes the data and resource hoarding, perverse capitalistic incentives, the spread of misinformation, and the loss of empathy and good citizenship.
I think people are still just beginning to understand the gravity of outbreak culture and how it is operating amidst COVID. For example, we all now know the importance of detecting outbreaks, through track-and-trace methods, before they have the chance to spread widely. But what is given less attention is how those efforts can be sidelined or undermined by many surrounding societal and political forces.
I always advocate for a massively increased effort for empathy during outbreaks. We need resilient communities to be able to do the best work against infectious disease. With our trust in our fellow citizens, our leaders, and our scientists undermined during this time, it is crucial to work within the community and low to the ground. We must listen to others, respect their opinions, and understand their fears. For that reason, I believe we must double down on empathy when it comes to community participation. If we do not work with communities and support them in the right ways, we end up causing more harm than good.
About Prof. Sabeti
Pardis Sabeti, MD, DPhil, MSc is a Professor at the Center for Systems Biology and Department of Organismic and Evolutionary Biology at Harvard University and the Department of Immunology and Infectious Disease at the Harvard School of Public Health. She was a 2016 and 2017 Finalist for the Academy’s Blavatnik National Award for Young Scientists. To learn more about Dr. Sabeti and her work, click here to listen to the “Deciphering Zika” podcast.
By definition, a rare disease is one that afflicts relatively few people compared to the general population. Collectively, though, there are over 7,000 of these conditions known, causing immense suffering for an estimated 300 million patients. Because most rare diseases stem from specific genetic mutations, they’ve proven difficult to treat.
Genome sequencing and molecular medicine might soon change those grim statistics, though. For example, using short DNA or RNA sequences complementary to the messenger RNA for a gene, researchers can inhibit the expression of the associated protein. These complementary sequences—called antisense oligonucleotides—could soon be delivered as drugs to treat many rare diseases.
On October 2, 2020, The New York Academy of Sciences and Takeda Pharmaceuticals hosted the Frontiers in Rare Diseases: 2020 Innovators in Science Award Symposium, an event highlighting breakthroughs in rare diseases research and honoring 2020 Innovators in Science Award Winners Adrian Krainer, PhD and Jeong Ho Lee, MD, PhD. Presentations, a panel discussion, and a virtual poster session covered the basic science, recent clinical breakthroughs, and remaining challenges in this rapidly evolving field.
Symposium Highlights
While many rare diseases are inherited, others arise through mutations in somatic cells during life.
Antisense oligonucleotides can alter the expression of specific genes, potentially mitigating or reversing many genetic diseases.
Clinical trials for rare disease therapies must be tailored to the pathogenesis of each disease.
Redirecting neural stem cells to become neurons could treat many neurodegenerative diseases.
The COVID-19 pandemic is inspiring new collaborations that could be adapted to rare disease research.
Speakers
Jeong Ho Lee, MD, PhD Korea Advanced Institute of Science and Technology
Adrian Krainer, PhD Cold Spring Harbor Laboratory
Annemieke Aartsma-Rus, PhD Leiden University Medical Center
Don Cleveland, MD, PhD University of California, San Diego
Huda Zoghbi, MD Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital
Brad Margus Cerevance
Graciana Diez-Roux, PhD Telethon Institute of Genetics and Medicine
David Fajgenbaum, MD University of Pennsylvania
Anne Heatherington, PhD Takeda Pharmaceuticals
Sponsors
The Winner’s Circle
Speakers
Jeong Ho Lee, MD, PhD Korea Advanced Institute of Science and Technology
Adrian Krainer, PhD Cold Spring Harbor Laboratory
Not Born This Way
Jeong Ho Lee, Early-Career Scientist winner of the 2020 Innovators in Science Award, discussed his work studying how somatic cell mutations—mutations that occur after development, during the normal process of cell division—result in rare neurological diseases caused by somatic cell mutations in the brain. Much of the recent boom in work on genetic diseases has focused on germline mutations. Because these mutations occur early in embryonic development, they show up in many types of cells throughout the body and are passed on to future offspring. These rare germline mutations can often be identified by sequencing the genomes of cells in easily accessible tissues, such as blood or skin. With advances in next-generation sequencing, “it’s become much easier to identify the germline mutations coding for many rare neurological disorders,” said Lee.
Somatic cell mutations occur throughout life, in every part of the body.
Nonetheless, germline mutations account for only a minority of rare neurological disorders. For example, 98% of epilepsy cases cannot be explained by germline mutations.
“We hypothesized that somatic cell mutations may be responsible for these unexplained neurological [diseases],” said Lee.
Somatic cell mutations occur during the ordinary cell division process that takes place billions of times in developing embryos, and continues to occur throughout life as somatic, or non-gamete, cells turn over.
DNA replication isn’t perfect, and human cells average 0.1 to 3 mutations per genome every time they divide. Lee theorized that a patient who hadn’t inherited an epilepsy-causing germline mutation might instead acquire somatic cell mutations in a subset of brain cells during development or later in life. If that happened, the mutation would only show up in the affected area of the brain, not in any other cells of the body.
One treatment for certain types of epilepsy is to resect the portion of the brain causing the seizures. Lee and his colleagues took samples of the brain tissue resected in these operations, along with blood samples from the same patients, and performed deep DNA sequencing to identify somatic mutations that occurred only in the affected brain tissue, not in the blood. They identified such mutations, including ones unique to genes involved in motor nerve activity, in 30% of the patients. When the scientists introduced the same mutations into a small percentage of neurons in developing mice, the animals developed epilepsy.
Next, the investigators looked at brain tumors, which can also cause epilepsy. One rare brain tumor type involves both glial cells and neurons, triggering epilepsy. Sequencing genetic information from cells in the tumors revealed that in 46% of affected patients, the glial cells and the neurons in the tumor shared an identical mutation.
“It means that the…neural stem cell already contained this… mutation,” said Lee, “and differentiated into the neuron and the glial cell.”
That could explain the high rate of disease recurrence in patients with these tumors; even if surgeons remove the entire tumor, the mutant stem cells might continue to produce more defective neurons and glial cells, which could then seed the growth of a new tumor.
To confirm that, Lee’s team collected an additional round of samples, this time sequencing cells not only from resected brain tumors and blood, but also from the subventricular zone in each patient’s brain, an area rich in undifferentiated neural stem cells. They found the tumor-associated mutation in the subventricular zone samples as well as the tumors, indicating that the error occurred in the neural stem cells, whose neuronal and glial descendants then migrated to where the tumor grew.
The researchers are also looking at neurodegenerative disorders such as Alzheimer’s disease.
“We hypothesized that somehow brain somatic mutation maybe accumulates over aging, and maybe associates with [Alzheimer’s disease development],” said Lee.
By performing deep sequencing on brain tissues from patients with and without Alzheimer’s disease, he and his colleagues identified somatic mutations unique to the diseased brains, supporting their theory.
In addition to identifying the underlying mechanisms behind neurological diseases, Lee is trying to help patients in other ways. In one effort, he has begun providing his results to clinicians to use in genetic counseling. Because conditions caused by somatic mutations aren’t heritable, while those caused by germline mutations are, patients who might be considering having children need to know which category they’re in.
The investigators are also trying to find ways to repair or mitigate the effects of somatic mutations in the brain, but it’s a tall order.
“Even if we found a molecular genetically validated target in the patient’s brain, it would be very difficult to develop a traditional drug to penetrate the blood-brain barrier and regulate the target,” Lee explained.
Instead, he’s hopeful that chemically modified strings of nucleic acids, called antisense oligonucleotides, will be able to target the somatic mutations he’s identified.
“I believe in the next five, ten, or twenty years, we probably can solve a lot of the rare neurological disorders,” he said.
Different Diseases, Different RNA Splices
Adrian Krainer won the 2020 Innovators in Science Senior Scientist Award, recognizing years of work spent developing treatments for rare diseases. Krainer and his colleagues were the first to develop an effective drug to treat spinal muscular atrophy. Affecting about 1 in 10,000 people worldwide, spinal muscular atrophy is an inherited genetic disease caused by a defect in the SMN1 gene. SMN1 encodes the SMN protein, which is essential for motor neuron survival. Patients with the mutation experience progressive loss of motor neurons, leading to loss of muscle control and, in most forms of the disease, early death.
Another gene, SMN2, also encodes the SMN protein, but cells usually splice out one of the protein coding sequences, or exons, from the SMN2 messenger RNA, preventing it from making the full-length protein. As a result, 80-90% of the protein translated from SMN2 RNA is truncated, nonfunctional, and rapidly degraded by the cell. Krainer reasoned that preventing the exon-skipping event might allow patients’ unmutated SMN2 genes to produce more functional SMN protein, overcoming the deficit caused by their mutated SMN1 genes. To do that, his team turned to antisense oligonucleotides, which encode the complementary, or antisense, sequence of a specific RNA target. When introduced into a cell, the antisense oligonucleotide binds specifically to its target sequence, triggering various cellular responses.
By designing an antisense oligonucleotide that altered the splicing of SMN2 messenger RNA, the researchers were able to get SMN1-mutant cells to produce more functional SMN protein. Subsequent preclinical and clinical trials proved that their antisense oligonucleotide also works in spinal muscular atrophy mouse models and in patients, respectively, significantly mitigating their motor neuron losses.
“Therefore this is a way that allows them to make closer to normal levels of a functional SMN protein in the presence of this drug,” said Krainer.
The oligonucleotide, now sold as nusinersen (Spinraza), was approved in the US in 2016 and the EU in 2017. Over 11,000 patients now receive it worldwide.
Nusinersen (Spinraza) pioneered many aspects of molecular medicine.
Based on the success of nusinersen, Krainer and his colleagues have begun looking at other RNA processing events to target with antisense oligonucleotides. One project focuses on familial dysautonomia, an inherited genetic disorder that affects only 310 known patients worldwide. These individuals have profound defects in their sensory neurons and autonomic nervous system, leading to symptoms that range from insensitivity to pain to difficulty swallowing.
“It is a very severe disease, a rare disease with a complex set of symptoms,” said Krainer, adding that “median survival is about 40 years of age.”
The condition is caused by a mutation in the gene for a protein called ELP1. As in SMN2, the mutation causes one exon of the gene’s messenger RNA to be spliced out, leading to a loss of functional ELP1 protein.
“So, we started targeting this aberrant splicing event using the same screening strategy and the same chemistry that we used…for spinal muscular atrophy,” said Krainer.
That effort identified an antisense oligonucleotide that can reverse the ELP1 RNA splicing defect in cultured cells from patients, as well as a transgenic mouse model.
“We feel that this is ready for clinical development; it is a challenge, though, because of the rarity of this disease,” said Krainer.
With only a few hundred patients in the US and Israel, the market for familial dysautonomia therapies is minuscule, and effective screening of potential carriers of the affected gene has led to very few new patients being born.
Not all RNA splicing-related diseases are rare, though. Work by several researchers has shown that in at least some cases, a change in the splicing of messenger RNA can help cancer cells grow. Alternatively, spliced forms of the messenger RNA for the PKM gene can produce two different isoforms of the metabolic enzyme pyruvate kinase. PKM1 predominates in normal adult tissues, while tumors and some developing tissues favor PKM2 production.
Using the same approach that worked in their rare disease work, Krainer’s team screened antisense oligonucleotides and identified candidates that bound the PKM messenger RNA and directed its splicing to favor PKM1 protein production. Putting these oligonucleotides into hepatocellular carcinoma cells causes the cells to shift their metabolism and slow their growth. In a mouse model of hepatocellular carcinoma, injecting the antisense oligonucleotides led to a significant reduction in tumor growth compared to control animals treated with saline solution.
Annemieke Aartsma-Rus, PhD Leiden University Medical Center
Don Cleveland, MD, PhD University of California, San Diego
The Kindest Cut
Annemieke Aartsma-Rus began the meeting’s third session with a presentation about her group’s efforts to address Duchenne muscular dystrophy with antisense oligonucleotides. While Krainer’s approach to rare diseases focuses on conditions where an exon needs to be added back into a messenger RNA, Aartsma-Rus described a case where it’s better to remove one.
Duchenne muscular dystrophy is an X-linked genetic disorder. In most cases, a mutation in the dystrophin gene shifts the messenger RNA’s reading frame, causing translation of the dystrophin protein to fail.
“Patients become wheelchair dependent around the age of 12, need assisted ventilation around the age of 20, and generally die in the second to fourth decade of life,” said Aartsma-Rus.
A related but milder disorder, Becker muscular dystrophy, also involves a deletion in the dystrophin gene but doesn’t shift the messenger RNA’s reading frame. As a result, patients with Becker muscular dystrophy produce partially functional dystrophin and exhibit a slower disease progression.
Skipping an exon in the RNA can fix a frame-shift mutation.
Looking at the affected DNA and RNA sequences, Aartsma-Rus reasoned that most Duchenne muscular dystrophy patients could make Becker-like dystrophin, if their cells could simply skip the affected exon in their dystrophin messenger RNA. To test that, she and her colleagues developed chemically modified antisense oligonucleotides that would remain stable in blood and tissues, and began testing them as potential drugs. By designing an oligonucleotide that targeted RNA splicing, the team restored dystrophin expression in cultured cells carrying a Duchenne muscular dystrophy mutation.
The researchers discovered a potential roadblock in a mouse model: antisense oligonucleotides injected into the animals’ tail veins were absorbed almost entirely by the liver and kidneys. The investigators could inject the molecules directly into muscles instead, but that clearly wouldn’t be a practical way to deliver treatment to patients.
“We have over 700 different muscles, and you’ll have to treat patients repeatedly,” said Aartsma-Rus, “so local injection of each and every muscle weekly or even monthly is likely not realistic.”
However, in a mouse model of Duchenne muscular dystrophy, the team discovered that oligonucleotides injected into the animals’ tail veins were absorbed into muscles ten times better than they had been in wild-type mice.
“The first time we thought we’d made a mistake, so we repeated it a couple of times,” said Aartsma-Rus, “but every time we saw that there was higher uptake by the dystrophic muscle than the healthy muscle.”
Dystrophin deficiency causes muscle cells to become more permeable, leading to leakage of cellular components, but this leakage works in both directions; the dystrophic cells readily absorbed oligonucleotides that healthy cells excluded.
Flush with this preclinical victory, the team began setting up clinical trials in 2007. The initial multi-center, open-label trial found that the antisense oligonucleotides caused no serious side effects, and eight of the twelve patients tested saw their conditions remain stable throughout the trial. To evaluate efficacy, the investigators moved into a phase 2b trial, which continued to show dose-dependent effectiveness in treated patients. However, a larger phase 3 trial yielded disappointment, with no significant difference in outcomes between treated and control patients.
“So, what happened [to explain why] we see these beneficial effects in the phase 2 trial, but in the phase 3 trial we see no effect?” Aartsma-Rus asked.
Analyzing the results and the disease further, she realized that the trials were built on the flawed assumption that the patients’ progression would be linear. Instead, they realized that younger patients tend to remain stable for an extended period, followed by a rapid decline later in life. By mixing different ages in the phase 3 trial, patients with worse disease symptoms likely masked any treatment benefits in those with milder symptoms.
Looking at the trial’s failure, Aartsma-Rus concluded that she and her colleagues should have opened discussions with regulatory agencies sooner, and studied the natural history of the disease more thoroughly, before initiating the phase 3 study. Unfortunately, the expensive late-stage failure has soured companies on further clinical development of exon-skipping antisense oligonucleotides for Duchenne muscular dystrophy. Aartsma-Rus has since focused on preventing such an outcome in the future.
“Now we have an open dialogue with academics, with patients, with regulators in the EU, and it is also starting in the US, developing new outcome measures,” she said, adding that “future trials will be better.”
Batting for Lou Gehrig
Don Cleveland discussed his group’s efforts to treat neurodegenerative diseases in the brain, especially those that develop gradually with age. In many of these conditions, such as Alzheimer’s and Parkinson’s disease and amyotrophic lateral sclerosis, “the genes that contribute to disease are all widely expressed…throughout the nervous system, not within individual classes of neurons,” said Cleveland. Mechanistic studies have suggested that decreasing the expression of the defective gene products in some of these cells could moderate the course of disease, so Cleveland and his colleagues set out to do just that.
The researchers first focused on amyotrophic lateral sclerosis (ALS), also known as Lou Gehrig’s disease. About four to five million people alive today will die of ALS, a progressive neurodegenerative condition that can be inherited or occur spontaneously in adults. One inherited form of the disease stems from a mutation in the gene for superoxide dismutase, which causes neurons to die through mechanisms that aren’t entirely clear yet.
Using the same strategy as his co-speakers, Cleveland’s lab designed antisense oligonucleotides that bind specifically to the superoxide dismutase messenger RNA and target it for degradation in the cell. That decreases the level of the enzyme, an intervention that had previously been shown to ameliorate ALS progression in a mouse model of the disease. The next challenge was delivering the oligonucleotides to affected neurons in the brain.
Antisense oligonucleotides have immense potential to be used as drugs against a wide range of diseases.
“These DNA drugs were ten to fifteen times the size of a typical drug, and they’re heavily charged,” said Cleveland, “so the pharmacology textbooks all said that there was no uptake mechanism that would permit them to be efficiently taken up [by neurons].” Nonetheless, he continued, “we tried it anyway, and it turns out that the cells of the nervous system hadn’t read the textbooks.”
Injecting the antisense oligonucleotides into the cerebrospinal fluid of mice genetically modified to develop severe ALS doubled the animals’ survival times.
While the superoxide dismutase defect was the first ALS mutation discovered, the most common cause of the inherited form of the disease is a mutation in a gene called ORF72, which inserts extra nucleotides into a non-coding region of the gene. This causes defective messenger RNA to accumulate, killing neurons because of a lack of functional ORF72 gene products and the accumulation of toxic byproducts of the altered gene. Antisense oligonucleotides targeting the defective RNA, however, inhibit its accumulation without reducing the production of working ORF72 gene products in cultured cells.
In an animal model of the ORF72 defect, the results were even more impressive.
“We dosed these animals [with the antisense oligonucleotides] at the age of disease onset and asked what happens, and the answer is we prevented further disease development for the life of those animals with a single dose injection applied at the initial signs of disease,” said Cleveland.
His team initiated clinical trials on this therapy in 2018, just seven years from the date when researchers had first published the data showing the ORF72 mutation caused ALS.
Although it’s an important target for research, inherited ALS accounts for only 10% of the disease’s total cases. In 90% of patients, the condition develops spontaneously due to somatic cell mutations later in life. Many of these cases involve mutations in the TDP-43 gene, which encodes a nuclear protein that regulates Stathmin-2, which in turn plays a critical role in regulating the cytoskeleton in neurons. TDP-43 normally binds the Stathmin-2 precursor RNA and ensures that it gets spliced properly into messenger RNA. Mutations that inactivate TDP-43 cause a loss of functional Stathmin-2, which is a hallmark of sporadic ALS.
Using cultured neurons, Cleveland and his colleagues found that a properly designed antisense oligonucleotide could compensate for the loss of TDP-43 activity, restoring normal RNA splicing and Stathmin-2 expression.
“This now enables a strategy for therapy for sporadic ALS,” said Cleveland.
If the result holds in other preclinical models, he expects to take that approach into clinical trials in 2023.
Besides correcting specific defects within a cell, antisense nucleotides can potentially redirect a cell’s fate entirely. That’s the central theme of another project in Cleveland’s lab, in which the team is causing astrocytes to change their identities. Astrocytes are companion cells in the nervous system that arise from the same stem cells as neurons. Using antisense oligonucleotides, the investigators can suppress two genes that direct cells into the astrocyte lineage, causing them to become neurons instead. Cleveland is initially focusing on treating Parkinson’s disease with this approach, but he explained that “this…conversion of astrocytes into replacement neurons may be broadly applicable for neurogenic disease.”
Huda Zoghbi Jan and Dan Duncan Neurological Research Institute, Texas Children’s Hospital
Maybe Not So Rare
Huda Zoghbi gave the meeting’s keynote presentation, which covered her work on Rett syndrome. Caused by spontaneous mutations in the MECP2 gene on the X chromosome, Rett syndrome is a progressive neurodegenerative disease that primarily manifests itself in girls. MECP2 is critical for gene regulation in neurons. Because females carry two copies of the X chromosome and an inactivate one in each cell, an inactivating mutation in MECP2 impairs the function of 50% of the affected individual’s neurons. That manifests itself as a rapid regression in motor and cognitive abilities by age two.
In boys, who have only one X chromosome, inactivation of MECP2 is generally lethal before age two. They don’t live long enough to develop the classic symptoms of Rett syndrome. However, recent work has revealed that some males acquire mutations that cause less severe defects in MECP2.
“What we’ve learned is when people carry milder mutations, we will see milder phenotypes, such as mild learning disability with…neuropsychiatric features,” said Zoghbi.
These individuals’ phenotypes can range from autism to hyperactivity or schizophrenia, but most die by middle age due to neurodegeneration. Females with mild defects in MECP2 show non-random inactivation of their X chromosomes, favoring the healthy copy of the gene and enabling them to develop and live normally. Some patients also have duplications in their MECP2 genes, often leading to severe neurological problems and premature death.
To understand the mechanisms driving Rett syndrome, Zoghbi and her colleagues developed a series of genetically modified mice carrying various duplications or mutations in MECP2. Consistent with the findings in humans, these animals display a spectrum of phenotypes depending on the severity of their MECP2 disruptions.
“The brain is very sensitive to the activities and functions of this protein, and we’ve done a lot of studies on both the loss and the gain models,” said Zoghbi.
She and her colleagues found that all types of neurons require functional MECP2 to operate normally.
Mutations affecting different types of neurons can cause a wide range of neurological phenotypes.
Next, Zoghbi and her colleagues tried inactivating MECP2 in excitatory and inhibitory neurons separately. They found that in both cases, animals developed obesity, lost motor coordination, and died young. However, targeting MECP2 only in inhibitory neurons led to more learning and social defects in the animals, while inactivating it only in excitatory neurons caused more anxiety and tremors. These phenotypes represent the downstream effects of the genes MECP2 would normally regulate.
“Given that it’s important for practically every cell, really there’s two major ways you can think of treating this disorder, either gene replacement therapy…or perhaps exploring modulation of the [MECP2 regulatory] circuit,” said Zoghbi.
Taking the latter approach, the investigators implanted electrodes into the brains of mice to deliver small electrical pulses. This type of deep brain stimulation, which has been shown to reverse many types of neuronal signaling and development defects, is already approved for human treatment of several neurological disorders. Stimulating the brains of Rett syndrome model mice leads to significant recovery in their learning, memory, and motor abilities that persists for weeks after treatment.
“It was really quite a dramatic rescue in that all these phenotypes normalize, and their normalization…lasted for several weeks,” said Zoghbi.
The treated animals’ neurons also displayed gene expression patterns similar to wild-type animals, whereas untreated animals showed significant gene dysregulation.
“The Rett brain, at least in mice, is responsive to neuromodulation,” said Zoghbi.
Looking at the MECP2 gene itself, Zoghbi’s team identified regulatory sequences that control its expression level. Altering these sequences to increase or decrease the amount of MECP2 expressed in mice underscored their earlier findings, showing that even modest changes in MECP2 levels led to detectable neurological phenotypes.
Like other speakers at the meeting, Zoghbi and her colleagues are also exploring the potential of antisense oligonucleotides as therapies. That approach seems especially promising for patients with duplications in MECP2 that lead to overexpression of the gene. In mice that recapitulate this condition, the researchers found that treatment with antisense oligonucleotides against MECP2 could reduce the amount of functional protein in neurons down to wild-type levels. The treatment reversed the animals’ motor defects.
Titrating the antisense oligonucleotide dosage also revealed that even modest decreases in excess MECP2 can lead to major improvements in symptoms.
“If you can even partially decrease the protein…you will probably rescue quite a bit of the features of the disease,” Zoghbi said. She added, “I’ve really never worked with a protein that is so exquisitely sensitive to the levels.”
Graciana Diez-Roux, PhD Telethon Institute of Genetics and Medicine
David Fajgenbaum, MD University of Pennsylvania
Anne Heatherington, PhD Takeda Pharmaceuticals
Silver Linings
The meeting’s general session concluded with a panel discussion led by Brad Margus, co-founder and CEO of Cerevance. With a background in business, Margus moved into rare disease drug development after his daughter was diagnosed with ataxia-telangiectasia, a genetic disorder that causes neurodegeneration and immune dysfunction. The panel also featured Graciana Diez-Roux, chief scientific officer at the rare disease-focused Telethon Institute; David Fajgenbaum, a physician-scientist who both studies and suffers from Castleman syndrome; and Anne Heatherington, a data scientist for Takeda Pharmaceuticals with extensive experience studying Duchenne muscular dystrophy.
Panel members discussed the need for improvement in collaborations between patients and researchers.
“There is a lot of miscommunication within the rare disease research space, [but] I think there’s been a really great trend for groups like Takeda and others toward engaging patients in the research process,” said Fajgenbaum, adding that “I also think clinicians can really be a part of this.”
Besides improving clinical trial recruitment, involving patients more directly in research can have far-reaching benefits for scientists.
“It’s incredible how our PhD students, when they have the chance [to interact] with the patients and [get] to know the patients’ organizations…how their motivation and their love for what they do changes,” Diez-Roux said.
Besides increasing collaborations between patients and scientists, all of the panelists endorsed the need for strong, well-defined partnerships with pharmaceutical companies. Margus described his company’s efforts to improve data collection and sharing for ataxia-telangiectasia, which included building a system that uses wearable devices to collect movement data from patients around the clock.
“The data [are] truly owned by the families and the community, and we can make decisions about sharing the data with academics or any researcher in the world in a matter of days,” said Margus.
Good partnerships require more than just good databases, though. Academic researchers accustomed to independent, curiosity-driven experimental design and flexible deadlines sometimes have trouble accommodating pharmaceutical companies’ urgent, goal-directed needs.
“I think the model has to be somewhere between…the industry knowing how to deal with the academic research and academic researchers being open to notice that industries have…different goals in some respects,” said Diez-Roux.
The group also discussed the impact of the COVID-19 pandemic. In the short term, of course, the global shutdown caused by the SARS-CoV-2 virus has halted or delayed many rare disease studies. However, panelists agreed that some of the innovative approaches developed for the pandemic response could transform many aspects of rare disease research in the future.
“We have been very involved in a lot of the COVID alliances, and have been steeped in novel ways of working,” said Takeda’s Heatherington.
As an example, she pointed to the company’s involvement in multi-corporation consortia to develop new therapies and even entirely new platforms for therapies.
“That’s a real breakthrough in terms of how we do our business, that extent of collaboration for [competitors to] come together,” Heatherington continued.
At the same time, “the public is realizing more how important research is, and this goes for COVID, but I think it goes for all diseases,” said Diez-Roux. Both she and Heatherington also pointed out that the pandemic has underscored the potential tradeoffs between speed and safety in therapeutic development, and highlighted the importance of oversight in clinical trials.
The meeting concluded with a virtual poster session, featuring rapid-fire presentations of some of the newest research in rare diseases and offering attendees the ability to interact with the presenters directly. Like the other presentations, the posters represented the diversity and enthusiasm of rare disease researchers.
“What makes me optimistic is the passion and the knowledge…and the fact that we have people that are so dedicated to rare diseases,” said Heatherington.
Lewis C. Cantley’s discovery of the enzyme phosphatidylinositol-3-kinase (PI3K) paved the way for a better understanding of cellular metabolism and its role in human diseases. In response to insulin, PI3K signals through lipids to activate a cellular cascade resulting in increased glucose uptake and subsequent cell growth and division. Cantley’s work has led to new cancer therapies, as PI3K pathway mutations are among the most common to drive cancer development, and a better understanding of insulin resistance in diabetes.
For his groundbreaking discovery of PI3K and critical body of research, Lewis C. Cantley, PhD, of Weill Cornell Medical College, received the 2020 Dr. Paul Janssen Award for Biomedical Research. On September 16, 2020, the New York Academy of Sciences hosted the award symposium to celebrate his achievements. Following Cantley’s award lecture, other experts in the field shared their work on the intersection between cellular metabolism, biology and disease.
Symposium Highlights
Differences in cell metabolism across cancer types may help explain why cancer cells are differentially sensitive to drugs that target metabolism.
The tumor suppressor protein p53 regulates several pathways that manage the production of reactive oxygen species in the mitochondria.
Metabolic pathways represent a powerful and to-date, underappreciated set of therapeutic targets for cancer.
Drugs that regulate metabolic pathways involved in cell differentiation are promising targets for acute myeloid leukemia.
Cancer researchers are coming to appreciate how a patient’s diet can be therapeutically applied and may directly modulate their disease progression and therapeutic response.
Speakers
Lewis C. Cantley, PhD Weill Cornell Medical College
Matthew Vander Heiden, PhD Massachusetts Institute of Technology
Karen Vousden, PhD Francis Crick Institute
Ulrike Philippar, PhD Johnson & Johnson
Costas Lyssiotis, PhD University of Michigan
Discovering Phosphatidylinositol-3-kinase and Its Link to Cancer
Speakers
Lewis C. Cantley Weill Cornell Medical College
Metabolism, Health, and Cancer
Lewis C. Cantley described how the discovery of the enzyme phosphatidylinositol-3-kinase (PI3K) and its signaling pathway led to the development of a new class of cancer drugs called phosphoinositide-3 kinase inhibitors.
He traced his foundational work on PI3K inhibitors to a single slide that his then-graduate student, Malcolm Whitman, produced for a lab meeting in 1987 when he was at Tufts University in Boston. The lab was studying phosphatidylinositol and its role in a signaling cascade regulated to insulin and cell growth. They purified an insulin-activated lipid kinase and found that when activated, it is associated with proteins linked to cancer.
The slide from a 1987 lab meeting that sparked decades of research.
At the time, researchers thought that phosphatidylinositol gets phosphorylated at position 4 on its sugar ring to produce a lipid called PI4P, which is involved in cell regulation. But Whitman found these phosphorylated lipids seemed to have two subtypes—one consistently ran a millimeter further on the gel. “That one-millimeter difference convinced me, as a chemist, that these had to be different species,” Cantley recalled. It turned out that their insulin-regulated kinase—now called PI3 kinase—produces a different molecule, called PI3P, and modulates a previously undiscovered insulin-regulated pathway that regulates the cell’s ability to take up glucose.
Cantley’s team went on to elucidate how PI3K regulates cells’ response to insulin signaling. “Almost every step in the first half of glycolysis is regulated by the PI3 kinase pathway,” he said. Collaborating with a team led by John Blenis, now at Weill Cornell Medicine, Cantley’s also fleshed out the connection between PI3K and multiple other cancer related signaling pathways, such as PTEN, Ras, and mTOR. Cantley called it “really quite remarkable” that “every tumor has at least one of the molecules in this pathway mutated—and many have several.” Indeed, accumulating research since then shows PI3K is one of the most frequently mutated oncogenes in cancer.
Although four different genes encode PI3Ks, just one of them—PIK3CA, which mediates insulin response, is widely mutated in cancers. Yet despite extensive effort, researchers have struggled to get PIK3CA-targeted drugs approved. Because this form of the protein mediates insulin activity, any drug that targets it makes patients insulin-resistant. “So you’re fighting the battle of insulin resistance as you are trying to treat the cancer, Cantley explained.
PIK3CA mutations are especially prevalent in uterine, cervical, breast, and ovarian cancer. In 2009, Cantley received a $12 million grant from the Stand Up To Cancer foundation to form a “dream team” of researchers to investigate how to develop PI3K inhibitors for women’s cancers and explore effective drug combinations. Through this grant, Cantley and his team conducted an early-stage trial of BYL719 (Alpelisib), a PIK3CA inhibitor developed by Novartis, in combination with a chemotherapy agent called letrozole, which lowers estrogen production, in people with metastatic breast cancer. The combination reduced glucose uptake to tumors and extended progression-free survival by roughly one year. The US Food and Drug Administration approved it in 2019.
PI3 Kinase inhibitor BYL719 combined with the chemotherapy agent letrozole reduced glucose uptake from tumors of a patient over the course of one month.
But in that early clinical trial, Cantley noticed that although the drugs’ response correlated with PIK3CA mutations overall, it was uneven; some patients with PIK3CA mutations didn’t respond, and some with mutations in unrelated genes did. “This was not a clear home run,” he said. “I argued, as we looked at the data, that maybe this has something to do with patients’ insulin levels.”
The insulin receptor (IR) is expressed in most tumors. Cantley and his colleagues wanted to interrupt the cycle in which high insulin levels trigger the IR and PI3K activity in the tumor, ultimately driving glucose uptake. To do so, the researchers tracked insulin levels in patients and mice taking PI3 kinase inhibitors. All three drugs they tested raised insulin levels 20-fold. In additional experiments, cells grown from patients with PIK3CA tumors died when exposed to PIK3CA inhibitors but were rescued when insulin was added. “So those ambient levels of insulin really are keeping the tumor alive,” Cantley said.
In subsequent studies, Cantley and his team tested the effects of insulin-lowering drugs, including metformin, as well as a ketogenic diet in mice. They found that a ketogenic diet effectively maintained low glucose and insulin levels while administering a PI3K inhibitor. In mice carrying tumors with a PIK3CA mutation, combining the PI3K inhibitor with a ketogenic diet significantly suppressed tumor growth compared to the control group. “This really tells us that keeping insulin down during the treatment with a PI3 kinase inhibitor could potentially have huge improvements in patient responses,” Cantley said.
Researchers are looking more closely at the effect of a ketogenic diet on the efficacy of PI3K inhibitors for endometrial cancer, breast cancer, and lymphoma. Previously a postdoc in Cantley’s lab and now at Weill Cornell Medicine, Marcus Goncalves is spearheading this effort with a trio of clinical trials currently enrolling patients.
Matthew Vander Heiden Massachusetts Institute of Technology
Karen Vousden Francis Crick Institute
Metabolic Limitations in Cancer
All cells in the body, including cancer cells, exist in different metabolic environments, and thus have different nutritional resources available to them, said Mathew Vander Heiden. Cancer cells have especially high metabolic needs because by definition, they proliferate—doubling the mass of proteins, nucleic acids, and lipids in order to go from one cell to two. Understanding how different types of cancer cells reorganize their metabolic pathways to accomplish this feat can bring insight into the role metabolism plays in cancer therapeutics.
Tissues solve their metabolic needs differently. Metabolism in brain cells, for example, differs from that in liver cells. These differences must be reflected in the gene expression patterns of the tissues’ metabolic networks. When a cell becomes cancerous, it takes its existing metabolic network, based on its environment, and reorganizes it to support its proliferation. That’s why cancers exhibit varied metabolisms and are sensitive to different therapies, Vander Heiden said.
His lab studies how a cell’s environment constrains its metabolic network. In one recent study, the researchers characterized the nutrients available in the interstitial fluid of pancreatic and lung tumors in mice. Transplanting a tumor into a different tissue site changed the nutrients available to it. But the mutational profile of a tumor had a much weaker effect on its metabolism. More recent human data suggests that nutrient profiles in the interstitial fluid of kidney cells and kidney tumor cells is similar. The findings suggest that nutrient availability is an intrinsic property of different tissues, and that cancer must adapt to the food available in a specific area. “That’s a cool idea, because it has a profound effect on how we think about metastasis,” said Vander Heiden. His lab has conducted metabolic phenotyping of mouse cancers transplanted to different tissue sites and found that the most significant outlier in the metabolic profile is in the brain. That matches clinicians’ experience with HER2-positive breast cancer, which tends to respond well to treatment except when the cancer reaches the brain.
To probe tissue-intrinsic metabolic responses, researchers in the lab of Rakesh Jain, also at Harvard, transplanted breast cancer tissue either into mammary fat pads or the brains of mice. They then treated the muse with PI3K inhibitor and found that the tumor shrunk in the mammary tissue but not the brain. Related studies showed that the brain microenvironment partially drove differences in response.
Fatty acid synthesis is increased in the breast cancer tissue transplanted into mouse brain.
In collaboration with Jain’s lab, Vander Heiden’s team implanted tumor cells into the brains and mammary fat pad tissue of mice. They found different metabolic gene expression and activity between the two sites. They also found more glucose uptake into fatty acids in the brain tumor tissue than in the mammary tissue, suggesting that fatty acid synthesis increases in brain metastases. Ultimately, they showed that these differences affect how tumors develop in the brain and facilitate tumor growth. The microenvironment and nutritional differences of tumor cells is key to understanding varied responses to metabolism-targeted therapies.
Metabolic Control of Tumor Progression
P53 is a key tumor suppressor gene. Some seven million cancer patients per year carry a mutant version of it. However, virtually all tumors, even those with a normal p53 gene, have lost p53’s tumor suppressive activity, said Karen Vousden. Early research showed that p53 suppresses tumors but activates cell death and senescence, killing tumor cells. But more recently, Vousden and others report another type of p53 activity, in which it contributes to the cell’s survival and its ability to adapt to stressful conditions, such as the accumulation of reactive oxygen species (ROS). Her lab is examining the gene’s role in metabolic signals such as nutrient fluctuation and oxidative stress.
Next, they looked at p53’s role in adapting to nutrient starvation. Comparing cell lines with p53 to without it, they found that p53 mitigates oxidative stress caused by cells lacking amino acids the body normally produces, such as serine. The absence of serine had stronger negative effects in cells without p53, reflecting the gene’s ability to drive antioxidant defense, Vousden said.
To look at how p53 protects cells from oxidation, the researchers focused on a downstream enzyme called TIGAR. Although the protein’s full suite of functions is still unknown, it promotes the production of NADPH, a molecule that supports antioxidant activity. Cells that lack TIGAR are more sensitive to oxidative stress and can’t maintain NADPH levels. Studies suggest that this effect is specific to mitochondria. Organoids lacking TIGAR have a much higher level of toxic ROS than organoids expressing the enzyme, and treating the cells with mitochondrial antioxidants — but not membrane antioxidant — mitigates this effect.
p53 controls reactive oxygen species via downstream molecules, including TIGAR.
Vousden’s lab also studied the effects of TIGAR during various stages of tumor development. They found that loss of TIGAR impedes tumor growth in both the pancreatic and intestinal adenoma models. Interestingly, mice with these tumors also had an increased rate of lung metastasis. Treating these mice with antioxidants mitigated the effect on lung metastasis, suggesting that ROS influences cancer development differently at different stages of the disease.
Targeting Metabolism and Differentiation Therapy in AML
Acute Myeloid Leukemia (AML) is a critical disease area in Janssen’s drug discovery effort. A clinically heterogenous disease with multiple subtypes, AML is characterized by a diverse landscape of chromosomal abnormalities and gene mutations. One hallmark of AML, myeloid differentiation block, is linked to metabolism, said Ulrike Philippar. Myeloid differentiation in normal cells results in the generation of mature blood cells from hematopoietic stem cells. This process is blocked in AML cells, resulting in uncontrolled proliferation of progenitors, and ultimately, aggressive leukemia. Differentiation therapies have thus been a recent focus of drug development for AML, and unmet medical need for the disease remains high.
Philippar first described two US Food and Drug Administration-approved differentiation therapies. The first, called all-trans retinoic acid, or ATRA, was first used in 1985 to treat acute promyelocytic leukemia (APL), a subtype of AML. ATRA binds to a translocated protein in APL, leading to disruption of the differentiation block. The second therapy, IDH1 and IDH2 inhibitors, treats AML patients with an IDH mutation. The drug came to be used in combination with arsenic trioxide. These agents bind to a translocated protein in APL, disrupting the differentiation block.
Philippar and her team conducted two CRISPR/Cas9 screens to identify novel targets that inhibit proliferation in cell lines in vitro and/or in a bone marrow transplant mouse model in vivo. They identified dihydroorotate dehydrogenase (DHODH), an enzyme located in the mitochondrial membrane involved in pyrimidine synthesis. DHODH is also essential for the production of the nucleotide uridine monophosphate (UMP), the first building block in the production of RNA and DNA synthesis. Researchers, including Philippar and her colleagues, have shown that malignant cells, particularly AML cells, are metabolically dependent on pyrimidine production because they cannot salvage sufficient uridine from the extracellular environment.
Janssen identified a potent, selective proprietary DHODH inhibitor. The molecule strongly inhibited proliferation in 25 AML cancer cell lines. Janssen’s DHODH inhibitor also showed strong antiproliferative activity in primary AML samples. A few patient samples appeared to be resistant to the drug, and Janssen researchers will continue to investigate the basis of that resistance.
Preclinical studies suggest that DHODH inhibition is an effective treatment strategy for AML
The researchers also probed the DHODH inhibitor’s mechanism of action. The molecule induced cell cycle arrest in S phase, resulting in increased apoptosis. Furthermore, treatment with the DHODH inhibitor also increased cellular differentiation in AML cell lines. Adding excess uridine can reverse these effects, which validates the inhibitor’s on-target activity. Janssen’s DHODH inhibitor led to tumor regression in subcutaneous AML xenografts and increased the lifespan of mice bearing disseminated AML cancer cells approximately twofold.
Philippar said that DHODH inhibitor therapy takes away the pyrimidine source and leads to starvation of cells by forcing them into a fasting state. The researchers hypothesize that normal cells can tolerate fasting and recover; however, leukemia and other cancer cells cannot tolerate fasting because they require higher pyrimidine synthesis. Hence, prolonged pyrimidine starvation of leukemia cells results in a cell fate decision leading to differentiation, and ultimately, apoptosis and tumor regression.
DHODH confirms the relevance of metabolic targets for AML. However, the clinical diversity of AML may be a challenge for metabolic drugs, and researchers will have to better understand patient selection as they develop these drugs in the clinic.
Lewis C. Cantley, PhD Weill Cornell Medical College
Matthew Vander Heiden, PhD Massachusetts Institute of Technology
Karen Vousden, PhD Francis Crick Institute
Ulrike Philippar, PhD Johnson & Johnson
Panelists reflected on what moderator Costas Lyssiotis, a former postdoctoral student in Cantley’s laboratory, framed as a key theme of the symposium: harnessing diet as a way to treat cancer.
Cantley stated that previous efforts to use diet to understand cancer have largely failed, because getting trial participants to adhere to a diet and accurately report what they eat is nearly impossible. In a recent clinical trial that tested a ketogenic diet as a neoadjuvant for endometrial cancer, Cantley’s collaborator Marcus Goncalves used the metabolic kitchen at Weill Cornell Medical School to generate 21 meals per week for each trial participant. Weekly blood draws made it possible for researchers to be sure patients adhered to the diet. “To show [that] a dietary intervention works, you have to do it in a very controlled manner, not just tell people what they should eat and hope that they find it at the market,” Cantley said. Ultimately, the aim is to get dietary interventions approved as therapies so that insurance companies will cover them. He noted that companies such as Bayer and Novartis are excited by the possibility because it can potentiate pharmaceutical therapies. “We should hold diets to the same standards we hold drugs to,” said Cantley.
Vander Heiden agreed about the need for more rigorous dietary trials. His lab found that alterations in diet led to changes in the composition of metabolic molecules in interstitial fluid, and presumably those changes extend to nutrients available in different tissues and in blood. Animal studies are especially useful for identifying how nutrient availability changes with diet, he said; and they might point to unexpected differences beyond those already identified for the glucose and insulin pathway and serine.
Translating dietary interventions from animals to humans may not be straightforward. “What happens in mice doesn’t necessarily happen in people,” Vousden said. She and her colleagues are developing a serine-depleted diet for people with cancer, along the lines of low phenylalanine diets prescribed for children with phenylketonuria, an inherited condition that raises the level of the amino acid building block phenylalanine in the blood.
Any nutrient may be modulated for cancer therapy, but researchers must have a firm mechanistic understanding of why that alteration might work. “I think that’s at the heart of whether it will be successful, and whether people will follow these diets,” Vousden said. A solid grounding in the mechanism underlying a possible dietary intervention’s effect can help motivate patients to follow a restrictive diet for their clinical benefit.
Philippar noted that so far, considering diet is not standard practice in industry drug development. But it is potentially important, not just for targets that modulate the metabolism but also for immunological approaches. Researchers now know that diet can affect gut bacteria, which might in turn affect drug response. In clinical trials run by Janssen, she explained, researchers generally check whether a patient is in a fed or fasted state when receiving an experimental therapy, but they do not check on what exactly the patient has eaten.
Single-cell RNA sequencing makes it possible to look at the tumor microenvironment and to determine how changes in diet affect the expression levels of immune genes. In 2019, Cantley, Goncalves and others showed that just one, 12-ounce serving of a sugary drink such as orange juice, apple juice, or soda each day enhances intestinal tumors in mice. A combination of fructose and glucose causes the effect, the researchers found, and knocking out a molecule called ketohexakinase, which phosphorylates fructose, limits tumor growth. The connection to sugary drinks “is very surprising,” says Cantley, “but I think it may explain why we’ve almost tripped the rate of early onset colorectal cancer in young adults in the last 20 years.”
Artificial intelligence is quickly becoming a ubiquitous part of our daily lives. What can we expect as this technology continues to grow? And how will it impact you?
Published September 14, 2020
By Liang Dong
Alexandra Boltasseva, PhD
From virtual assistants like Siri to self-driving cars and computer-aided medical diagnoses, artificial intelligence (AI) affects our lives with unprecedented speed. Slowly but steadily, scientists in a broad range of fields have started to embrace AI in their research, hoping to significantly reduce the time needed to achieve new discoveries. This trend has become more obvious in the physical sciences, and in the field of materials science in particular, which is focused on the discovery and production of new, advanced materials imbued with desirable properties or functions. Think: screens of foldable smartphones; batteries that power electric cars; or materials that bend light around them, rendering them invisible.
How exactly could AI help materials scientists? We recently interviewed three honorees of the Blavatnik Awards for Young Scientists, Alexandra Boltasseva, PhD, Professor of Electrical and Computer Engineering at Purdue University; Léon Bottou, PhD, Principal Researcher at Facebook AI Research; and Sergei V. Kalinin, PhD, Corporate Fellow at Oak Ridge National Laboratory, who are contributing to an upcoming virtual symposium on October 6 and 7, AI for Materials: From Discovery to Production. Here’s what they had to say about the opportunities, as well as the challenges, in this rising field.
It is only recently that researchers in the physical sciences, like materials scientists, have begun to incorporate AI techniques into their work. Why do we need to take advantage of AI for this field? What benefits may AI offer materials science?
Kalinin
Sergei V. Kalinin, PhD
AI offers a set of powerful tools to explore large volumes of multidimensional data in the physical sciences, and promises to uncover hidden functional relationships between the physical properties that we can observe. As such, AI methods are poised to become an inseparable part of all physical sciences, to enable discovery and hypothesis-driven research and to guide planning of experiments. We can take advantage of a broad range of AI techniques—from multivariate statistics to convolutional networks, unsupervised and semi-supervised methods, Gaussian processing, and reinforcement learning.
In addition, the proliferation of laboratory automation in areas from materials synthesis to imaging of materials’ molecular structures opens up broad opportunities for AI-driven experiments. For example, we will be able to adopt large-scale robotic systems or the microscale lab-on-a-chip platforms in our experiments, producing thousands or more materials in a single process.
Boltasseva
My own field, photonics, has truly been transformed by the concept of “inverse design,” meaning scientists input desired performances of photonic systems into computers and run physics-informed algorithms to figure out the best possible optical designs. The daunting challenge of this field lies in the inconceivably high computational power required for an exhaustive search within the extremely large, hyper-dimensional space of optical design parameters and constituent materials. Merging AI techniques with photonics is expected to not only enhance and enrich the design space, but, most importantly, to unlock novel functionalities and bring about disruptive performance improvements.
As compared to life sciences and pharmaceutical sciences, the application of AI in physical sciences is at least 10 years behind. What do you think is the biggest challenge for applying AI in physical sciences? How could the AI and physical sciences communities work together to address these challenges?
Bottou
Léon Bottou, PhD
Using machine learning in physical sciences is not an obvious proposition. Recent advances in AI have shown how tasks in computer science, such as computer vision and machine translation, can be achieved using big data. Yet it would be unwise to claim that this success can be replicated in all scientific fields. Big data only reveal statistical correlations that are not always indicative of the causal relations that physicists often seek. To solve this question, the AI and physics communities may take the strategy of defining a hierarchy of problems for which one could envision using AI, such as:
Visualizing or measuring an ongoing physical phenomenon. These problems are the most accessible to AI/machine learning because they can directly leverage recent advances in computer vision and signal analysis in collecting data from physical experiments and computations.
Explaining a physical phenomenon. These problems belong to the next rung of difficulty because we need AI/machine learning systems that incorporate enough of our current knowledge of physics, and can then clarify the phenomenon of interest by constructing something interpretable on top of our current knowledge.
Designing a physical system that leverages a certain phenomenon in new ways. These are by far the most difficult problems, because they require AI/machine learning systems to accurately predict how the physical phenomenon will be affected by changes that are not included or prominent in the experimental data on which AI models have been trained.
Boltasseva
The physical sciences community should ultimately build extensive databases to unleash the power of AI. We should even set up an ‘optical structures and materials genome’ project to construct a comprehensive dataset of photonic concepts, architectures, components, and photonic materials to enable hierarchical machine learning algorithms that could provide ultimate-efficiency devices.
Kalinin
I agree with Alexandra. AI tends to proliferate in the communities that adopt the model of open sharing of codes and data. While some areas of physics research have undergone this transformation, many more require both enabling tools and proof-of-benefit to accelerate this process.
I also want to add on to Léon’s comment on the fundamental difference between the AI and physics communities. AI starts with purely correlative models, and tends to rely on big data. In comparison, research in physical sciences is strongly based on prior knowledge to explore the cause and effect relationships, and often assumes the presence of simple rules or descriptors that can give rise to complex behaviors in macroscopic systems. Experiments in physical sciences can give rise to huge data volumes, but these data can pertain only to one specific situation of the system and hence are not “big.”
In order to further leverage the benefits of AI in physical sciences, researchers have to possess both sufficient domain knowledge in physical sciences and expertise in machine learning, or forge robust interdisciplinary collaborations. Conferences like AI for Materials will help researchers in both fields form these kinds of interdisciplinary teams.
The 2020 Innovators in Science Award winners include a biochemist/molecular geneticist from Cold Spring Harbor Laboratory and brain disorder researcher from the Korea Advance Insitute of Science and Technology.
New York, NY | July 8, 2020 and Osaka, Japan | July 8, 2020 – Takeda Pharmaceutical Company Limited (“Takeda”) (TSE:4502) and the New York Academy of Sciences announced today the Winners of the third annual Innovators in Science Award for their excellence in and commitment to innovative science that has significantly advanced the field of rare disease research. Each Winner receives a prize of US $200,000.
Senior Scientist Award: Adrian R. Krainer
The 2020 Winner of the Senior Scientist Award is Adrian R. Krainer, Ph.D., St. Giles Foundation Professor at Cold Spring Harbor Laboratory. Prof. Krainer is recognized for his outstanding research on the mechanisms and control of RNA splicing, a step in the normal process by which genetic information in DNA is converted into proteins. Prof. Krainer studies splicing defects in patients with spinal muscular atrophy (SMA), a devastating, inherited pediatric neuromuscular disorder caused by loss of motor neurons, resulting in progressive muscle atrophy and eventually, death. Prof. Krainer’s work culminated notably in the development of the first drug to be approved by global regulatory bodies that can delay and even prevent the onset of an inherited neurodegenerative disorder.
“Collectively, rare diseases affect millions of families worldwide, who urgently need and deserve our help. I’m extremely honored to receive this recognition for research that my lab and our collaborators carried out to develop the first approved medicine for SMA,” said Prof. Krainer. “As basic researchers, we are driven by curiosity and get to experience the thrill of discovery; but when the fruits of our research can actually improve patients’ lives, everything else pales in comparison.”
Early-Career Scientist Award: Jeong Ho Lee
The 2020 Winner of the Early-Career Scientist Award is Jeong Ho Lee, M.D., Ph.D, Associate Professor, Korea Advanced Institute of Science and Technology (KAIST). Prof. Lee is recognized for his research investigating genetic mutations in stem cells in the brain that result in rare developmental brain disorders.
He was the first to identify the causes of intractable epilepsies and has identified the genes responsible for several developmental brain disorders, including focal cortical dysplasias, Joubert syndrome—a disorder characterized by an underdevelopment of the brainstem—and hemimegalencephaly, which is the abnormal enlargement of one side of the brain. Prof. Lee also is the Director of the National Creative Research Initiative Center for Brain Somatic Mutations, and Co-founder and Chief Technology Officer of SoVarGen, a biopharmaceutical company aiming to discover novel therapeutics and diagnosis for intractable central nervous system (CNS) diseases caused by low-level somatic mutation.
“It is a great honor to be recognized by a jury of such globally respected scientists whom I greatly admire,” said Prof. Lee. “More importantly, this award validates research into brain somatic mutations as an important area of exploration to help patients suffering from devastating and untreatable neurological disorders.”
The 2020 Innovators in Science Award Ceremony and Symposium
The 2020 Winners will be honored at the virtual Innovators in Science Award Ceremony and Symposium in October 2020. This event provides an opportunity to engage with leading researchers, clinicians and prominent industry stakeholders from around the world about the latest breakthroughs in the scientific understanding and clinical treatment of genetic, nervous system, metabolic, autoimmune and cardiovascular rare diseases.
“At Takeda, patients are our North Star and those with rare diseases are often underserved when it comes to the discovery and development of transformative medicines,” said Andrew Plump, M.D., Ph.D., President, Research & Development at Takeda. “Insights from the ground-breaking research of scientists like Prof. Krainer and Prof. Lee can lead to pioneering approaches and the development of novel medicines that have the potential to change patients’ lives. That’s why we are proud to join with the New York Academy of Sciences to broadly share and champion their work — and hopefully propel this promising science forward.”
“Connecting science with the world to help address some of society’s most pressing challenges is central to our mission,” said Nicholas Dirks, Ph.D., President and CEO, the New York Academy of Sciences. “In this third year of the Innovators in Science Award we are privileged to recognize two scientific leaders working to unlock the power of the genome to bring innovations that address the urgent needs of patients worldwide affected by rare diseases.”
About the Innovators in Science Award
The Innovators in Science Award grants two prizes of US $200,000 each year: one to an Early-Career Scientist and the other to a well-established Senior Scientist who have distinguished themselves for the creative thinking and impact of their research. The Innovators in Science Award is a limited submission competition in which research universities, academic institutions, government or non-profit institutions, or equivalent from around the globe with a well-established record of scientific excellence are invited to nominate their most promising Early-Career Scientists and their most outstanding Senior Scientists working in one of four selected therapeutic fields of neuroscience, gastroenterology, oncology, and regenerative medicine.
Prize Winners are determined by a panel of judges, independently selected by The New York Academy of Sciences, with expertise in these disciplines. The New York Academy of Sciences administers the Award in partnership with Takeda.
For more information please visit the Innovators in Science Award website.
About Takeda Pharmaceutical Company Limited
Takeda Pharmaceutical Company Limited (TSE:4502/NYSE:TAK) is a global, values-based, R&D-driven biopharmaceutical leader headquartered in Japan, committed to bringing Better Health and a Brighter Future to patients by translating science into highly-innovative medicines. Takeda focuses its R&D efforts on four therapeutic areas: Oncology, Rare Diseases, Neuroscience, and Gastroenterology (GI).
We also make targeted R&D investments in Plasma-Derived Therapies and Vaccines. We are focusing on developing highly innovative medicines that contribute to making a difference in people’s lives by advancing the frontier of new treatment options and leveraging our enhanced collaborative R&D engine and capabilities to create a robust, modality-diverse pipeline. Our employees are committed to improving quality of life for patients and to working with our partners in health care in approximately 80 countries. For more information, visit https://www.takeda.com.
On March 5, 2020, the New York Academy of Sciences celebrated the Laureates and Finalists and winners of the 2020 Blavatnik Awards for Young Scientists in the United Kingdom. The one-day symposium featured fast-paced, engaging research updates from nine scientists working in diverse fields within life sciences, chemistry, and physical sciences and engineering. This year’s Blavatnik UK honorees are probing the deepest mysteries ranging from the universe to the human mind, tackling longstanding questions that have occupied scientists and philosophers for millennia. Is there life beyond our Solar system? How is knowledge organized in the brain? What is the fundamental nature of gravity? Find out how this game-changing group of young scientists is working to answer these questions in this summary of the symposium.
Symposium Highlights
Environmental factors can influence the defense strategies bacteria use to fend off invading viruses. Insights into this process are advancing the potential for phage therapy as an alternative to antibiotics.
New analytical and computational tools are revealing the neural machinery that allows the brain to create models of the world and facilitates decision-making and behavior.
Chemists can exploit chirality to create novel molecules with a wide variety of applications in drug design, consumer electronics, and catalysis.
The scientific community is closer now than ever to realizing the commercial potential of nuclear fusion as a source of clean energy.
The first viable theory of massive gravity might help explain some of the biggest mysteries in physics, including the accelerated expansion of the universe.
Hosted By
Victoria Gill Science Correspondent BBC News
Speakers
Tim Behrens, DPhil University of Oxford and University College London
Ian Chapman, PhD UK Atomic Energy Authority
Matthew J. Fuchter, PhD Imperial College London
Stephen M. Goldup, PhD University of Southampton
Kirsty Penkman, PhD University of York
Claudia de Rham, PhD Imperial College London
Eleanor Stride, PhD University of Oxford
Amaury Triaud, PhD University of Birmingham
Edze Westra, PhD University of Exeter
Program Supporter
Changing the Game in Life Sciences
Speakers
Eleanor Stride, PhD University of Oxford
Edze Westra, PhD University of Exeter
Tim Behrens, DPhil University of Oxford & University College London
Engineering Bubbles
Mechanical engineer Eleanor Stride never planned to design drug delivery systems. She was “convinced I wanted to spend my career designing Aston Martins,” until a chance discussion with a supervisor piqued her interest in therapeutic applications of engineered microbubbles. Just two microns in diameter, microbubbles can be used as ultrasound contrast agents, but Stride sees a role for these tiny tools in the fight against cancer. “In many cases, the problem with cancer drugs [is] how we deliver them,” she said, explaining that systemic chemotherapy agents often cannot penetrate far enough into tumors to be effective. These drugs can also cause side effects and damage healthy tissues.
Microbubbles can help sidestep these challenges, safely encapsulating drug molecules within a stabilizing shell. The shell can be functionalized with magnetic nanoparticles, allowing clinicians to direct the bubbles’ aggregation at tumor sites and visualize them with ultrasound. As the bubbles compress and release in response to the ultrasound beam, the oscillation helps the bubbles penetrate into the surrounding tissue. “If we increase the ultrasound energy, we can destroy the bubble, allowing us to release the drugs on demand,” said Stride, noting that molecules released from a single 2-micron microbubble can circulate up to 100 times that diameter, pumping drugs deep into tumor tissues. This approach is highly localized—drugs are only released at the tumor site—which eliminates the potential for systemic toxic effects.
Ultrasound-stimulated oscillation of microbubbles creates a vortex in surrounding fluids. The vortex pumps drug molecules deep into tumor sites.
In 2019, Stride and a team of collaborators published the results of trials using oxygen-loaded magnetic microbubbles to treat malignant pancreatic tumors. In animal models, tumors treated with microbubble-delivered drugs showed dramatic spikes in cell death and also shrank in size, “which can mean the difference between a surgeon being able to remove a tumor or not,” said Stride. Additional experiments have helped hone techniques for external magnetic control of microbubbles within blood vessels to ensure precise, targeted drug delivery—a critical step toward tailoring this method for use in humans. Stride and her collaborators aim to launch a clinical trial in pancreatic cancer patients “in the very near future.”
Insights From Bacteria-Phage Interactions
As the fight against viruses dominates the news cycle, 2020 Blavatnik Awards UK Finalist Edze Westra shared an update from the front lines of a viral war billions of years in duration: the “evolutionary arms race” between bacteria and the viruses that infect them, called phages. The interactions between bacteria and phages—the most abundant biological entities on Earth—have profound implications for the development of phage-based therapies as alternatives to antibiotics.
Phages are often successful killers, but bacteria have evolved sophisticated immune strategies to resist attacks. Understanding how and when bacteria deploy each of these defensive tactics is key to designing phage therapies to treat bacterial infections.
Like humans, bacteria utilize both innate and adaptive immune responses to invading pathogens. In bacteria, innate immunity relies on the modification of surface structures to prevent phages from attaching. This system is effective, yet it creates no “record,” or memory, of which phages it encounters. The adaptive immune system, however, allows bacteria to build a database of previously encountered pathogens in the form of bits of genetic material snipped from invading phages and incorporated into the bacterium’s own DNA. The adaptive immune system, known as CRISPR immunity, forms the basis of CRISPR-Cas genome editing techniques. “There’s a critical balance between these two systems, and both are critical for survival,” said Westra, whose research aims to determine the factors that influence whether a bacterium mounts an innate or adaptive immune defense against a particular phage.
Using Pseudomonas aeruginosa, an antibiotic-resistant pathogen that often infects cystic fibrosis patients, Westra determined that a bacterium’s environment—specifically, the level of available nutrients—determined which defensive strategy was utilized. In high-nutrient environments, almost all bacteria deployed an innate immune response to phage attacks, whereas in lower nutrient settings, CRISPR immunity dominated.
The level of available nutrients influences which immune strategy bacteria use to defend against phage attacks.
In experiments using moth larvae, Westra discovered that infections were more severe when bacteria utilized CRISPR immunity, whereas bacteria that evolved innate immunity often caused less aggressive infections. “If we can manipulate how bacteria evolve resistance to phages, this could potentially revolutionize the way we approach antimicrobial resistance, with major benefits to our healthcare,” Westra said.
Building Models of the World
Computational neuroscientist Timothy Behrens is fascinated with the basic functions and decisions of everyday life—the process of navigating our home or city, the steps involved in completing household tasks, the near-subconscious inferences that inform our understanding of the relationships between people and things. Behrens designs analytical tools to understand how neuronal activity in the brain gives rise to these thought processes and behaviors, and his research is illuminating how knowledge is organized in the brain.
The activities of grid cells and place cells are well understood. By creating spatial maps of the world, grid and place cells allow us to navigate familiar spaces and locate items, such as car keys. Behrens explained that much less is known about how the brain encodes non-spatial, abstract concepts and sequence-based tasks, such as loading, running, and emptying a dishwasher. Over the past several years, Behrens and his collaborators have demonstrated that abstract information is similarly mapped as grid-like codes within the brain. “On some level, all relational structures are the same, and all are handled by the same neural machinery,” he said. This insight helps explain the effects of diseases like Alzheimer’s, which targets grid and place cells first and impacts both spatial and non-spatial knowledge.
Relational information is encoded by the same neural machinery that encodes spatial and navigational maps.
In another line of research, Behrens is probing a phenomenon called replay, during which the brain revisits recent memories as a means to consolidate knowledge about current events and anticipate future ones. Behrens illustrated the concept by showing patterns of neuronal activity as a rat runs around a track, then rests. Even at rest, the rat’s brain displays millisecond-long flashes of neuronal activity that mimic those that take place during running. “He’s not running down the track anymore, but his brain is,” said Behrens. Replay also underlies the human ability to understand a simple story even when it’s told in the wrong order. “Our knowledge of the world tells us…what the correct order is, and replay will rapidly stitch together the events in the correct order.”
Computational tools developed in Behrens’ lab have been shared with thousands of scientists around the globe as they pursue new hypotheses about the neural computations that control cognition and behavior. “It’s an exciting time to be thinking about the brain,” Behrens said.
Exploiting Molecular Shape to Develop Materials and Medicines
Consider the handshake: a greeting so automatic it takes place without thinking. Two right hands extend and naturally lock together, but as Matthew Fuchter explained, that easy connection becomes impossible if one party offers their left hand instead. The fumbling that ensues stems from a type of asymmetry called chirality. Chiral objects, such as hands, are mirror-image forms that cannot be superimposed or overlapped, and when one chiral object interacts with another, their chirality dictates the limits of their interaction. Chirality can be observed throughout nature, from the smallest biological molecules to the structures of skyscrapers.
In organic chemistry, molecular chirality can be exploited to tremendous advantage. Fuchter explained that the shape of molecules “is not only critical for their molecular properties, but also for how they interact with their environment.” By controlling subtle aspects of molecular shape, Fuchter is pioneering new strategies in drug design and devising solutions to technological problems that plague common electronic devices.
The notion of pairing complementary molecular geometries to achieve a specific effect is not unique to drug design—such synchronicities can be found throughout nature, including in the “lock and key” structure of enzymes and their substrates. Fuchter’s work aims to invent new drug molecules with geometries perfectly suited to bind to specific biological targets, including those implicated in diseases such as malaria and cancer.
Only one of these two chiral molecules has the correct orientation, or “handedness” to bind to the receptor site on the target protein.
Fuchter is also exploring applications for chirality in a field where the concept is less prominent—consumer electronics. Organic LED, or OLED, technology has “revolutionized the display industry,” allowing manufacturers to create ultra-thin, foldable screens for smartphones and other displays. Yet these features come at a steep efficiency cost—more than half of the light generated by OLED pixels is blocked by anti-glare filters added to the screens to minimize reflectiveness. A novel solution, in the form of chiral molecules bound to non-chiral OLED-optimized polymers, induces a chiral state of light called circularly polarized light. These circularly polarized, chiral light molecules are capable of bypassing the anti-glare filter on OLED screens. Fuchter noted that displays are far from the only technology that stands to be impacted by the introduction of chiral molecules. “Our research is generating new opportunities for chiral molecules to control electron transport and electron spin, which could lead to new approaches in data storage,” he said.
Making Use of the Mechanical Bond
Most molecules are bound by chemical bonds—strong, glue-like connections that maintain the integrity of molecules, which can be both simple, such as hydrogen, and highly complex, such as DNA. 2020 Blavatnik Awards UK Finalist Stephen Goldup’s work focuses on a less familiar bond. Mechanical bonds join molecules in a manner akin to an interconnected chain of links—the components retain movement, yet cannot separate.
Mechanically interlocked molecules have the potential to yield materials with “exciting properties,” according to Goldup, but in the decades since they were first synthesized, they have largely been regarded as “molecular curiosities.” Goldup’s lab is working to push these molecules beyond the laboratory bench by characterizing the properties of interlocked molecules and probing their potential applications in unprecedented ways. His work focuses on two types of mechanically bound molecules—catenanes, in which components are linked together like a chain, and rotaxanes, which consist of a ring component threaded through a dumbbell-shaped axle.
Goldup’s lab has taken cues from nature to introduce additional elements into rotaxanes, resulting in novel molecules with a variety of potential applications. For example, much as enzymes contain “pockets” within which small molecules can bind, rotaxanes too contain a space that can trap a molecule or ion of interest. Rotaxanes that bind metal ions have unique magnetic and electronic properties that could be used in memory storage devices or medical imaging. Inspired by proteins and enzymes that bind DNA, Goldup’s lab has also designed rotaxanes in which DNA itself is the “axle.” In theory, these molecules can be used to effectively “hide” portions of DNA and alter its biological behavior.
Just as enzymes bind small molecules with their structures, rotaxanes can bind molecules in the cavity between the ring and the axle.
Perhaps most significantly, Goldup’s lab has solved a longstanding obstacle to studying rotaxanes: the difficulty of making them. The problem lies in the fact that rotaxanes can be chiral even when their components are not, making it extremely challenging to synthesize a distinct “hand,” or version, of the molecule. Recalling Matthew Fuchter’s example of how an awkward left-hand/right-hand handshake differentiates the “handedness” of two chiral objects, Goldup explained how his lab developed a technique for synthesizing distinctly “left” or “right” handed rotaxanes by utilizing a chiral axle to build the molecules. “Our insight was that by making the axle portion chiral on its own, when we thread the axle into the ring, the rotaxanes we make are no longer mirror-images of each other. They have different properties, and they can now be separated,” he said. Once separate, the chiral portion of the axle can be chemically removed and replaced with other functional groups.
Goldup’s lab is conducting experiments with new mechanically-locked molecules—including chiral rotaxane catalysts— to determine where they may outperform existing catalysts.
Amino Acids as a Portal to the Past
Scientists have multiple methods for peering into the history of Earth’s climate, including sampling marine sediment and ice cores that encapsulate environmental conditions stretching back millions of years. “But this is an incomplete picture—akin to a musical beat with no notes,” said Kirsty Penkman, the 2020 Blavatnik Awards UK Laureate in Chemistry. The records of life on land—fossil records—provide “the notes to our tune, and if we know the timing, that gives us the whole melody,” she said. Archaeologists, paleontologists, and climate scientists can harmonize fossil records with climate history to understand the past, yet their efforts stall with fossils older than 50,000 years—the limit of radiocarbon dating.
Penkman’s lab is developing dating methods for organic remains that reach far deeper into the history of life on Earth. Their strategy relies not on the decay of carbon, but the conversion of amino acid molecules from one form to another. Continuing the theme of chirality from previous presentations, Penkman explained that amino acids exist in two mirror-image forms. However, the body only synthesizes amino acids in the “left-handed,” or L-form. This disequilibrium shifts after death, when a portion of L-amino acids begins a slow, predictable conversion to the right-handed, or D-form. The older the fossil, the greater the balance between D and L isomers. This conversion process, called racemization, was first proposed as a dating method in the 1960s. Yet, it became clear that some of the fossil amino acids were vulnerable to environmental factors that impact the racemization rate, and therefore the date.
About 15 years ago, Penkman discovered that minute stores of proteins within the remains of snail shells are entrapped in intracrystalline voids. These tiny time capsules are unaffected by environmental factors. Studies have since confirmed that shells found in older horizons, for example deeper underground, contain higher ratios of D-amino acids versus those found at younger sites, thus validating the technique.
Calcitic snail shells found at older horizons have higher ratios of D-amino acids than those found at younger horizons.
Snail shells are often found in archeological sites, a serendipity that has led to astonishing findings about early human migration. Shells found alongside several Paleolithic tools “dated as far back as 700,000 years,” according to Penkman. “We’ve successfully shown that early humans were living in Northern Europe 200,000 years earlier than previously believed,” she said.
Penkman’s team has analyzed remains of ostrich eggshells at some of the earliest human sites in Africa, discovering fully preserved, stable sequences of proteins in shells dating back 3.8 million years. Mammalian remains are the next frontier for Penkman’s lab. They have analyzed amino acids in ancient tooth enamel—including that of a 1.7-million-year-old rhinoceros—and are developing microfluidic techniques to sample enamel from early human remains.
Changing the Game in Physical Sciences and Engineering
Speakers
Amaury Triaud University of Birmingham
Ian Chapman UK Atomic Energy Authority and Culham Centre for Fusion Energy
Claudia de Rham Imperial College London
Worlds Beyond Our Solar System
For millennia, humans have wondered whether life exists beyond our planet. Amaury Triaud, 2020 Blavatnik Awards UK Finalist believes we are closer to answering that question now than at any other time in history. The study of exoplanets—planets that orbit stars other than the Sun—offers what Triaud believes is “the best hope for finding out how often genesis happens, and under what conditions.”
The search for exoplanets has revealed remarkable variety among stars and planets in our galaxy. “The universe is far more surprising and diverse than we anticipated,” said Triaud. Astronomers have identified thousands of exoplanets since 1995, and now estimate that there are more planets in the Milky Way than stars—”something we had no idea about ten years ago,” Triaud said. Many exoplanets orbit stars so much smaller than the Sun that these stars cannot be seen with the naked eye. Yet these comparatively small stars provide “optimal conditions” for exoplanet hunters.
Exoplanets are often detected using the transit method—as an orbiting planet passes in front of a star, its shadow temporarily dims the star’s brightness. The larger the planet relative to the star, the greater its impact on the brightness curve and the easier for astronomers to detect. While monitoring a small star 39 light-years from Earth, TRAPPIST-1, a team of astronomers, including Triaud, discovered an exoplanet system comprised of seven rocky planets similar in size to Earth, Venus, and Mercury.
“The next question is to find out whether biology is happening out there,” said Triaud, joking that the biology of interest is not little green men, but rather green algae or microbes similar to the ones that fill our atmosphere with oxygen. The presence of oxygen “acts like a beacon through space, broadcasting that here on Earth, there is life,” said Triaud, explaining that the only way to gauge the presence of life on exoplanets is through atmospheric analysis. Using transmission spectroscopy, Triaud and other astronomers will look for exoplanets that possess an atmosphere and chemical signatures of life, such as oxygen, ozone, or methane, in the atmospheric composition of exoplanets.
Measurements of spectral signatures in a planet’s atmosphere can reveal the presence of gases associated with life, including oxygen and methane.
Such analyses will begin with the launch of the James Webb telescope in 2021. In the meantime, a land-based mission called Speculoos, based partially in Chile’s Atacama desert, is monitoring 1,400 stars in search of additional exoplanets. “It’s rather poetic that from one of the most inhospitable places on Earth, we are on the path to investigating habitability and the presence of life in the cosmos,” Triaud said.
The Path to Delivering Fusion Power
“There’s an old joke that nuclear fusion is 30 years away and somehow always will be,” said 2020 Blavatnik Awards UK Finalist Ian Chapman, but he insists that the joke will end soon. According to Chapman, the “ultimate energy source” is entering the realm of reality. “We’re now in the delivery era, where fusion lives up to its potential,” he said. Low-carbon, low-waste, capable of producing tremendous amounts of energy from an unlimited fuel source—seawater—and far safer than nuclear fission, fusion power has a long list of desirable qualities. Chapman is the first to acknowledge that fusion is “really hard,” but his work is helping to ease the challenges and bring a future of fusion into focus.
Nuclear fusion relies on the collision of two atoms—deuterium, or “heavy” hydrogen, and tritium, an even heavier isotope of hydrogen. Inside the Sun, these atoms collide and fuse, producing the heat and energy that powers the star. Replicating that process on Earth requires enough energy to heat the fuel. of deutrium and tritium gases to temperatures ten times hotter than the Sun, a feat that Chapman admits “sounds bonkers, but we do it every day.”
Within fusion reactors called tokamaks, this superhot fuel is trapped between arrays of powerful magnets that “levitate” the jet as it spins around a central magnetic core, preventing the fuel from melting reactor walls. Yet this is an imperfect process, explained Chapman, and due to fuel instabilities, eruptions akin to “throwing a hand grenade into the bottom of the machine” happen as often as once per second. Chapman devised a method based on his numerical calculations for preventing these eruptions using additional magnet arrays that induce three-dimensional perturbations, or “lobes” at the edge of the plasma stream. Just as a propped-open lid on a pot of boiling water allows steam to escape, these lobes provide a path to release excess pressure.
An array of magnets near the plasma edge creates perturbations in the fuel stream, allowing pressure to escape safely.
Chapman’s technique has been incorporated into the “the biggest scientific experiment ever undertaken by humankind”—a massive tokamak called ITER, roughly the size of a football stadium and equipped with a central magnet strong enough to lift an aircraft carrier. Scheduled to begin producing power in 2025, ITER aims to demonstrate the commercial viability of nuclear fusion. “We can put 50 megawatts of power into the machine, and it produces 500 megawatts of power out,” said Chapman. “That’s enough to power a medium-sized city for a day.”
Even before ITER’s completion, Chapman and others are setting their sights on designing less expensive fusion devices. Late last year, the UK committed to building a compact tokamak that offers the benefits of fusion with a smaller footprint, and Chapman is the leader of this project.
The Nature of Gravity
Claudia de Rham, the 2020 Blavatnik Awards UK Laureate in Physical Sciences and Engineering, concluded the day’s research presentations with an exploration of nothing less than “the biggest mystery in physics today.” For decades, cosmologists and physicists have grappled with discrepancies between observations about the universe—for example, its accelerated expansion— and Einstein’s general theory of relativity, which dictates that gravity should gradually slow that expansion. “The universe is behaving in unexpected ways,” said de Rham, whose efforts to resolve this question stand to profoundly impact all areas of physics.
Understanding the fundamental nature of gravity is key to understanding the origin and evolution of the universe. As de Rham explained, gravity can be detected in the form of gravitational waves, which are produced when two black holes or neutron stars rotate around each other, perturbing the fabric of spacetime and sending rippling waves outward like a stone tossed into a pond. But gravity can also be represented as a fundamental particle, the graviton, similar to the way light can be considered as a particle, the photon, or an electromagnetic wave. Unlike the other fundamental particles such as the photon, the electron, the neutrino, or even the famously elusive Higgs boson, the graviton has never been observed. In theory, the graviton would, like all fundamental particles, exist even in a perfect vacuum, a phenomenon known as vacuum quantum fluctuation. Unknown in Einstein’s day, vacuum quantum fluctuations, when factored into the general theory of relativity, do predict an accelerated expansion of the universe. “That’s the good news,” said de Rham. “The bad news is that the predicted rate of expansion is too fast by at least 28 orders of magnitude.”
This raises the possibility that “general relativity may not be the correct description of gravity on large cosmological scales,” said de Rham. If the graviton had mass, however, it would impact the behavior of gravity on the largest scales and could explain the observed rate of expansion.
Signal patterns from gravitational wave events can serve as models for estimating the mass of the graviton. By comparing the expected signals produced by either a massless particle or a high-mass particle with actual signal patterns from detected events, physicists can place an upper and lower boundary on the graviton’s potential mass.
The idea of a massive graviton has been considered—and refuted—by physicists as far back as the 1930s. Several years ago, de Rham, along with collaborators Andrew Tolley and Gregory Gabadadze, “realized a loophole that had evaded the whole community.” Together, they derived the first theory of massive gravity. “Through gravity, we can now connect small vacuum fluctuations with the acceleration of the universe, linking the infinitely small with the infinitely large,” de Rham said.
Determining the mass of the graviton requires the most precise scale imaginable, and de Rham believes that gravitational wave observatories are perfectly suited to the task. Whether her theory will hold up in future tests remains to be seen, but when it comes to solving this epic mystery, “the possibility is now open.”
Several Laureates and Finalists of the 2020 Blavatnik Awards in the UK joined BBC science reporter Victoria Gill for the final session of the day, a wide-ranging panel discussion that touched on issues both current and future-looking.
Two themes—fear and opportunity— emerged as powerful forces shaping science and society, especially as it relates to climate change and the threat of emerging infectious disease. Gill noted that climate change is “the biggest challenge ever to face humanity,” and that many efforts to raise awareness of its impacts focus on bleak projections for the future. Asked for insights on shifting the tone of climate change communications, Kirsty Penkman acknowledged that “there needs to be a certain level of fear to get people’s attention.” She then advocated for a solutions-oriented plan rooted in the fast pace of scientific progress in clean energy, among other areas. “This is an amazing opportunity,” she said. “Humans are ingenious….in the last 120 years we’ve moved from a horse-drawn economy to a carbon-based economy, and in 5 or 20 years we could be in a fusion-based economy. We have the potential to open up a whole new world.” Eleanor Stride suggested combatting complacency by emphasizing the power of small changes in mitigating the impact of climate change. “One billion people making a tiny change has a huge impact,” she said.
The specter of a coronavirus pandemic had not yet become a reality at the time of the symposium. But Edze Westra presciently detailed the challenges of containing a highly contagious emerging pathogen in a “tightly connected world.” He commented that detecting and containing emerging diseases hinges on the development of new diagnostics, and that preventing future outbreaks will require cultural shifts to limit high-risk interactions with wildlife. For zoonotic diseases such as the novel coronavirus, “it’s all about opportunity,” Westra said.
Panelists also looked to the future of science, touching on issues of equality, discrimination, and diversity, and emphasizing the importance of raising the bar for science education. Stride noted that children are natural scientists, gravitating toward problem-solving and puzzles regardless of nationality or gender. “But something happens later,” she said, lamenting the drop in interest in science as children progress in school. “One of the things that gets lost is that creativity, which is what science really is—we’re coming up with a guess and trying to gather evidence for it—we’re not just learning a huge number of facts and regurgitating them,” she said.
In the wake of Brexit, panelists expressed concern about potential difficulties in attracting international students to their labs. “Diversity is so important,” said Penkman. “Getting ideas from all around the world from people with different backgrounds is essential to making science in the UK—and the world—the best it can be.” In her closing comments, Penkman said that ultimately, the trajectory of science comes down to the people in the field. “My eternal optimism is in the people I work with and the people I talk to when I visit schools—it’s that innate interest and curiosity. Whenever I see it, I feel that is the future of science,” she said.
Turning data into predictive models is not a simple task.
Published April 14, 2020
By Roger Torda
Shelf life is an important variable when it comes to snack foods. But how can shelf life be predicted when new products are being developed?
The starting point is often data from taste tests. Turning that data into a predictive model is not a simple task. And that is why PepsiCo, teaming with The New York Academy of Sciences, posed the problem as a challenge to young scientists.
Pallavi Gupta, who is pursuing her PhD in Informatics at the University of Missouri, Columbia, was the Grand Prize winner in the Data Science in Research & Development Challenge. And as a result she will head to Valhalla, New York in the Summer of 2020, for an internship with PepsiCo’s R&D Data Analytics team.
“I love to analyze data,” Pallavi said, quickly breaking into laughter. “I am looking forward to the internship with PepsiCo, to test my skills and to gain additional experience with data analytics using machine learning techniques.”
Competing Against Hundreds of Innovators
Pallavi was among 1,235 registrants in the Challenge. Jhansi Kurma, who recently earned a master’s degree in Business Information Systems from the New Jersey Institute of Technology, came in second.
PepsiCo turned to the Academy to host the competition because of its experience running innovation challenges in science and technology, dating back to 2010. Many of the Academy’s challenges target early career scientists. Other Academy challenges are for high school students.
“The New York Academy of Science-led data challenge has proven to be an excellent way to reach talented data scientists from around the world and have them work on real life challenges together with PepsiCo’s experts. We are looking forward to the 2020 edition and are committed to make this an annual tradition,” says Ellen de Brabander, PepsiCo’s Senior Vice President for Research and Development, said the Data Science Challenge.
The Value of STEM Skills
Large, diverse companies like PepsiCo, value STEM skills across a wide range of job functions.
“In global research and development, our number one output is innovation, and STEM [skills] are critically important competencies to drive innovation,” the company’s James Yuan said in a NYAS webinar titled “Why STEM Professionals are Valuable Across Industries.”
Yuan, Pepsico’s Senior Director, Data Science & Analytics, went on to explain that students joining R&D at the company can pursue work in a wide variety of areas, including product formulation, packaging, process engineering, food safety, quality control, and regulatory affairs.
“In e-commerce and in global business, there are also a lot of opportunities to leverage STEM capabilities for business optimization,” said Eric Higgins, PepsiCo VP, Data Science and Analytics. “We’re talking about media buys, we’re talking about identifying how to best place our products, product assortment, and supply chain optimization.”
A lot of product innovation within this company comes through simply hypothesis testing,” Higgins continued. “Using data science and STEM disciplines, we’re able to automate that process and expand capability, so we can find new ways of innovating. So, in both R&D and on the business side, there are opportunities across the board for people using new methodologies in mathematics, statistics, and computer science.”
Developing a Useful Shelf-Life Model
Competitors in the Challenge were each given a data set from 81 individual shelf-life studies. The data came from evaluations of changes in the taste of snack products as they aged. The goal was to develop a useful shelf-life model that would allow a product developer to predict shelf life based on the product, process, packaging information, and storage conditions related to where the product would be sold.
The competitors had 14 days to complete the challenge. Ten finalists then presented their solutions virtually to a panel of judges, made up of PepsiCo employees from Data Science, R&D, and Human Resources departments.
Pallavi is working toward her PhD, and is using computational and machine learning approaches to study how small non-coding RNA (also known as “small RNAs) – are involved in gene expression regulation. Pallavi said she would take skills from her upcoming internship and apply them to her own research in genomics.
The Data Science in Research and Development Challenge drew entries from 42 countries, especially from the US, Ireland, the UK, Canada and India.
Mammalian cells can make up to 20,000 different proteins, which are responsible for a wide range of cellular functions, including structure, catalysis, transport, and signaling. Proteins are synthesized as linear chains, but to carry out their myriad roles, they must then fold into complex three-dimensional configurations.
Franz-Ulrich Hartl, MD, of the Max Planck Institute of Biochemistry and Arthur Horwich, MD, of Yale School of Medicine and Howard Hughes Medical Institute, have dedicated their careers to better understanding the molecular machinery that drives protein folding, and the implications when a protein misfolds. In doing so, they discovered a new class of proteins, part of the chaperone family, responsible for protein folding.
Chaperones bind to peptide chains as they are being transcribed to prevent them from aggregating and to give them an isolated, quiet space, shielded from the hubbub of the crowded cytoplasm, in which to fold properly. This process is essential to human biology and health, because misfolded proteins are associated with aging and diseases including Alzheimer’s disease, Parkinson’s disease, Huntington’s disease, and prion disease.
On October 4, 2019, prominent scientists gathered at the New York Academy of Sciences to grant the 2019 Dr. Paul Janssen Award to Hartl and Horwich for their groundbreaking insights into chaperone-mediated protein folding. The symposium included award lectures from the honorees, as well as presentations on several aspects of protein folding, from basic biology to the implications for human disease.
Symposium Highlights
While studying mitochondrial protein import, Horwich and Hartl hypothesized that the process may not be spontaneous but dependent on cellular machinery. They discovered a new class of proteins responsible for protein folding.
Hsp60, its bacterial homolog GroEL, and its eukaryotic homolog TRiC have a double ring structure that forms a chamber in which a peptide substrate can fold into its proper shape.
The unfolded protein response of the endoplasmic reticulum responds to the presence of misfolded proteins, which accrue with age. The response itself declines with age.
Hsp70 is a diverse family of monomeric chaperones that binds to polypeptide chains as they’re being translated or when they misfold from mutation or stress and prevents them from collapsing into aggregates.
Clinically relevant receptors that have been difficult to treat require specific chaperones that may provide more easily druggable targets for neurological and psychiatric disorders.
Honorees
Franz-Ulrich Hartl, MD Max Planck Institute of Biochemistry
Arthur Horwich, MD Yale School of Medicine and Howard Hughes Medical Institute
Speakers
David S. Bredt, MD, PhD Janssen Pharmaceutical Companies of Johnson & Johnson
Andrew Dillin, PhD University of California, Berkeley and Howard Hughes Medical Institute
Judith Frydman, PhD Stanford University
Lila M. Gierasch, PhD University of Massachusetts Amherst
Event Sponsors
This symposium was made possible with support from:
Dr. Paul Janssen Award Lectures
Speakers
Franz-Ulrich Hartl Max Planck Institute of Biochemistry
Arthur Horwich Yale School of Medicine and Howard Hughes Medical Institute
Highlights
Chaperones prevent the formation of toxic protein aggregates, and failure of the chaperone system is associated with numerous age-dependent proteopathies and neurodegenerative diseases.
GroEL mediates two key actions on a substrate polypeptide: binding in the open ring forestalls aggregation and can exert unfolding, while binding in the closed ring holds the polypeptide in “solitary confinement,” giving it a chance to fold on its own and alleviating the risk of aggregation.
Molecular Chaperones — Central Players of the Proteostasis Network
“Protein folding is the final step in the information transfer from gene to functional protein, and as such is of fundamental biological importance,” began Franz-Ulrich Hartl.
In the 1950s, biochemist Christian Anfinsen showed that denatured proteins could refold spontaneously in vitro, thus revealing that all of the information required for a protein to attain its final structure is contained in its amino acid sequence. The study was somewhat misleading, however, as it only used small proteins — under 100 amino acids long — and it started with a completely synthesized amino acid chain. This hardly recapitulates the conditions under which proteins must fold in the cell, where many proteins are large, have multiple domains, fold as they are being synthesized on the ribosome, and are in the very crowded cytoplasm.
In the late 1980s, growing evidence showed that cellular machines were required to help proteins fold “at biologically relevant timescales.” These machines were deemed molecular chaperones, as they help proteins achieve their final active conformations but are not themselves part of the final structure. Hartl and Horwich initially discovered chaperones using mitochondria as a model system.
Mitochondria import about 1,000 proteins from the cytoplasm, and these proteins must be unfolded to get across the mitochondrial membranes. Based on Anfinsen’s experiments, it was thought that they would then spontaneously fold properly once inside the mitochondria. But proteins in yeast with mutant Hsp60 got into the mitochondria but failed to fold, identifying Hsp60 as a required chaperone.
Chaperones like Hsp60 prevent the formation of protein aggregates. Aggregation can occur in the intermediate stages of multidomain protein folding when hydrophobic regions might become exposed; chaperones protect these hydrophobic regions through multiple rounds of binding and releasing the partially folded proteins.
ATP binding and hydrolysis often mediate these bind-and-release cycles. The chaperones provide a safe space for the proteins to fold, sequestered away from the hubbub of the cytoplasm. Proteins revisit the quiet chambers that chaperones provide throughout their lifetimes, not only as they are being synthesized.
In the current model, while an amino acid chain is being translated, it interacts with a nascent-chain-binding protein like Hsp70, a type of chaperone that binds to hydrophobic peptide segments. Hsp70 prevents premature misfolding, only allowing the protein to fold when enough structural information for productive folding becomes available — when the protein chain gets long enough.
Most proteins only require this type of chaperone to fold efficiently. But some have more complicated structures and need to fold in the isolated, constrained cage of a cylindrical chaperonin complex like Hsp60, the chaperone that Hartl and Horwich first isolated from mitochondria. Bacterial GroEL and its cofactor GroES are the most well-studied of this class of chaperones; the eukaryotic cytoplasmic versions are called TRiC or CCT.
Chaperones are only one facet of cellular regulation of proteostasis, or protein quality control. They prevent proteins from misfolding, and the degradation machinery eliminates proteins that do not misfold.
There is an age-dependent decline in chaperone function, though. Since chaperones are required for protein maintenance, this decline can lead to a buildup of protein aggregates — which then further strains the already declining chaperones.
These protein aggregates lead to neurodegenerative diseases like Alzheimer’s disease and Huntington’s disease. Aggregates of different disease proteins have the same amyloid fibrillar structure, which suggests that a basic pathological mechanism may underlie all of these diseases. Hartl found that the aggregates interfere with almost every aspect of cellular machinery — transcription, translation, nuclear translocation, DNA maintenance, protein degradation, cytoskeletal organization, and vesicle transport —not only chaperones. But as they overwhelm the chaperone system, toxic aggregates build up until they cause cell death.
Thus, he suggests that rebalancing the proteostasis network may be a means of treating these neurodegenerative diseases.
Chaperonin-mediated Protein Folding
Arthur Horwich described how, in a classic bedside-to-bench approach, he discovered that chaperonin ring machines function to mediate protein folding. He studied the lethal X linked inherited metabolic disease caused by the mutant mitochondrial enzyme OTC. OTC is the second step in the urea cycle; when it is defective, cells can’t clear urea.
Since it is X linked, baby boys with nonfunctional OTC die. Horwich isolated the OTC cDNA and found its mitochondrial transport signal, then looked for a yeast mutant that could transport unfolded human OTC into the mitochondria but in which the transported OTC would not then fold. The yeast mutant he found lacked Hsp60.
Mitochondrial Hsp60, and its bacterial counterpart GroEL, performs two vital functions: they bind to polypeptides to prevent the formation of protein aggregates, and they help polypeptides achieve their functional state. In 1994 and 1997, the X-ray structures of both GroEL alone and in complex with its cochaperonin single ring GroES were presented along with structure-function studies in collaborative work with the late Paul Sigler, providing insight into how the machinery works.
The Binding of GroES to one end of the GroEL cylinder widely expands the folding chamber, giving the substrate space to fold in isolation from the busy cytosolic environment.
GroEL is a cylinder made of 14 identical subunits arranged into two back-to-back 7-membered rings. Each of the subunits is folded into: an equatorial domain, at the waistline of the cylinder, the collective of which hold the assembly together via side-by-side contacts within a ring and contacts of subunits between the two rings; a hinge like “intermediate” domain interconnecting the equatorial and apical domain; and a terminal “apical” domain at an end of the cylinder.
The equatorial domains each house an ATP binding pocket at the inside aspect and the cooperative binding of 7 ATP’s in a GroEL ring causes the terminal GroEL apical domains, attached to the equatorial domains through the slender intermediate domains, to open up like flower petals. In their “unopened” position the apical domains surround an open central cavity of 45 Angstrom diameter and each apical domain proffers sticky “hydrophobic” surface at its cavity-facing aspect.
The continuous hydrophobic surface around the ring specifically captures an unfolded protein species via its own exposed hydrophobic surface (that will become buried to the interior in the final folded “native” form). Thus the binding of a non-native protein by an open GroEL ring serves to capture the protein’s sticky hydrophobic surfaces, masking them, and preventing them from interacting with other unfolded proteins which can lead to aggregation.
When a polypeptide-bound ring of GroEL binds the cochaperonin ring, GroES, a smaller 7-membered single ring of identical subunits, in the presence of ATP, now a large movement of the apical domains occurs, both clockwise rotation and further elevation (see Figure; GroES is colored gold and the GroEL ring undergoing large movements is green). The large movements remove the hydrophobic polypeptide binding surface from facing the cavity, and the lining of the now GroES-encapsulated GroEL cavity becomes watery (hydrophilic) in character.
The large twisting apical domain movements strip the polypeptide off of the cavity wall into the now encapsulated and watery (hydrophilic) cavity where the protein folds in “solitary confinement,” as Horwich phrased it, without any chance of aggregation. Subsequently, after this longest step of the reaction cycle (~10 sec), ATP hydrolyzes, GroES releases, and out from the cavity comes the polypeptide whether properly folded or not. If it has not reached native form, it can make another try at proper folding, either by entering another GroEL cavity, or becoming bound to a different chaperone.
Andrew Dillin University of California, Berkeley and Howard Hughes Medical Institute
Highlights
There are a considerable variety of chaperones that are structurally and functionally different from recognizing and binding nonnative proteins in all of their various stages and processes.
The endoplasmic reticulum unfolded protein response evolved to protect the organism from infection. In the nervous system, it can act in a non-autonomous manner to promote transcription in response to stress.
The TRiCKy Business of Folding Proteins in the Cell
“Proteins are astoundingly complex,” said Judith Frydman. As an example, she pointed to the mammalian respiratory complex I, the 45-subunit complex that drives protons across the inner mitochondrial membrane. Thus, the potential problems with protein folding are not limited to the folding process.
Chaperones bind unfolded polypeptides to help them achieve their native state. Still, much more than that, they engage polypeptides at every stage of their existence in the cell, waiting to receive them as they’re translated and monitoring for damage throughout their lifespans.
TRiC, or CCT, is the stacked chaperone in eukaryotic cells — the equivalent of GroEL. However, unlike GroEL, it does not have a separate cap. It requires ATP hydrolysis, which closes the lid to allow folding; but ATP binding is not sufficient. TRiC binds nascent chains when they are almost complete, while they are still on the ribosome but after they have interacted with Hsp70.
The complex only binds precise types of folding intermediates — notably those with complex topologies like p53, tubulin, actin, telomerase, F box proteins, and others — and then comes off once that folding intermediate has resolved into its properly folded domain. It also suppresses amyloid aggregation, but is overexpressed in many cancers and has been linked to poor prognosis in lung and breast cancer.
Subunit diversity confers unique molecular features to TRiC-mediated folding.
TRiC descends from the chaperone in archaea, which only has one type of subunit. The heteromeric nature of eukaryotic TRiC allows it to form an asymmetrical complex. TRiC has eight subunits, and each subunit has a different affinity for ATP; these subunits are arranged with high-affinity subunits around one side of the ring and low-affinity subunits around the other side.
The subunits have varying degrees of affinity for substrates as well, with each subunit’s binding site presenting a distinct and evolutionarily conserved surface of polar and hydrophobic residues. Their combination thus broadens TRiC’s binding specificity.
Once the binding chamber is closed, one hemisphere is positively charged and the other is negatively charged, further orienting how the substrate can bind and influencing its folding trajectory. Frydman called it a “chaperone with an opinion,” rather than a cage, “that guides the substrate where it needs to go.”
Prefoldin is a cofactor for TRiC, so named because it was thought to facilitate substrate transfer to TRiC before the substrate folded. It binds to TRiC in TRiC’s open state, and, like TRiC, it has a charge asymmetry and a specific pattern of polar and hydrophobic residues that contribute to the inner surface of TRiC’s binding chamber. Prefoldin seems to enhance both the yield and the rate of folding. In vivo, it must bind to TRiC, or else massive protein aggregation builds up in the cell.
Perceiving ER Stress
As many as thirteen million proteins fold and mature in the endoplasmic reticulum (ER) every minute. It is no wonder then that defects in ER function are strongly associated with metabolic and age-related disorders. The unfolded protein response in the ER (UPRER) responds to the presence of unfolded proteins by inducing the transcription of chaperones, and it declines with age. Andrew Dillin wondered how this UPRER works in multicellular organisms.
Are unfolded proteins detected in each individual cell by its own machinery, in a stochastic manner? Or might there be a higher order of regulation, coordinating protein folding mechanisms across the whole system? He turned to C. elegans to figure it out. Since all of the cells in the adult C. elegans are post mitotic, the worm provides a great model system for studying proteome maintenance.
The Dillin lab demonstrated that the neuronal transcription factor XBP-1 could rescue the age-dependent decline in ER proteostasis. Overexpression of XBP-1 extends the worm’s life. XBP-1 — which has the very unusual property that its mRNA is spliced in the cytoplasm instead of the nucleus — senses unfolded proteins and induces the UPRER in nerve cells. These nerves then send signals to peripheral and distal cells, causing them to activate their own UPRER.
Only neuronal cells, both neurons and glia, respond to XBP by inducing the UPR. The peripheral cells don’t sense the unfolded proteins and respond to them; they respond to the signal from the brain. Neurons require small, clear vesicles to send this signal, indicating that neurotransmitters are involved. Unlike neurons, glia need dense core vesicles, suggesting that they signal through neuropeptides or biologic amines rather than neurotransmitters. The neuronal and glial effects are synergistic, and the mechanism is conserved in mice.
XBP-1 induces the UPR from both neurons and glia, but uses different pathways to signal from the different cell types.
The UPRER “only deals with the challenge after the damage has occurred” said Dillin. Wouldn’t a protective system be preferable?
Thus, he conducted a CRISPR screen to find such a system, of UPRER regulators that would identify and protect the organism from ER stress instead of just responding after it happens. In doing so, Dillin found TMEM2, a transmembrane hyaluronidase that had not been previously implicated in ER stress. It does not activate the UPRER, which can induce apoptosis. Rather, it acts through the MAP kinase pathway to promote stress resistance in the ER and survival of the organism.
By breaking down extracellular hyaluronan, it generates a smaller product that increases ER stress resistance. TMEM2 is conserved from worms all the way through humans; it senses the stress from outside the plasma membrane of brain cells, before the stress hits, and then sends the signal to the periphery. Dillin does not yet know how TMEM protects the ER from stress, but he knows that it is not through chaperones.
Franz-Ulrich Hartl Max Planck Institute of Biochemistry
Arthur Horwich Yale School of Medicine and Howard Hughes Medical Institute
Lila M. Gierasch University of Massachusetts Amherst
David S. Bredt Janssen Pharmaceutical Companies of Johnson & Johnson
Seema Kumar (Moderator) Johnson & Johnson
Highlights
The Hsp70 allosteric cycle involves major conformational changes, alternating between a docked state with bound ATP and low affinity for unfolded protein substrates and an undocked state in which the α-helical lid rotates out of the way to allow substrate binding and ATP hydrolysis.
Receptors implicated in neuronal and psychiatric disorders often require specific chaperones to help them fold; these chaperones are often expressed only in specific areas of the brain, and thus may provide appropriate drug targets.
The Versatile Hsp70 Molecular Chaperones Machine
Lila Gierasch introduced Hsp70 as the “early greeting committee” for nascent polypeptide chains. It can maintain the chains in an unfolded state for transport across membranes and meet them on the other side. Hsp70 can also give them a second chance to fold if things don’t go right the first time around. Like all chaperones, it prevents aggregation. It acts as a monomer, but that hardly makes it simple.
Hsp70 activities depend on intramolecular allostery controlled by ligand modulation of an energy landscape. The C-terminal substrate-binding domain (SBD) binds to short hydrophobic stretches of a polypeptide chain. ATP binding to the N-terminal nucleotide-binding domain (NBD) reorients the NBD actin fold. It decreases the affinity of the SBD for the substrate, and the substrate activates the NBD ATPase activity. The α-helical lid can rotate, allowing access to either the SBD or the NBD.
Hsp70 shifts between a docked, ATP bound state with low substrate affinity and an undocked, ADP bound state with high substrate affinity.
Hsp70 allosteric landscapes can be shaped by the strength of interdomain interfaces and as well as ligand binding, making them “tunable molecular machines.” They must have promiscuous selectivity because they bind an immense number of substrates with varying affinities.
There are Hsp70 molecules bound approximately every 40 amino acids throughout the proteome, and there is evidence that more than one Hsp70 molecule can bind to one substrate, mainly to keep it unfolded as it is translocated. And there are many isoforms of eukaryotic Hsp70 with different allosteries. These could have evolved through interactions with co-chaperones, post-translational modifications like phosphorylation, and even the sequence of the substrate.
Gierasch suggested that tweaking its allostery might modulate Hsp70 activity, or one class of Hsp70 could be targeted over another to treat particular diseases. It is tempting to think of activating the chaperone network to prevent neurodegeneration, but it is risky, too, since cancer cells often rely on mutant chaperones.
Getting a Handle on Neuropharmacology by Targeting Receptor Chaperones
Abnormalities in psychiatric diseases are heterogeneous across brain regions, with increased activity in some areas and decreased activity in others. It has been very difficult to find small molecules that can affect synaptic transmission in these different regions.
Stargazer mutant mice, that constantly look up because they have epilepsy, don’t have functional AMPARs (a type of glutamate receptor) on their cerebellar granule cells. David Brendt found that the receptors didn’t work because the mice lacked a chaperone he named stargazin. Stargazin is a Transmembrane AMPAR Regulatory Protein, or TARP, a family of proteins that Bredt said, “act more like escorts than chaperones.”
TARPs take the AMPARs from the endoplasmic reticulum to the cell surface at the synapse of cerebellar granule cells. Different TARPs are distributed to different brain regions, making them attractive drug targets. A molecule that disrupts the interaction between TARP-γ8 and AMPAR has been shown to inhibit neurotransmission in the hippocampus.
Thus, TARPs could be key to treating epilepsy without the terrible side effects of current anticonvulsants, and could possibly be used to treat bipolar disorder, schizophrenia, and anxiety.
Clinically relevant receptors that have been difficult to treat pharmacologically, like AMPAR and nAChRs, have specific required chaperones — TARPS and NACHO, in this case — that may provide more easily druggable targets.
Acetylcholine receptors are the site of action for a number of Alzheimer’s drugs that induce modest but reproducible improvements in cognition. These pentameric receptors have been very difficult to study in the lab, though, because they only fold properly in neuronal cells.
Bredt recognized this as an opportunity in addition to a challenge. His lab cotransfected a library of 4,000 transmembrane proteins along with the acetylcholine receptor into HEK cells and screened for any that would help the receptors fold. Only one did, a novel transmembrane protein with no homology to anything, found in one copy in mammals and Drosophila and not found in worms or yeast at all. They named it NACHO. It resides in the membrane of the endoplasmic reticulum in neuronal cells, and it mediates the folding of nicotinic acetylcholine receptors.
Panel Discussion
Highlights
We don’t know why protein aggregates are toxic, or why chaperones’ ability to prevent their formation wanes with age.
Future research should focus on understanding the proteostasis network in a physiological context and figuring out if, and how, it is an appropriate clinical target.
The day ended with a panel discussion in which Hartl and Horwich fielded questions. Many of them focused on the role misfolded proteins play in disease, why they accumulate with age, and if, when, and how the proteostasis machinery can be targeted therapeutically.
Moderator Seema Kumar began the panel by asking about the greatest challenges and limitations in the field. Horwich replied that we don’t understand the toxicity of misfolded proteins; we don’t even know if they themselves are toxic, or if they are recruiting other toxic mediators. He speculated that it would be great if we could monitor single polypeptide chains as they fold, to see which ones go astray and how that makes them toxic.
Since antibodies against amyloid plaques have been ineffective in Alzheimer’s disease, enhancing multiple parts of the proteostasis network might be a better strategy than targeting specific misfolded proteins or chaperones. Horwich also pointed out that we don’t know why aging thwarts chaperones: does their ability to handle their task decline, or are there genomic or proteomic issues? Hartl added that we don’t understand neurodegenerative diseases nearly well enough to know the role that protein folding plays in their development; Parkinson’s disease, for instance, is likely more than one monolithic disease.
As for how the field will unfold in the future, Horwich noted that most of what we know about protein folding mechanisms comes from in vitro studies with purified components. So we need to know more about how the cellular milieu affects binding affinities and folding. It would be helpful to determine how many times a particular ligand comes back to a particular chaperone. Hartl explained the importance of figuring out who the first responders are, who the next responders are, and if we can develop small molecules to affect the proteostasis machinery.
The New York Academy of Sciences and the Blavatnik Family Foundation hosted the annual Blavatnik Science Symposium on July 15–16, 2019, uniting 75 Finalists, Laureates, and Winners of the Blavatnik Awards for Young Scientists. Honorees from the UK and Israel Awards programs joined Blavatnik National and Regional Awards honorees from the U.S. for what one speaker described as “two days of the impossible.” Nearly 30 presenters delivered research updates over the course of nine themed sessions, offering a fast-paced peek into the latest developments in materials science, quantum optics, sustainable technologies, neuroscience, chemical biology, and biomedicine.
Symposium Highlights
Computer vision and machine learning have enabled novel analyses of satellite and drone images of wildlife, food crops, and the Earth itself.
Next-generation atomic clocks can be used to study interactions between particles in complex many-body systems.
Bacterial communities colonizing the intestinal tract produce bioactive molecules that interact with the human genome and may influence disease susceptibility.
New catalysts can reduce carbon emissions associated with industrial chemical production.
Retinal neurons display a surprising degree of plasticity, changing their coding in response to repetitive stimuli.
New approaches for applying machine learning to complex datasets is improving predictive algorithms in fields ranging from consumer marketing to healthcare.
Breakthroughs in materials science have resulted in materials with remarkable strength and responsiveness.
Single-cell genomic studies are revealing some of the mechanisms that drive cancer development, metastasis, and resistance to treatment.
Speakers
Emily Balskus, PhD Harvard University
Chiara Daraio, PhD Caltech
William Dichtel, PhD Northwestern University
Elza Erkip, PhD New York University
Lucia Gualtieri, PhD Stanford University
Ive Hermans, PhD University of Wisconsin – Madison
Liangbing Hu, PhD University of Maryland, College Park
Jure Leskovec, PhD Stanford University
Heather J. Lynch, PhD Stony Brook University
Wei Min, PhD Columbia University
Seth Murray, PhD Texas A & M University
Nicholas Navin, PhD, MD MD Anderson Cancer Center
Ana Maria Rey, PhD University of Colorado Boulder
Michal Rivlin, PhD Weizmann Institute of Science
Nieng Yan, PhD Princeton University
Event Sponsor
Technology for Sustainability
Speakers
Heather J. Lynch Stony Brook University
Lucia Gualtieri Stanford University
Seth Murray Texas A & M University
Highlights
Machine learning algorithms trained to analyze satellite imagery have led to the discovery of previously unknown colonies of Antarctic penguins.
Seismographic data can be used to analyze more than just earthquakes—typhoons, hurricanes, iceberg-calving events and landslides are reflected in the seismic record.
Unmanned aerial systems are a valuable tool for phenotypic analysis in plant breeding, allowing researchers to take frequent measurements of key metrics during the growing season and identify spectral signatures of crop yield.
Satellites, Drones, and New Insights into Penguin Biogeography
Satellite images have been used for decades to document geological changes and environmental disasters, but ecologist and 2019 Blavatnik National Awards Laureate in Life Sciences, Heather Lynch, is one of the few to probe the database in search of penguin guano. She opened the symposium with the story of how the Landsat satellite program enabled a surprise discovery of several of Earth’s largest colonies of Adélie penguins, a finding that has ushered in a new era of insight into these iconic Antarctic animals.
Steady streams of high quality spatial and temporal data regularly support environmental science. In contrast, Lynch noted that wildlife biology has advanced so slowly that many field techniques “would be familiar to Darwin.” Collecting information on animal populations, including changes in population size or migration patterns, relies on arduous and imprecise counting methods. The quest for alternative ways to track wildlife populations—in this case, Antarctic penguin colonies—led Lynch to develop a machine learning algorithm for automated identification of penguin guano in high resolution commercial satellite imagery, which can be combined with lower resolution imagery like that coming from NASA’s Landsat program. Pairing measurements of vast, visible tracts of penguin guano—the excrement colored bright pink due to the birds’ diet—with information about penguin colony density yields near-precise population information. The technique has been used to survey populations in known penguin colonies and enabled the unexpected discovery of a “major biological hotspot” in the Danger Islands, on the tip of the Antarctic Peninsula. This Antarctic Archipelago is so small that it is doesn’t appear on most maps of the Antarctic continent, yet it hosts one of the world’s largest Adélie penguin hotspots.
Satellite images of the pink stains of Antarctic penguin guano have been used to identify and track penguin populations.
Lynch and her colleagues are developing new algorithms that utilize high-resolution drone and satellite imagery to create centimeter-scale, 3D models of penguin terrain. These models feed into detailed habitat suitability and population-tracking analyses that further basic research and can even influence environmental policy decisions. Lynch noted that the discovery of the Danger Island colony led to the institution of crucial environmental protections for this region that may have otherwise been overlooked. “Better technology actually can lead to better conservation,” she said.
Listening to the Environment with Seismic Waves
The study of earthquakes has dominated seismology for decades, but new analyses of seismic wave activity are broadening the field. “The Earth is never at rest,” said Lucia Gualtieri, 2018 Blavatnik Regional Awards Finalist, while reviewing a series of non-earthquake seismograms that show constant, low-level vibrations within the Earth. Long discarded as “seismic noise,” these data, which comprise more than 90% of seismograms, are now considered a powerful tool for uniting seismology, atmospheric science, and oceanography to produce a holistic picture of the interactions between the solid Earth and other systems.
In addition to earthquakes, events such as hurricanes, typhoons, and landslides are reflected in the seismic record.
Nearly every environmental process generates seismic waves. Hurricanes, typhoons, and landslides have distinct vibrational patterns, as do changes in river flow during monsoons and “glacial earthquakes” caused by ice calving events. Gualtieri illustrated how events on the surface of the Earth are reflected within the seismic record—even at remarkably long distances—including a massive landslide in Alaska detected by a seismic sensor in Massachusetts. Gualtieri and her collaborators are tapping this exquisite sensitivity to create a new generation of tools capable of measuring the precise path and strength of hurricanes and tropical cyclones, and for making predictive models of cyclone strength and behavior based on decades of seismic data.
Improving Crop Yield Using Unmanned Aerial Systems and Field Phenomics
Plant breeders like Seth Murray, 2019 Blavatnik National Awards Finalist, are uniquely attuned to the demands a soaring global population places on the planet’s food supply. Staple crop yields have skyrocketed thanks to a century of advances in breeding and improved management practices, but the pressure is on to create new strategies for boosting yield while reducing agricultural inputs. “We need to grow more plants, measure them better, use more genetic diversity, and create more seasons per year,” Murray said. It’s a tall order, but one that he and a transdisciplinary group of collaborators are tackling with the help of a fleet of unmanned aerial systems (UAS), or drones.
Drones facilitate frequent measurement of plant height, revealing variations between varietals early in the growth process.
Genomics has transformed many aspects of plant breeding, but phenotypic, rather than genotypic, information is more useful for predicting crop yield. Using drones equipped with specialized equipment, Murray has not only automated many of the time-consuming measurements critical for plant phenotyping, such as tracking height, but has also identified novel metrics that can accelerate the development of new varietals. Spectral signatures obtained via drone can be used to identify top-yielding varietals of maize even before the plants are fully mature. Phenotypic features distilled from drone images are also being used to determine attributes such as disease resistance, which directly influence crop management. Murray’s team is modeling the influence of thousands of phenotypes on overall crop performance, paving the way for true phenomic selection in plant breeding.
Quantum mechanics underlies the technologies of modern computing, including transistors and integrated circuits.
Most quantum insights are derived from studies of single quantum particles, but understanding interactions between many particles is necessary for the development of devices such as quantum computers.
Atoms cooled to one billionth of a degree above absolute zero obey the laws of quantum mechanics, and can be used as quantum simulators to study many-particle interactions.
Atomic Clocks: From Timekeepers to Quantum Computers
The discovery of quantum mechanics opened “a new chapter in human knowledge,” said 2019 Blavatnik National Awards Laureate in Physical Sciences & Engineering, Ana Maria Rey, describing how the study of quantum phenomena has revolutionized modern computing, telecommunications, and navigation systems. Transistors, which make up integrated circuits, and lasers, which are the foundation of the atomic clocks that maintain the precision of satellites used in global positioning systems, all stem from discoveries about the nature of quantum particles.
The next generation of innovations—such as room temperature superconductors and quantum computers—will be based on new quantum insights, and all of this hinges on our ability to study interactions between many particles in quantum systems. The complexity of this task is beyond the scope of even the most powerful supercomputers. As Rey explained, calculating the possible states for a small number of quantum particles (six, for example) is simple. “But if you increase that by a factor of just 10, you end up with a number of states larger than the number of stars in the known universe,” she said.
Calculating the number of possible states for even a small number of quantum particles is a task too complex for even the most powerful supercomputer.
Researchers have developed several experimental platforms to clear this hurdle and explore the quantum world. Rey shared the story of how her work developing ultra-precise atomic clocks inadvertently led to one experimental platform that is already demystifying some aspects of quantum systems.
Atomic clocks keep time by measuring oscillations of atoms—typically in cesium atoms—as they change energy levels. Recently, Rey and her collaborators at JILA built the world’s most sensitive atomic clock using strontium atoms instead of cesium and using many more atoms that are typically found in these clocks. The instrument had the potential to be 1,000 times more sensitive than its predecessors, yet collisions between the atoms compromised its precision. Rey explained that by suppressing these collisions, their clock became “a window to explore the quantum world.” Within this framework, the atoms can be manipulated to simulate the movement and interactions of quantum particles in solid-state materials. Rey reported that this clock-turned-quantum simulator has already generated new findings about phenomena including superconductivity and quantum magnetism.
The human gut is colonized by trillions of bacteria that are critical for host health, yet may also be implicated in the development of diseases including colorectal cancer.
For over a decade, chemists have sought to resolve the structure of a genotoxin called colibactin, which is produced by a strain of E. coli commonly found in the gut microbiome of colorectal cancer patients.
By studying the specific type of DNA damage caused by colibactin, researchers found a trail of clues that led to a promising candidate structure of the colibactin molecule.
Gut Reactions: Understanding the Chemistry of the Human Gut Microbiome
The composition of the trillions-strong microbial communities that colonize the mammalian intestinal tract is well characterized, but a deeper understanding of their chemistry remains elusive. Emily Balskus, the 2019 Blavatnik National Awards Laureate in Chemistry, described her lab’s hunt for clues to solve one chemical mystery of the gut microbiome—a mission that could have implications for colorectal cancer (CRC) screening and early detection.
Some commensal E. coli strains in the human gut produce a genotoxin called colibactin. When cultured with human cells, these strains cause cell cycle arrest and DNA damage, and studies have shown increased populations of colibactin-producing E. coli in CRC patients. Previous studies have localized production of colibactin within the E. coli genome and hypothesized that the toxin is synthesized through an enzymatic assembly line. Yet every attempt to isolate colibactin and determine its chemical structure had failed.
Balskus’ group took “a very different approach,” in their efforts to discover colibactin’s structure. By studying the enzymes that make the toxin, the team uncovered a critical clue: a cyclopropane ring in the structure of a series of molecules they believed could be colibactin precursors. This functional group, when present in other molecules, is known to damage DNA, and its detection in the molecular products of the colibactin assembly line led the researchers to consider it as a potential mechanism of colibactin’s genotoxicity.
In collaboration with researchers at the University of Minnesota School of Public Health, Balskus’ team cultured human cells with colibactin-producing E. coli strains as well as strains that cannot produce the toxin. They identified and characterized the products of colibactin-mediated DNA damage. “Starting from the chemical structure of these DNA adducts, we can work backwards and think about potential routes for their production,” Balskus explained.
A proposed structure for the genotoxin colibactin, which is associated with colorectal cancer, features two cyclopropane rings capable of interacting with DNA to generate interstrand cross links, a type of DNA damage.
Further studies revealed that colibactin triggers a specific type of DNA damage that requires two reactive groups—likely represented by two cyclopropane rings in the final toxin structure—a pivotal discovery in deriving what Balskus believes is a strong candidate for the true colibactin structure. Balskus emphasized that this work could illuminate the role of colibactin in carcinogenesis, and may lead to cancer screening methods that rely on detecting DNA damage before cells become malignant. The findings also have implications for understanding microbiome-host interactions. “These studies reveal that human gut microbiota can interact with our genomes, compromising their integrity,” she said.
The chemical industry is a major producer of carbon dioxide, and efforts to create more efficient and sustainable chemical processes are often stymied by cost or scale.
Boron nitride is not well known as a catalyst, yet experiments show it is highly efficient at converting propane to propylene—one of the most widely used chemical building blocks in the world.
Two-dimensional polymers called covalent organic frameworks (COFs) can be used for water filtration, energy storage, and chemical sensing.
Until recently, researchers have struggled to control and direct COF formation, but new approaches to COF synthesis are advancing the field.
Boron Nitride: A Surprising Catalyst
Industrial chemicals “define our standard of living,” said Ive Hermans, 2019 Blavatnik National Awards Finalist, before explaining that nearly 96% of the products used in daily life arise from processes requiring bulk chemical production. These building block molecules are produced at an astonishingly large scale, using energy-intensive methods that also produce waste products, including carbon dioxide.
Despite pressure to reduce carbon emissions, the pace of innovation in chemical production is slow. The industry is capital-intensive — a chemical production plant can cost more than $2 billion—and it can take a decade or more to develop new methods of synthesizing chemicals. Concepts that show promise in the lab often fail at scale or are too costly to make the transition from lab to plant. “The goal is to come up with technologies that are both easily implemented and scalable,” Hermans said.
Catalysts are a key area of interest for improving chemical production processes. These molecules bind to reactants and can boost the speed and efficiency of chemical reactions. Hermans’ research focuses on catalyst design, and one of his recent discoveries, made “just by luck,” stands to transform production of one of the most in-demand chemicals worldwide—propylene.
Historically, propylene was one product (along with ethylene and several others) produced by “cracking” carbon–carbon bonds in naphtha, a crude oil component that has since been replaced by ethane (from natural gas) as a preferred starting material. However, ethane yields far less propylene, leaving manufacturers and researchers to seek alternative methods of producing the chemical.
Boron nitride catalyzes a highly efficient conversion of propane to propylene.
Enter boron nitride, a two-dimensional material whose catalytic properties took Hermans by surprise when a student in his lab discovered its efficiency at converting propane, also a component of natural gas, to propylene. Existing methods for running this reaction are endothermic and produce significant CO2. Boron nitride catalysts facilitate an exothermic reaction that can be conducted at far cooler temperatures, with little CO2 production. Better still, the only significant byproduct is ethylene, an in-demand commodity.
Hermans sees this success as a step toward a more sustainable future, where chemical production moves “away from a linear economy approach, where we make things and produce CO2 as a byproduct, and more toward a circular economy where we use different starting materials and convert CO2 back into chemical building blocks.”
Polymerization in Two Dimensions
William Dichtel, a Blavatnik National Awards Finalist in 2017 and 2019, offered an update from one of the most exciting frontiers in polymer chemistry—two-dimensional polymerization. The synthetic polymers that dominate modern life are comprised of linear, repeating chains of linked building blocks that imbue materials with specific properties. Designing non-linear polymer architectures requires the ability to precisely control the placement of components, a feat that has challenged chemists for a decade.
Dichtel described the potential of a class of polymers called covalent organic frameworks, or COFs—networks of polymers that form when monomers are polymerized into well-defined, two-dimensional structures. COFs can be created in a variety of topologies, dictated by the shape of the monomers that comprise it, and typically feature pores that can be customized to perform a range of functions. These materials hold promise for applications including water purification membranes, energy and gas storage, organic electronics, and chemical sensing.
Dichtel explained that COF development is a trial and error process that often fails, as the mechanisms of their formation are not well understood. “We have very limited ability to improve these materials rationally—we need to be able to control their form so we can integrate them into a wide variety of contexts,” he said.
Two-dimensional polymer networks can be utilized for water purification, energy storage, and many other applications, but chemists have long struggled to understand their formation and control their structure.
A breakthrough in COF synthesis came when chemist Brian Smith, a former postdoc in Dichtel’s lab, discovered that certain solvents allowed COFs to disperse as nanoparticles in solution rather than precipitating as powder. These particles became the basis for a new method of growing large, controlled crystalline COFs using nanoparticles as structural “seeds,” then slowly adding monomers to maximize growth while limiting nucleation. “This level of control parallels living polymerization, with well-defined initiation and growth phases,” Dichtel said.
More recently, Dichtel’s group has made significant advances in COF fabrication, successfully casting them into thin films that could be used in membrane and filtration applications.
Further Readings
Hermans
Zhang Z, Jimenez-Izal E, Hermans I, Alexandrova AN.
The 80 subtypes of retinal ganglion cells each encode different aspects of vision, such as direction and motion.
The “preferences” of these cells were believed to be hard-wired, yet experiments show that retinal ganglion cells can be reprogrammed by exposure to repetitive stimuli.
Sodium ion channels control electrical signaling in cells of the heart, muscles, and brain, and have long been drug targets due to their connection to pain signaling.
Cryo-electron microscopy has allowed researchers to visualize Nav 7, a sodium ion channel implicated in pain syndromes, and to identify molecules that interfere with its function.
Retinal Computations: Recalculating
The presentation from Michal Rivlin, the Life Sciences Laureate of the 2019 Blavatnik Awards in Israel, began with an optical illusion, a dizzying exercise during which a repetitive, unidirectional pattern of motion appeared to rapidly reverse direction. “You probably still perceive motion, but the image is actually stable now,” Rivlin said, completing a powerful demonstration of the action of direction-sensitive retinal ganglion cells (RGCs), whose mechanisms she has studied for more than a decade. The approximately 80 subtypes of RGCs each encode a different aspect, or modality of vision—motion, color, and edges, as well as perception of visual phenomena such as direction. These modalities are hard-wired into the cells and were thought to be immutable—a retinal ganglion cell that perceived left-to-right motion was thought incapable of responding to visual signals that move right-to-left. Rivlin’s research has challenged not only this notion, but also many other beliefs about the function and capabilities of the retina.
Rather than simply capturing discrete aspects of visual information like a camera and relaying that information to the visual thalamus for processing, the cells of the retina actually perform complex processing functions and display a surprising level of plasticity. Rivlin’s lab is probing both the anatomy and functionality of various types of retinal ganglion cells, including those that demonstrate selectivity, such as a preference for movement in one direction or attunement to increases or decreases in illumination. By exposing these cells to various repetitive stimuli, Rivlin has shown that the selectivity of RGCs can be reversed, even in adult retinas.
Direction-selective retinal ganglion cells that prefer left-to-right motion (Before) can change their directional preference (After) following a repetitive visual stimulus.
These dynamic changes in cells whose preferences were believed to be singular and hard-wired have implications not just for understanding retinal function but for understanding the physiological basis of visual perception. Stimulus-dependent changes in the coding of retinal ganglion cells also have downstream impacts on the visual thalamus, where retinal signals are processed. This unexpected plasticity in retinal cells has led Rivlin and her collaborators to investigate the possibility that the visual thalamus and other parts of the visual system might also display greater plasticity than previously believed.
Targeting Sodium Channels for Pain Treatment
Nature’s deadliest predators may seem an unlikely inspiration for developing new analgesic drugs, but as Nieng Yan, 2019 Blavatnik National Awards Finalist, explained, the potent toxins of some snails, spiders, and fish are the basis for research that could lead to safer alternatives to opioid medications.
Voltage-gated ion channels are responsible for electrical signaling in cells of the brain, heart, and skeletal muscles. Sodium channels are one of many ion channel subtypes, and their connection to pain signaling is well documented. Sodium channel blockers have been used as analgesics for a century, but they can be dangerously indiscriminate, inhibiting both the intended channel as well as others in cardiac or muscle tissues. The development of highly selective small molecules capable of blocking only channels tied to pain signaling seemed nearly impossible until two breakthroughs—one genetic, the other technological—brought a potential path for success into focus.
A 2006 study of families with a rare genetic mutation that renders them fully insensitive to pain turned researchers’ focus to the role of the gene SCN9A, which codes for the voltage-gated sodium ion channel Nav 1.7, in pain syndromes. Earlier studies showed that overexpression of SCN9A caused patients to suffer extreme pain sensitivity, and it was now clear that loss of function mutations resulted in the opposite condition.
A powerful natural toxin derived from corn snails blocks the pore of a voltage-gated sodium channel, halting the flow of ions and inhibiting the initiation of an action potential.
As Yan explained, understanding this channel required the ability to resolve its structure, but imaging techniques available at that time were poorly suited to large, membrane-bound proteins. With the advent of cryo-electron microscopy, Yan and other researchers have not only resolved the structure of Nav 1.7, but also characterized small molecules—mostly derived from animal toxins—that precisely and selectively interfere with its function. Developing synthetic drugs based on these molecules is the next phase of discovery, and it’s one that may happen more quickly than expected. “When I started my lab, I thought resolving this protein’s structure would be a lifetime project, but we shortened it to just five years,” said Yan.
A novel approach to developing machine learning algorithms has improved applications for non-linear datasets.
Neural networks can now be used for complex predictive tasks, including forecasting polypharmacy side effects.
5G wireless networks will expand the capabilities of internet-connected devices, providing dramatically faster data transmission and increased reliability.
Tools used to design wireless networks can also be used to understand vulnerabilities in the design of online platforms and social networks, particularly as it pertains to user privacy and data anonymization.
Machine Learning with Networks
“For the first time in history, we are using computers to process data at scale to gain novel insights,” said Jure Leskovec, a Blavatnik National Awards Finalist in 2017, 2018, and 2019, describing one aspect of the digital transformation of science, technology, and society. This shift, from using computers to run calculations or simulations to using them to generate insights, is driven in part by the massive data streams available from the Internet and internet-connected devices. Machine learning has catalyzed this transformation, allowing researchers to not only glean useful information from large datasets, but to make increasingly reliable predictions based on it. Just as new imaging techniques reveal previously unknown structures and phenomena in biology, astronomy, and other fields, so too are big data and machine learning bringing previously unobservable models, signals, and patterns to the surface.
This “new paradigm for discovery” has limitations, as Leskovec explained. Machine learning has advanced most rapidly in areas where data can be represented as simple sequences or grids, such as computer vision, image analysis, and speech processing. Analysis of more complex datasets—represented by networks rather than linear sequences—was beyond the scope of neural networks until recently, when Leskovec and his collaborators approached the challenge from a different angle.
The team considered networks as computation graphs, recognizing that the key to making predictions was understanding how information propagates across the network. By training each node in the network to collect information about neighboring nodes and aggregating the resulting data, they can use node-level information to make predictions within the context of the entire network.
Each node within a network collects information from neighboring nodes. Together, this information can be used to make predictions within the context of the network as a whole.
Leskovec shared two case studies demonstrating the broad applicability of this approach. In healthcare, a neural network designed by Leskovec is identifying previously undocumented side effects from drug-drug interactions. Each network node represents a drug or a protein target of a drug, with links between the nodes emerging based on shared side effects, protein targets, and protein-protein interactions. This type of polydrug side effects analysis is infeasible through clinical trials, and Leskovec is working to optimize it as a point-of-care tool for clinicians.
A similar system has been deployed on the online platform Pinterest, where Leskovec serves as Chief Scientist. It has improved the site’s ability to classify users’ preferences and suggest additional content. “We’re generalizing deep learning methodologies to complex data types, and this is leading to new frontiers,” Leskovec said.
Understanding and Engineering Communications Networks
Elza Erkip has never seen a slide rule. In two decades as a faculty researcher and electrical and computer engineer, Erkip, 2010 Blavatnik Awards Finalist, has corrected her share of misconceptions about her field, and about the role of engineering among the scientific disciplines. She joked about stereotypes portraying engineers—most of them men—wielding slide rules or wearing hard hats, but emphasized the importance of raising awareness about the real-life work of engineers. “Scientists want to understand the universe, but engineers use existing scientific knowledge to design and build things,” she explained. “We contribute to discovery, but mostly we want to solve problems, to find solutions that work in the real world.”
Erkip focuses on one of the most impactful areas of 21st century living—wireless communication—and the ever-evolving suite of technologies that support it. She reviewed the rapid progression of wireless device capabilities, from phones that featured only voice calling and text messaging, through the addition of Wi-Fi capability and web browsing, all the way to the smartphones of today, which boast more computing power than the Apollo 11 spacecraft that landed on the moon. She described the next revolution in wireless—5G networks and devices—which promises higher data rates and significant increases in speed and reliability. Tapping the millimeter-wave bands of the electromagnetic spectrum, 5G will rely on different wireless architectures featuring massive arrays of small antennae, which are better suited to propagating shorter wavelengths. The increased bandwidth will enable many more devices to come online. “It won’t just be humans communicating—we’ll have devices communicating with each other,” Erkip said, describing the future connectivity between robots, autonomous cars, home appliances, and sensors embedded in transportation, manufacturing, and industrial equipment.
Despite efforts to anonymize data, many social media sites and online databases remain vulnerable to efforts to match users’ identities across platforms.
Erkip also discussed the application of tools used to understand and build wireless networks to gain insight into privacy issues within social networks. De-anonymization of user data has long plagued online platforms. Studies have shown that it’s often possible to identify and match users across multiple social platforms or databases using publicly available information—a breach that has greater implications for a database of health or voting records than it does for a consumer-oriented site such as Netflix. Erkip is working to understand the fundamental properties of these networks to elucidate the factors that predispose them to de-anonymization attacks.
IEEE International Symposium on Information Theory. 2018.
Materials Science
Speakers
Chiara Daraio Caltech
Liangbing Hu University of Maryland, College Park
Highlights
Computer-aided manufacturing is enabling researchers to design materials with precisely tuned properties, such as responsiveness to light, temperature, or moisture.
Structured materials can mimic robots or machines, changing shape and form repeatedly in the presence of various stimuli.
Ultra-strong, lightweight wood-based materials made of nanocellulose fibers may one day resolve some of the world’s most pressing challenges in water, energy and sustainability, replacing transparent plastic packaging, window glass, and even steel and other alloys in vehicles and buildings.
Mechanics of Robotic Matter
Chiara Daraio’s work challenges the traditional definition of words like material, structure, and robot. Working at the intersection of physics, materials science, and computer science, she designs materials with novel properties and functionalities, enabled by computer-aided design and 3D fabrication. Rather than considering a material as the foundation for assembling a structure, Daraio, 2019 Blavatnik National Awards Finalist, designs materials with intricate structures in unique and complex geometries.
Daraio demonstrated a series of responsive materials—those that morph in the presence of stimuli such as temperature, light, moisture, or salinity. In their simplest forms, these materials change shape—a piece of heat-responsive material folds and unfolds as air temperature changes, or a leaf-shaped hydro-sensitive material opens and closes as it transitions from wet to dry. In more complex forms, materials can display time-dependent responses, as shown in a video demonstration of a row of polymer strips changing shape at different rates, depending on their thickness. Daraio showed how computer-graphical approaches allow researchers to design a single material with different properties in different regions, allowing complex actuation in a time-dependent manner, such as a polymer “flower” with interconnecting leaves taking shape and a polymer “ribbon” slowly interweaving a knot.
A thin foil elastomer comprised of materials with alternating temperature-sensitivity (heat and cold) folds up and “walks” across a table as the temperature varies.
Conventional ideas dictate that a robot is a programmable machine capable of completing a task. “But what if the material is the machine?” asked Daraio, showing the remarkable capabilities of a thin liquid crystal elastomer foil composed of one heat-sensitive and one cold-sensitive material. At room temperature, the foil is flat. Heat from a warm table causes it to curl upward, turn over, and “walk” forward. “As long as there’s some kind of external environmental stimulus, we can design a material that can repeatedly perform actions in time,” Daraio said. Similar responsive materials have been used in a self-deploying solar panel that [remove folds and] unfolds in response to heat.
Materials have been “the seeds of technological innovation” throughout human history, and Daraio believes that structured materials will enable new functionalities at the macroscale—for use in wearables such as helmets as well as in smart building technologies—and at the microscale, where responsive materials could be used for medical diagnostics or drug delivery.
Sustainable Applications for Wood Nanotechnologies
Wood, glass, plastic, and steel are among the most ubiquitous materials on Earth, and Liangbing Hu, 2019 Blavatnik National Awards Finalist, is rethinking them all. Inspired by the global need to develop sustainable materials, Hu turned to the most plentiful source of biomass on Earth— trees—to create a new generation of wood-based materials with astonishing properties. Hu relies on nanocellulose fibers, which can be engineered to serve as alternatives to commonly used unsustainable or energy-intensive materials.
Hu introduced a transparent film that could pass for plastic and can be used for packaging, yet is ten times stronger and far more versatile. This transparent nanopaper, made of nanocellulose fibers, could also be used as a display material in flexible electronics or as a photonic overlay that boosts the efficiency of solar cells by 30%.
Hu has also tested transparent wood—a heavier-gauge version of nanopaper made by removing lignin from wood and injecting the channels with a clear polymer—as an energy-saving building material. More than half of home energy loss is due to poor wall insulation and leakage through window glass. By Hu’s calculations, replacing glass windows with transparent wood would also provide a six-fold increase in thermal insulation. Pressed, delignified wood has also proven to be a superior material for wall insulation. Used on roofs, it is a highly efficient means of passive cooling—the material absorbs heat and then re-radiates it, cooling the surface below it by about ten degrees.
White delignified wood is pressed to increase its strength. It can be used on roofs to passively cool homes by absorbing and re-radiating light, cooling the area below it by about ten degrees.
Comparisons of mechanical strength between wood and steel are almost laughable, unless the wood is another of Hu’s creations—the aptly named “superwood.” Delignified and compressed to align the nanocellulose fibers, even inexpensive woods become thinner and 10-20 times stronger. Superwood rivals steel in strength and durability, and could become a viable alternative to steel and other alloys in buildings, vehicles, trains, and airplanes. Sustainable sourcing would eliminate pollution and carbon dioxide associated with steel production, and its lightweight profile could drastically improve vehicle fuel efficiency.
Tumor cells are genetically heterogeneous, complicating efforts to sequence DNA from tumor tissue samples.
Techniques for isolating and sequencing single-cell samples have transformed the study of cancer genetics.
Stimulated Ramen scattering, a non-invasive imaging technique, can visualize processes including glucose uptake and fatty acid metabolism within living cells.
Single Cell Genomics: A Revolution in Cancer Biology
Nicholas Navin, 2019 Blavatnik National Awards Finalist, doesn’t use the word “revolution” lightly, but when it comes to the field of single-cell genomics and its impact on cancer research, he stands by the term. Over the past ten years, DNA sequencing of single tumor cells has led to major discoveries about the progression of cancer and the process by which cancer cells resist treatment.
Unlike healthy tissue cells, tumor cells are characterized by genomic heterogeneity. Samples from different areas of the same tumor often contain different mutations or numbers of chromosomes. This diversity has long piqued researchers’ curiosity. “Is it stochastic noise generated as tumor cells acquire different mutations, or could this diversity be important for resistance to therapy, invasion, or metastasis?” Navin asked.
Answering that question required the ability to do comparative studies of single tumor cells, a task that was long out of reach. DNA sequencing technologies historically required a large sample of genetic material—a tricky proposition when sampling a highly diverse population of tumor cells. Some mutations, which could drive invasion or resistance, may be present in just a few cells and thus not be represented in the results. Navin was part of the first team to develop a method for excising a single cancer cell from a tumor, amplifying the DNA, and producing an individualized genetic sequence. As amplification and sequencing methods have improved, so too have the insights gleaned from single-cell genomic studies, which Navin likens to “paleontology in tumors”—the notion that a sample taken at a single point in time can allow researchers to make inferences about tumor evolution.
Single-cell genomic studies reveal that some cancer cells have innate mechanisms of resistance to chemotherapy, and undergo further transcriptional changes that enhance this resistance.
Single-cell studies have contradicted the idea of a stepwise evolution of cancer cells, with one mutation leading to another and ultimately tipping the scales toward malignancy. Instead, Navin’s studies reveal a punctuated evolution, whereby many cells simultaneously become genetically unstable. Longitudinal studies of single-cell samples in patients with triple-negative breast cancer are beginning to answer questions about how cancer cells evade treatment, showing that cells that survive chemotherapy have innate resistance, and then undergo further transcriptional changes during treatment, which increase resistance.
Translating these findings to the clinic is a longer-term process, but Navin envisions single-cell genomics will significantly impact strategies for targeted therapy, non-invasive monitoring, and early cancer detection.
Chemical Imaging in Biomedicine
Wei Min, a Blavatnik Awards Finalist in 2012 and 2019, concluded the session with a visually striking glimpse into the world of stimulated Raman scattering (SRS) microscopy. This noninvasive imaging technique provides both sub-cellular resolution and chemical information about living cells, while transcending some of the limitations of fluorescence-based optical microscopy. The probes used to tag molecules for fluorescent imaging can alter or destroy small molecules of interest, including glucose, lipids, amino acids, or neurotransmitters. Rather than using tags, SRS builds on traditional Raman spectroscopy, which captures and analyzes light scattered by the unique vibrational frequencies between atoms in biomolecules. The original method, first pioneered in the 1930s, is slow and lacks sensitivity, but in 2008, Min and others improved the technique.
SRS has since become a leading method for label-free visualization of living cells, providing an unprecedented window into cellular activities. Using SRS and a variety of custom chemical tags—“vibrational tags,” as Min described them—bound to biomolecules such as DNA or RNA bases, amino acids, or even glucose, researchers can observe the dynamics of biological functions. SRS has visualized glucose uptake in neurons and malignant tumors, and has been used to observe fatty acid metabolism, a critical step in understanding lipid disorders. Imaging small drug molecules is notoriously difficult, but Min reported the results of experiments using SRS to tag therapeutic drug molecules and study their activity within tissues.
Stimulated Raman scattering microscopy uses chemical tags to image small biological molecules in living cells. The technique can visualize cellular processes including glucose uptake in healthy cells and tumor cells.
A recent breakthrough in SRS technology involves pairing it with Raman dyes to break the “color barrier” in optical imaging. Due to the width of the fluorescent spectrum, labels are limited to five or six colors per sample, which prevents researchers from imaging many structures within a tissue sample simultaneously. Min has introduced a hybrid imaging technique that allows for super-multiplexed imaging—up to 10 colors in a single cell image—and utilizes a dramatically expanded palette of Raman frequencies that yield at least 20 distinct colors.
New breakthroughs in controlling mosquito populations, quantum gravity and reducing chemical byproduct waste are among the cutting edge research being honored by the 2019 Blavatnik Regional Awards for Young Scientists.
Published September 14, 2019
By Kamala Murthy
This year the Blavatnik Regional Awards for Young Scientists received 137 nominations from 20 academic institutions in the tri-state area. A jury of distinguished senior scientists and engineers from leading academic institutions selected three outstanding scientists as Winners who will each receive a $30,000 unrestricted prize, and six Finalists (two from each category) who each will collect a $10,000 unrestricted prize.
Supporting outstanding scientists from academic research institutions across New York, New Jersey, and Connecticut since 2007, the Blavatnik Regional Awards for Young Scientists recognize and honor postdoctoral researchers in three scientific disciplinary categories: Life Sciences, Physical Sciences & Engineering, and Chemistry.
The 2019 Blavatnik Regional Awards Winners are:
Life Sciences: Laura Duvall, PhD, nominated by The Rockefeller University (now at Columbia University). Dr. Duvall’s discovery of two key molecules in mosquitos that inhibit blood-feeding and breeding has worldwide implications for controlling mosquito populations and the spread of diseases such as Dengue and Zika. At the time of nomination, Dr. Duvall was a trainee of the 2007 Blavatnik Regional Awards Faculty Winner, Leslie Vosshall of The Rockefeller University.
Physical Sciences & Engineering: Netta Engelhardt, PhD, nominated by Princeton University (now at Massachusetts Institute of Technology). Dr. Engelhardt’s research at the interface of general relativity and quantum field theory is answering complex questions about the fundamentals of our universe, including the remarkable explanation for the origin of black hole entropy. Her work is integral to the understanding of how the fabric of the universe at large-scale is encoded in quantum gravity.
Chemistry: Juntao Ye, PhD, nominated by Cornell University (now at Shanghai Jiao Tong University in China). Improving synthetic efficiency while lowering the cost of synthesis is a primary goal for pharmaceutical industries. Ye invented several new methods that allow for converting readily available chemicals into value-added and pharmaceutically relevant products in a highly efficient and economical manner, while reducing chemical byproduct waste. These methods could accelerate the pace of drug discovery through improving efficiency in synthesizing complex and bioactive compounds.
The cutting-edge discoveries being recognized this year cover an incredibly disparate breadth of work in quantum gravity, drug discovery, control of mosquito populations and underwater photographic imagery. These are the advances that will change our world.
Ellis Rubinstein
2019 Blavatnik Regional Awards Finalists
Life Sciences
Carla Nasca, PhD, nominated by The Rockefeller University — recognized for the discovery of acetyl-L-carnitine (LAC) as a novel modulator of brain rewiring and a possible new treatment for depression that acts by turning on and off specific genes related to the neurotransmitter glutamate.
Liling Wan, PhD, nominated by The Rockefeller University (currently transitioning to the University of Pennsylvania) — recognized for identifying a previously unknown function of a protein called ENL, which has the ability to read epigenetic information on our chromosomes and activate genes that perpetuate tumor growth. Elucidating the structure and mechanism of ENL has guided ongoing development of drugs to treat cancers.
Physical Sciences & Engineering
Derya Akkaynak, PhD, nominated by Princeton University — recognized for significant breakthroughs in computer vision and underwater imaging technologies, resolving a fundamental technological problem in the computer vision community — the reconstruction of lost colors and contrast in underwater photographic imagery — which will have real implications for oceanographic research.
Matthew Yankowitz, PhD, nominated by Columbia University (now at the University of Washington) — recognized for groundbreaking experimental work modifying the electronic properties of a new class of two-dimensional materials, known as van der Waal materials. van der Waal materials have generated tremendous interest due to their properties and the promise they show for use in next-generation optoelectronic and electronic devices, future computing, and telecommunications technologies. Dr. Yankowitz’s work led to the discovery that applied pressure can be used to induce superconductive properties in multi-layer graphene, and has significantly advanced a new area of research recently coined “twistronics.”
Chemistry
Yaping Zang, PhD, nominated by Columbia University — recognized for innovatively using electrochemistry and electrical fields in conjunction with scanning tunneling microscopy techniques to drive chemical reactions. This work provides a deeper understanding of the reaction mechanisms and opens new avenues for the use of electricity as a catalyst in chemical reactions.
Igor Dikiy, PhD, nominated by the Advanced Science Research Center at The Graduate Center, CUNY — recognized for completing the first study of G-protein–coupled receptor (GPCR) fast sidechain dynamics using NMR (nuclear magnetic resonance) spectroscopy to shed light on the molecular mechanisms of cell signaling. GPCRs control a variety of processes in the human body and are targets for over 30% of all FDA-approved drugs. Elucidating the mechanisms of GPCR signaling will enable researchers to design more effective drugs.
Honoring the Blavatnik Regional Award Winners and Finalists
The 2019 Blavatnik Regional Awards Winners and Finalists will be honored at the New York Academy of Sciences’ Annual Gala at Cipriani 25 Broadway in New York on Monday, November 11, 2019.
