Posted on

Delivering genes across the blood-brain barrier to treat brain diseases

As a test of penetrating the blood-brain barrier using a harmless virus (AAV-PHP.B), green fluorescent protein (GFP) was created in the mouse cerebellum via delivery of the GFP gene via the virus. Purkinje neurons (shown in magenta) are responsible for sending signals out of the cerebellum. (credit: Ben Deverman and the Gradinaru laboratory/Caltech)

Caltech biologists have modified a harmless virus to allow it to enter the adult mouse brain through the bloodstream and deliver genes to cells of the nervous system.

The modified virus could lead to novel therapeutics to address diseases such as Alzheimer’s and Huntington’s, help researchers map the brain, and target cells in other organs, according to Ben Deverman, a senior research scientist at Caltech and lead author of a paper describing the work in the February 1 online publication of the journal Nature Biotechnology.

A viral penetration strategy

The blood-brain barrier (BBB) allows the body to keep pathogens and potentially harmful chemicals circulating in the blood from entering the brain and spinal cord. But that also makes it nearly impossible for many drugs and other molecules to be delivered to the brain via the bloodstream.

One variant of the AAV virus (credit: Jazzlw/Creative Commons)

The small, harmless adeno-associated virus (AAV) has been used to transport specific genes into the nuclei of cells; once there, the genes can be expressed (translated from DNA into proteins). AAVs can carry functional copies of genes to replace mutated forms present in individuals with genetic diseases. AAVs can also be used to deliver genes that provide instructions for generating molecules such as antibodies or fluorescent proteins that help researchers study, identify, and track certain cells.

Unfortunately, the blood-brain barrier has made it difficult to deliver AAVs and their genetic cargo to the central nervous system. So researchers have mostly relied on surgical injections, which deliver high concentrations of the virus at the injection site but little to the outlying areas.

That method is also extremely invasive. “One has to drill a hole through skull, then pierce tissue with a needle to the injection site,” explains Viviana Gradinaru, assistant professor of biology and biological engineering at Caltech and senior author on the paper. “The deeper the injection, the higher the risk of hemorrhage.”

Guradinaru says delivering via the BBB would avoid those problems, and “many disorders are not tightly localized. Neurodegenerative disorders like Huntington’s disease affect very large brain areas. Also, many complex behaviors are mediated by distributed interacting networks.”

Testing millions of self-replicating viral variants

In 2009, a group led by Brian Kaspar of Ohio State University published a paper, also in Nature Biotechnology, showing that an AAV strain called AAV9 injected into the bloodstream could make its way into the brain — but it was only efficient when used in neonatal (infant) mice. “The big challenge was how do we achieve the same efficiency in an adult,” says Gradinaru.

To do that efficiently, the researchers developed a high-throughput selection assay called CREATE (Cre REcombinase-based AAV Targeted Evolution) that allowed them to test millions of self-replicating viral variants of AAV in vivo simultaneously and to identify those that were best at entering the brain and delivering genes to a specific class of brain cells known as astrocytes.*

They also used PARS CLARITY, a technique developed previously in the Gradinaru lab to make normally opaque mammalian tissues transparent, allows organs to be examined without the laborious task of making thin slide-mounted sections. This allows researchers to more quickly screen the viral vectors for those that best target the cells and organs of interest.**

They narrowed it down to one variant, which they named AAV-PHP.B — it  delivers genes to the brain and spinal cord efficiently through the vasculature into the brain (and to other tissues in the body).

In collaboration with colleagues from Stanford University, the researchers also showed that AAV-PHP.B is better than AAV9 at delivering genes to human neurons and glia.

Using a new selection method, Caltech researchers have evolved the protein shell of a harmless virus, AAV9, so that it can more efficiently cross the blood brain barrier and deliver genes, such as the green fluorescent protein (GFP), to cells throughout the central nervous system. Here, GFP expression in naturally occurring AAV9 (left) can be seen distributed sparsely throughout the brain. The modified vector, AAV-PHP.B (right), provides more efficient GFP expression. (credit: Ben Deverman and the Gradinaru laboratory/Caltech)

The researchers hope to begin testing AAV-PHP.B’s ability to deliver potentially therapeutic genes in disease models. They are also working to further evolve the virus to make even better performing variants and to produce variants that target certain cell types with more specificity.

The Beckman Institute at Caltech recently opened a resource center called CLOVER (CLARITY, Optogenetics, and Vector Engineering Research Center) to support such research efforts involving tissue clearing and imaging, optogenetic studies, and custom gene-delivery vehicle development. Deverman is the center’s director, and Gradinaru is the principal investigator.

The work was supported by funding from the Hereditary Disease Foundation and the Caltech-City of Hope Biomedical Initiative, a National Institutes of Health (NIH) Director’s New Innovator Award, the NIH’s National Institute of Aging and National Institute of Mental Health, the Beckman Institute, and the Gordon and Betty Moore Foundation.

* They started with the AAV9 virus and modified a gene fragment that codes for a small loop on the surface of the capsid — the protein shell of the virus that envelops all of the virus’ genetic material. Using a common amplification technique, known as polymerase chain reaction (PCR), they created millions of viral variants. Each variant carried within it the genetic instructions to produce more capsids like itself.

Then they used their novel selection process to determine which variants most effectively delivered genes to astrocytes in the brain. Importantly, the new process relies on strategically positioning the gene encoding the capsid variants on the DNA strand between two short sequences of DNA, known as lox sites. These sites are recognized by an enzyme called Cre recombinase, which binds to them and inverts the genetic sequence between them. By injecting the modified viruses into transgenic mice that only express Cre recombinase in astrocytes, the researchers knew that any sequences flagged by the lox site inversion had successfully transferred their genetic cargo to the target cell type—here, astrocytes.

After one week, the researchers isolated DNA from brain and spinal cord tissue, and amplified the flagged sequences, thereby recovering only the variants that had entered astrocytes.

Next, they took those sequences and inserted them back into the modified viral genome to create a new library that could be injected into the same type of transgenic mice. After only two such rounds of injection and amplification, a handful of variants emerged as those that were best at crossing the blood-brain barrier and entering astrocytes.

“We went from millions of viruses to a handful of testable, potentially useful hits that we could go through systematically and see which ones emerged with desirable properties,” says Gradinaru.

Through this selection process, the researchers identified a variant dubbed AAV-PHP.B as a top performer. To test AAV-PHP.B, the researchers used it to deliver a gene that codes for a protein that glows green, making it easy to visualize which cells were expressing it. They injected the AAV-PHP.B or AAV9 (as a control) into different adult mice and after three weeks used the amount of green fluorescence to assess the efficacy with which the viruses entered the brain, the spinal cord, and the retina.

“We could see that AAV-PHP.B was expressed throughout the adult central nervous system with high efficiency in most cell types,” says Gradinaru. Indeed, compared to AAV9, AAV-PHP.B delivers genes to the brain and spinal cord at least 40 times more efficiently.

“What provides most of AAV-PHP.B’s benefit is its increased ability to get through the vasculature into the brain,” says Deverman. “Once there, many AAVs, including AAV9 are quite good at delivering genes to neurons and glia.”

** Making tissues transparent

Gradinaru notes that since AAV-PHP.B is delivered through the bloodstream, it reaches other parts of the body. “Although in this study we were focused on the brain, we were also able to use whole-body tissue clearing to look at its biodistribution throughout the body,” she says.

“In this case, the priority was to express the gene in the brain, but we can see by using whole-body clearing that you can actually have expression in many other organs and even in the peripheral nerves,” explains Gradinaru. “By making tissues transparent and looking through them, we can obtain more information about these viruses and identify targets that we might overlook otherwise.”

The biologists conducted follow-up studies up to a year after the initial injections and found that the protein continued to be expressed efficiently. Such long-term expression is important for gene therapy studies in humans.

Deverman says that the CREATE system could indeed be applied to develop AAVs capable of delivering genes specifically to many different cell types. “There are hundreds of different Cre transgenic lines available,” he says. “Researchers have put Cre recombinase under the control of gene regulatory elements so that it is only made in certain cell types. That means that regardless of whether your objective is to target liver cells or a particular type of neuron, you can almost always find a mouse that has Cre recombinase expressed in those cells.”

“The CREATE system gave us a good hit early on, but we are excited about the future potential of using this approach to generate viruses that have very good cell-type specificity in different organisms, especially the less genetically tractable ones,” says Gradinaru. “This is just the first step. We can take these tools and concepts in many exciting directions to further enhance this work, and we—with the Beckman Institute and collaborators—are ready to pursue those possibilities.”


Ben Deverman and the Gradinaru laboratory/Caltech and Nature Biotechnology | Green fluorescent protein (GFP) expression in the spinal cord of an adult mouse that received an intravenous injection of the modified virus, AAV-PHP.B, carrying the gene for GFP. The spinal cord sample was rendered transparent for imaging using a whole-body tissue clearing method called PARS CLARITY previously developed by the Gradinaru lab.

Abstract of Cre-dependent selection yields AAV variants for widespread gene transfer to the adult brain

Recombinant adeno-associated viruses (rAAVs) are commonly used vehicles for in vivogene transfer. However, the tropism repertoire of naturally occurring AAVs is limited, prompting a search for novel AAV capsids with desired characteristics. Here we describe a capsid selection method, called Cre recombination–based AAV targeted evolution (CREATE), that enables the development of AAV capsids that more efficiently transduce defined Cre-expressing cell populations in vivo. We use CREATE to generate AAV variants that efficiently and widely transduce the adult mouse central nervous system (CNS) after intravenous injection. One variant, AAV-PHP.B, transfers genes throughout the CNS with an efficiency that is at least 40-fold greater than that of the current standard, AAV9, and transduces the majority of astrocytes and neurons across multiple CNS regions. In vitro, it transduces human neurons and astrocytes more efficiently than does AAV9, demonstrating the potential of CREATE to produce customized AAV vectors for biomedical applications.

Posted on

Future of drug delivery seen in a crystal ball

Researchers have developed a way to grow crystals in a spherical shape — a possible future drug-delivery platform. (credit: Drexel University)

A Drexel University materials scientist has discovered a way to encapsulate medication to deliver it more effectively inside the body.

Until now, crystals have grown in rigid, structured formations (like the snowflake) — with a web of straight lines connecting to making a grid that grows into the crystalline flake.*

But the formation of a crystal is affected by the environment in which it forms. And Christopher Li, PhD, a professor in the College of Engineering and head of the Soft Materials Lab in the Department of Materials Science & Engineering, uses this workaround to engineer hollow crystal spheres. He recently reported his finding in Nature Communications (open access).

Introducing …. the “crystalsome”

Li was able to overcome crystal’s edge-forming tendencies by creating a tiny bubble of oil to encase water molecules. When the surfactant bubble was cooled to the appropriate temperature, the molecules inside began to crystalize. But rather than forming an angular web of connections, the molecules, instead, lined up along the interior of the oil bubble — crystallizing in a hollow, spherical shape.

Liposomes, composite structures made of phospholipids, can carry drugs to target tissues (Kosi Gramatikoff/Wikimedia Commons)

Early tests indicate that Li’s “crystalsome” (named for their similarity to liposomes — tiny bubbles with the same membrane as cells that are being explored for use as biological packages for delivering drug treatments) is a few hundred times stronger than liposomes, making them a sturdier option for medicine encapsulation.

Crystalsome fabriation process: 1) Emulsification; (2) quench to the crystallization temperature; and (3–5) different stages of crystal growth (credit: Wenda Wang et al./Nature Communications)

With funding from the National Science Foundation, Li’s team is now exploring ways to control the shape and strength of the spheres by making them out of various different molecules.

* Crystals form this way because their molecules are predisposed to align themselves in a way that links them via the strongest electrochemical bond available. If molecules are floating freely, as they are in a water vapor for example, they are able to follow this default course to connect with other molecules and, eventually, form a crystal — an ice crystal, or snowflake, in the case of water molecules.

Abstract of Highly robust crystalsome via directed polymer crystallization at curved liquid/liquid interface

Lipids and amphiphilic block copolymers spontaneously self-assemble in water to form a plethora of micelles and vesicles. They are typically fluidic in nature and often mechanically weak for applications such as drug delivery and gene therapeutics. Mechanical properties of polymeric materials could be improved by forming crystalline structures. However, most of the self-assembled micelles and vesicles have curved surfaces and precisely tuning crystallization within a nanoscale curved space is challenging, as the curved geometry is incommensurate with crystals having three-dimensional translational symmetry. Herein, we report using a miniemulsion crystallization method to grow nanosized, polymer single-crystal-like capsules. We coin the name crystalsome to describe this unique structure, because they are formed by polymer lamellar crystals and their structure mimics liposomes and polymersomes. Using poly(L-lactic acid) (PLLA) as the model polymer, we show that curved water/p-xylene interface formed by the miniemulsion process can guide the growth of PLLA single crystals. Crystalsomes with the size ranging from ~148 nm to over 1 μm have been formed. Atomic force microscopy measurement demonstrate a two to three orders of magnitude increase in bending modulus compared with conventional polymersomes. We envisage that this novel structure could shed light on investigating spherical crystallography and drug delivery.

Posted on

How to efficiently convert carbon dioxide from air to methanol fuel

The carbon dioxide-to-methanol process (credit: Surya Prakash)

Researchers at the University of Southern California (USC) Loker Hydrocarbon Research Institute have created fuel out of thin air — directly converting carbon dioxide from air into methanol at relatively low temperatures for the first time. While methanol can’t currently compete with oil, it will be there when we run out of oil, the researchers note.

The researchers bubbled air through an aqueous solution of pentaethylenehexamine (PEHA), adding a Ru-Macho-BH ruthenium catalyst to encourage hydrogen to latch onto the CO2 under pressure. They then heated the solution, converting 79 percent of the CO2 into methanol.

Though mixed with water, the resulting methanol can be easily distilled, said G.K. Surya Prakash, professor of chemistry and director of the Loker Hydrocarbon Research Institute.

Proposed reaction sequence for CO2 capture and in situ hydrogenation to methanol (credit: J. Kothandaraman et al./Journal of the American Chemical Society)

Scaling up

Methanol (aka “wood alcohol” or “rubbing alcohol”) is attractive because it can be directly used as a clean-burning liquid fuel for internal combustion engines and for fuel cells. It’s also a hydrogen storage medium and a chemical feedstock for producing a myriad of chemicals and products, including ethylene and propylene. It’s one of the most important building blocks in the chemical industry, with an annual production of more than 70 million tons.

The research is part of a broader effort to use renewable energy to transform greenhouse gas into its combustible form —  attacking global warming from two angles simultaneously.

Prakash and Olah hope to refine the process to the point that it could be scaled up for industrial use within five to 10 years. “Of course it won’t compete with oil today, at around $30 per barrel,” Prakash said. “But right now we burn fossilized sunshine. We will run out of oil and gas, but the sun will be there for another five billion years. So we need to be better at taking advantage of it as a resource.”

Lower temperatures

Previous efforts have required a slower multistage process with the use of high temperatures and high concentrations of CO2, meaning that renewable energy sources would not be able to efficiently power the process.

The new system operates at around 125 to 165 degrees Celsius (257 to 359 degrees Fahrenheit), minimizing the decomposition of the catalyst — which occurs at 155 degrees Celsius (311 degrees Fahrenheit). The system uses a homogeneous catalyst, making it a quicker “one-pot” process. In a lab, the researchers demonstrated that they were able to run the process five times with only minimal loss of the effectiveness of the catalyst.

The new process was published in the Journal of the American Chemical Society on Dec. 29. The research was supported by the USC Loker Hydrocarbon Research Institute.

Abstract of Conversion of CO2 from Air into Methanol Using a Polyamine and a Homogeneous Ruthenium Catalyst

A highly efficient homogeneous catalyst system for the production of CH3OH from CO2 using pentaethylenehexamine and Ru-Macho-BH (1) at 125–165 °C in an ethereal solvent has been developed (initial turnover frequency = 70 h–1 at 145 °C). Ease of separation of CH3OH is demonstrated by simple distillation from the reaction mixture. The robustness of the catalytic system was shown by recycling the catalyst over five runs without significant loss of activity (turnover number > 2000). Various sources of CO2 can be used for this reaction including air, despite its low CO2 concentration (400 ppm). For the first time, we have demonstrated that CO2 captured from air can be directly converted to CH3OH in 79% yield using a homogeneous catalytic system.

Posted on

NASA engineers to build first integrated-photonics modem

NASA laser expert Mike Krainak and his team plan to replace portions of this fiber-optic receiver with an integrated-photonic circuit (its size will be similar to the chip he is holding) and will test the advanced modem on the International Space Station. (credit: W. Hrybyk/NASA)

A NASA team plans to build the first integrated-photonics modem, using an emerging, potentially revolutionary technology that could transform everything from telecommunications, medical imaging, advanced manufacturing to national defense.

The cell phone-sized device incorporates optics-based functions, such as lasers, switches, and fiber-optic wires, onto a microchip similar to an integrated circuit found in all electronics hardware.

The device will be tested aboard the International Space Station beginning in 2020 as part of NASA’s multi-year Laser Communications Relay Demonstration (LCRD). The Integrated LCRD LEO (Low-Earth Orbit) User Modem and Amplifier (ILLUMA) will serve as a low-Earth-orbit terminal for NASA’s LCRD, demonstrating another capability for high-speed, laser-based communications.

Space communications requires 100 times higher data rates

Since its inception in 1958, NASA has relied exclusively on radio frequency (RF)-based communications. Today, with missions demanding higher data rates than ever before, the need for LCRD has become more critical, said Don Cornwell, director of NASA’s Advanced Communication and Navigation Division within the space Communications and Navigation Program, which is funding the modem’s development.

LCRD, expected to begin operations in 2019, promises to transform the way NASA sends and receives data, video and other information. It will use lasers to encode and transmit data at rates 10 to 100 times faster than today’s communications equipment, requiring significantly less mass and power.

Such a leap in technology could deliver video and high-resolution measurements from spacecraft over planets across the solar system — permitting researchers to make detailed studies of conditions on other worlds, much as scientists today track hurricanes and other climate and environmental changes here on Earth.

A payload aboard the Lunar Atmosphere and Dust Environment Explorer (LADEE) demonstrated record-breaking download and upload speeds to and from lunar orbit at 622 megabits per second (Mbps) and 20 Mbps, respectively, in 2013 (see “NASA laser communication system sets record with data transmissions to and from Moon“).

LCRD, however, is designed to be an operational system after an initial two-year demonstration period. It involves a hosted payload and two specially equipped ground stations. The mission will dedicate the first two years to demonstrating a fully operational system, from geosynchronous orbit to ground stations. NASA then plans to use ILLUMA to test communications between geosynchronous and low-Earth-orbit spacecraft, Cornwell said.

Laser Communications Relay Demonstration (LCRD), artist’s illustration (credit: NASA)

Integrated photonics: transforming light-based technologies

ILLUMA incorporates an emerging technology — integrated photonics — that is expected to transform any technology that employs light. This includes everything from Internet communications over fiber optic cable to spectrometers, chemical detectors, and surveillance systems.

“Integrated photonics are like an integrated circuit, except they use light rather than electrons to perform a wide variety of optical functions,” Cornwell said. Recent developments in nanostructures, metamaterials, and silicon technologies have expanded the range of applications for these highly integrated optical chips. Furthermore, they could be lithographically printed in mass — just like electronic circuitry today — further driving down the costs of photonic devices.

“The technology will simplify optical system design, said Mike Krainak, who is leading the modem’s development at NASA’s Goddard Space Flight Center in Greenbelt, Maryland. “It will reduce the size and power consumption of optical devices, and improve reliability, all while enabling new functions from a lower-cost system.”

Krainak also serves as NASA’s representative on the country’s first consortium to advance integrated photonics. Funded by the U.S. Department of Defense, the non-profit American Institute for Manufacturing Integrated Photonics, with headquarters in Rochester, New York, brings together the nation’s leading technological talent to establish global leadership in integrated photonics. Its primary goal is developing low-cost, high-volume, manufacturing methods to merge electronic integrated circuits with integrated photonic devices.

NASA’s Space Technology Mission Directorate (STMD) also appointed Krainak as the integrated photonics lead for its Space Technology Research Grants Program, which supports early-stage innovations. The program recently announced a number of research awards under this technology area (see related story).

Photonics on a chip

Under the NASA project, Krainak and his team will reduce the size of the terminal, now about the size of two toaster ovens — a challenge made easier because all light-related functions will be squeezed onto a microchip. Although the modem is expected to use some optic fiber, ILLUMA is the first step in building and demonstrating an integrated photonics circuit that ultimately will embed these functions onto a chip, he said.

ILLUMA will flight-qualify the technology, as well as demonstrate a key capability for future spacecraft. In addition to communicating to ground stations, future satellites will require the ability to communicate with one another, he said.

“What we want to do is provide a faster exchange of data to the scientific community. Modems have to be inexpensive. They have to be small. We also have to keep their weight down,” Krainak said. The goal is to develop and demonstrate the technology and then make it available to industry and other government agencies, creating an economy of scale that will further drive down costs. “This is the payoff,” he said.

Although integrated photonics promises to revolutionize space-based science and inter-planetary communications, its impact on terrestrial uses also is equally profound, Krainak added. One such use is with data centers. These costly, very large facilities house servers that are connected by fiber optic cable to store, manage and distribute data.

Integrated photonics promises to dramatically reduce the need for and size of these behemoths — particularly since the optic hardware needed to operate these facilities will be printed onto a chip, much like electronic circuitry today. In addition to driving down costs, the technology promises faster computing power.

“Google, Facebook, they’re all starting to look at this technology,” Krainak said. “As integrated photonics progresses to be more cost effective than fiber optics, it will be used,” Krainak said. “Everything is headed this way.”


NASA | Laser Comm: That’s a Bright Idea

Posted on

US could see substantial impact of Zika virus, warns researcher

Countries that have past or current evidence of Zika virus transmission (as of January 2016) (credit: CDC)

A researcher at Boston Medical Center (BMC) and the Boston University School of Public Health (BUSPH) warns that Zika virus could spread quickly to the U.S. There is currently no vaccine or cure.

WHO director general Margaret Chan, M.D., declared on Feb. 1 that the recent cluster of microcephaly and other neurological abnormalities reported in Latin America, following a similar cluster in French Polynesia in 2014, constitutes “a public health emergency of international concern.”

The mosquito-borne virus is believed to cause microcephaly in infants who are exposed in utero and causes rash and flu-like symptoms in adults and children who have been infected.

An outbreak in French Polynesia in 2013 was responsible for 19,000 suspected cases, and since October 2015, nearly 4,000 cases of Zika virus-related microcephaly have been reported in Brazil. Microcephaly is abnormal smallness of the head, a congenital condition associated with incomplete brain development and a range of neurological complications.

The findings are published online (open access) in advance of print in the Annals of Internal Medicine.

Rapidly emerging in the Western Hemisphere via mosquitos or sexual activity

Zika virus has been rapidly emerging in the Western Hemisphere in the last few months, and as of Jan. 22, 2016, there were 20 countries and territories in the Americas with Zika virus in circulation. Currently, it can be found in Central America, the Caribbean and Mexico, and transmission has occurred in travelers to these areas returning to non-endemic countries, including the U.S., Canada, Japan, Western Europe, and Israel.

“At this time, we believe that Zika virus is primarily transmitted via infected mosquitoes, and therefore people living in or traveling to impacted areas are strongly encouraged to protect themselves against mosquitoes by using an effective insect repellent (containing DEET or picaridin),” said senior author Davidson Hamer, MD, director of the Travel Clinic at BMC, and professor of global health and medicine at the Boston University School of Public Health and School of Medicine.

“However, there is some evidence to suggest that Zika virus could be transmitted via blood transfusion and sexual activity, so researchers are trying to determine if these are meaningful pathways to transmission.” There is also evidence of mother-to-child transmission, which appears to be responsible for the surge in cases of microcephaly being seen in Brazil.

Hamer and his co-author, Lin Chen, MD, of the Mt. Auburn Hospital Travel Clinic, say there is substantial risk of introduction of the Zika virus in the U.S., given the presence of the mosquito species that carry the virus, Aedes aegypti and Ae. albopictus, in many states.

“If you are pregnant, put off travel to the endemic areas,” Hamer said. “If you absolutely must go, be sure to protect yourself against mosquitoes. For those who are not pregnant, it’s still a good idea to delay travel so that you don’t risk getting infected and transferring the virus back home — there are many unknowns about its transmission, so there is still a risk.”

In 2007, the first case was detected in a human, leading to an outbreak on an island in Micronesia. An estimated 73 percent of the island residents age 3 or older became infected; however, about 80 percent of these cases did not present significant symptoms.

Zika virus is generally mild and typically resolves itself within a week. Symptoms can include rash, conjunctivitis, muscle and joint pain, headache, joint swelling, dizziness and vomiting. However, neurological and autoimmune complications have been linked to the French Polynesia outbreak, particularly development of Guillain-Barre syndrome, a neurological illness that may result in temporary paralysis. Microcephaly has been reported in thousands of cases in Brazil, and recently in a newborn in Hawaii.

There currently is no vaccine or cure for the Zika virus.

Abstract of Zika Virus: Rapid Spread in the Western Hemisphere

Zika virus, a mosquito-borne flavivirus that causes febrile illness associated with rash, has been rapidly emerging in the Western Hemisphere over the past few months. The virus was rarely identified until outbreaks occurred on Yap Island in the Federated States of Micronesia in 2007, French Polynesia in 2013, and Easter Island in 2014. It was initially detected in Brazil in 2015, in the northeast, and was subsequently identified in other states and several South American countries, including Colombia, Ecuador, Suriname, Venezuela, French Guyana, and Paraguay. Local transmission has been documented in Central America (Panama, El Salvador, Honduras, and Guatemala), the Caribbean (Martinique, Puerto Rico, Dominican Republic, and Haiti), and Mexico. Transmission has also occurred in travelers returning from the infected regions to nonendemic countries, including the United States, Canada, Japan, and Western Europe. As of 22 January 2016, a total of 20 countries and territories in the Americas have Zika virus circulation. The explosive spread mirrors the emergence of chikungunya, which was first detected in the Americas (St. Martin) in 2013 and rapidly disseminated throughout the region.

Posted on

Swarm of aquatic robots learns to cooperate by themselves

Aquatic surface robot (credit: Biomachines Lab)

Portuguese researchers have demonstrated the first swarm of intelligent aquatic surface robots in a real-world environment.

Swarms of aquatic robots have the potential to scale to hundreds or thousands of robots and cover large areas, making them ideal for tasks such as environmental monitoring, search and rescue, and maritime surveillance. They can replace expensive manned vessels and can put the crew out of danger in many maritime missions.

“Swarm robotics is a paradigm shift: we rely on many small, simple and inexpensive robots, instead of a single or a few large, complex and expensive robots,” said Anders Christensen, Ph.D., the principal investigator in the project.

The researchers, from the Institute of Telecommunications at University Institute of Lisbon and from University of Lisbon in Portugal, used nature-inspired approaches for designing their robotic swarm. Instead of manually programming the robots to carry out a mission, evolutionary algorithms are used to synthesize the controller of each robot.

Evolutionary algorithms mimic Darwinian evolution to automatically generate the artificial neural network that allows the robots to carry out the missions autonomously, without a human operator or a central control station. The team demonstrated the capabilities of a swarm with up to ten robots in various collective tasks, including area monitoring, navigation to waypoint, aggregation, and dispersion.

Each robot costs only €300 in materials. The hull of the robots is built from CNC-machined polystyrene foam, and fitted with 3D-printed components. Each robot is equipped with GPS and compass, it can communicate with neighboring robots using Wi-Fi, and the software runs on an onboard $40 Raspberry Pi 2 computer.

The team is now developing the second generation of their aquatic robots, which will be equipped with more advanced sensors and able to carry out long-term missions at sea. The research was funded by Fundação para a Ciência e Tecnologia, Portugal.


AAAI Video Competition | A Sea of Robots

Abstract of Evolution of Collective Behaviors for a Real Swarm of Aquatic Surface Robots

Studies on swarm robotics systems have shown the potential of large-scale multirobot systems based on decentralized control. So far, studies have been mostly conducted in simulation, and the few that have been conducted on real robots have been confined to laboratory environments. In this paper, we present experiments with an autonomous swarm of up to ten aquatic surface robots, conducted in a real and uncontrolled environment. We evolve neural network-based controllers in simulation for canonical swarm robotics tasks, and then assess their performance on real hardware. Our results show that the evolved controllers generally transfer well to the real robots, displaying a similar performance to that obtained in simulation, and similar behavior patterns. We also show that the evolved control maintain key properties of swarm intelligence-based control, namely scalability, flexibility, and robustness. We conclude with a proof-of-concept experiment in which the swarm performs an environmental monitoring task by combining multiple evolved controllers.

Posted on

Scientists discover how the human brain folds

This 3D gel model of a smooth fetal brain is coated with a thin layer of elastomer gel and immersed in solvent. The gel absorbs the solvent, causing it to swell relative to the deeper regions. Within minutes, the resulting compression leads to the formation of folds similar in size and shape to real brains. (credit: Mahadevan Lab/Harvard SEAS)

Folded brains likely evolved to fit a large cortex into a small volume, with the added benefit of reducing neuronal wiring length and improving cognitive function. But how does the brain fold?

A simple mechanical instability associated with buckling, researchers at the Harvard John A. Paulson School of Engineering and Applied Sciences, collaborating with scientists in Finland and France, have discovered in research published in Nature Physics.

Expanded gray matter

“We found that we could mimic cortical folding using a very simple physical principle and get results qualitatively similar to what we see in real fetal brains,” said L. Mahadevan, the Lola England de Valpine Professor of Applied Mathematics, Organismic and Evolutionary Biology, and Physics.

The number, size, shape, and position of neuronal cells present during brain growth all lead to the expansion of the gray matter (cortex) relative to the underlying white matter. This puts the cortex under compression, leading in turn to a mechanical instability that causes it to crease locally.

“This simple evolutionary innovation, with iterations and variations, allows for a large cortex to be packed into a small volume, and is likely the dominant cause behind brain folding, known as gyrification,” said Mahadevan, who is also a core faculty member of the Wyss Institute for Biologically Inspired Engineering and a member of the Kavli Institute for Bionano Science and Technology, both at Harvard University.

Mahadevan’s previous research found that the growth differential between the brain’s outer cortex and the soft tissue underneath explains the variations in the folding patterns across organisms in terms of just two parameters: the relative size of the brain and the relative expansion of the cortex.

Testing a 3D gel model

Left: All notable gyri as identified on a real fetal brain. Right: Analogous regions shown in a simulated brain driven by relative constrained growth. In both cases, the coloring is based on visual identification of the major gyri. (credit: adapted from P. D. Griffiths et al./Atlas of fetal and neonatal brain MR imaging, with permission from Elsevier; and Tuomas Tallinen et al./Nature Physics)

Building on this, the team collaborated with neuroanatomists and radiologists in France and directly tested this theory using data from human fetuses. The team made a 3D gel model of a smooth fetal brain based on MRI images. The model’s surface was coated with a thin layer of elastomer gel, as an analog of the cortex. To mimic cortical expansion, the gel brain was immersed in a solvent that is absorbed by the outer layer, causing it to swell relative to the deeper regions. Within minutes of being immersed in liquid solvent, the resulting compression led to the formation of folds similar in size and shape to real brains.

The extent of the similarities surprised even the researchers. “When I put the model into the solvent, I knew there should be folding but I never expected that kind of close pattern compared to human brain,” said Jun Young Chung, a postdoctoral fellow and co-first author of the paper. “It looks like a real brain.”

The key to those similarities lies in the unique shape of the human brain. “The geometry of the brain is really important because it serves to orient the folds in certain directions,” said Chung. “Our model, which has the same large-scale geometry and curvature as a human brain, leads to the formation of folds that matches those seen in real fetal brains quite well.

“Brains are not exactly the same from one human to another, but we should all have the same major folds in order to be healthy,” said Chung. “Our research shows that if a part of the brain does not grow properly, or if the global geometry is disrupted, we may not have the major folds in the right place, which may cause dysfunction in the brain.”

The distinctive folds of the human brain are not present in most animals — only in a handful of species, including some primates, dolphins, elephants, and pigs. In humans, folding begins in fetal brains around the 20th week of gestation and is completed only when the child is about 18 months old.

Scientists at Jyvaskyla University in Finland, the Institut Mines-Télécom in France, and the University of Aix-Marseille, France were also involved in the research, which was supported by the Academy of Finland, the Wyss Institute for Biologically Inspired Engineering, and fellowships from the MacArthur Foundation and the Radcliffe Institute.

Abstract of On the growth and form of cortical convolutions

Posted on

Graphene is ideal substrate for brain electrodes, researchers find

This illustration portrays neurons interfaced with a sheet of graphene molecules in the background (credit: Graphene Flagship)

An international study headed by the European Graphene Flagship research consortium has found that graphene is a promising material for use in electrodes that interface with neurons, based on its excellent conductivity, flexibility for molding into complex shapes, biocompatibility, and stability within the body.

The graphene-based substrates they studied* promise to overcome problems with “glial scar” tissue formation (caused by electrode-based brain trauma and long-term inflammation). To avoid that, current electrodes based on tungsten or silicon use a protective coating on electrodes, which reduces charge transfer. Current electrodes are also rigid (resulting in tissue detachment and preventing neurons from moving) and generate electrical noise, with partial or complete loss of signal over time, the researchers note in a paper published recently in the journal ACS Nano.

Electrodes are used as neural biosensors and for prosthetic applications — such as deep-brain intracranial electrodes used to control motor disorders (mainly epilepsy or Parkinson’s) and for brain-computer interfaces (BCIs), used to recover sensory functions or control robotic arms for paralyzed patients. These applications require an interface with long-term, minimal interference.

Interfacing graphene to neurons directly

Scanning electron microscope image of rat hippocampal neurons grown in the lab on a graphene-based substrate, showing normal morphology characterized by well-defined round neural soma, extended neurite arborization (branching), and cell density similar to control substrates (credit: A. Fabbro et al./ACS Nano)

“For the first time, we interfaced graphene to neurons directly, without any peptide-coating,” explained lead neuroscientist Prof. Laura Ballerini of the International School for Advanced Studies (SISSA/ISAS) and the University of Trieste.

Using electron microscopy and immunofluorescence, the researchers found that the neurons remained healthy, transmitting normal electric impulses and, importantly, no adverse glial reaction, which leads to damaging scar tissue, was seen.

Atomic force microscope (AFM) image of graphene-based substrate created using liquid phase exfoliation (credit: A. Fabbro et al./ACS Nano)

As a next step, Ballerini says the team plans to investigate how different forms of graphene, from multiple layers to monolayers, are able to affect neurons,  and “whether tuning the graphene material properties might alter the synapses and neuronal excitability in new and unique ways.”

Prof. Andrea C. Ferrari, Director of the Cambridge Graphene Centre and Chair of the Graphene Flagship Executive Board, said the Flagship will “support biomedical research and development based on graphene technology with a new work package and a significant cash investment from 2016.”

The interdisciplinary collaboration also included the University Castilla-La Mancha and the Cambridge Graphene Centre.

* The study used two methods of creating graphene-based substrates (GBSs).  Liquid phase exfoliation (LPE) — peeling off graphene from graphite — can be performed without the potentially hazardous chemical treatments involved in graphene oxide production, is scalable, and operates at room temperature, with high yield. LPE dispersions can also be easily deposited on target substrates by drop-casting, filtration, or printing. Ball milling (BM), with the help of melamine (which forms large hydrogen-bond domains, unlike LPE), can be performed in a solid environment. “Our data indicate that both GBSs are promising for next-generation bioelectronic systems, to be used as brain interfaces,” the paper concludes.

Posted on

Scientists decode brain signals to recognize images in real time

Using electrodes implanted in the temporal lobes of seven awake epilepsy patients, University of Washington scientists have decoded brain signals (representing images) at nearly the speed of perception for the first time* — enabling the scientists to predict in real time which images of faces and houses the patients were viewing and when, and with better than 95 percent accuracy.

Multi-electrode placements on thalamus surface (credit: K.J. Miller et al./PLoS Comput Biol)

The research, published Jan. 28 in open-access PLOS Computational Biology, may lead to an effective way to help locked-in patients (who were paralyzed or have had a stroke) communicate, the scientists suggest.

Predicting what someone is seeing in real time

“We were trying to understand, first, how the human brain perceives objects in the temporal lobe, and second, how one could use a computer to extract and predict what someone is seeing in real time,” explained University of Washington computational neuroscientist Rajesh Rao. He is a UW professor of computer science and engineering and directs the National Science Foundation’s Center for Sensorimotor Engineering, headquartered at UW.

The study involved patients receiving care at Harborview Medical Center in Seattle. Each had been experiencing epileptic seizures not relieved by medication, so each had undergone surgery in which their brains’ temporal lobes were implanted (temporarily, for about a week) with electrodes to try to locate the seizures’ focal points.

Temporal lobes process sensory input and are a common site of epileptic seizures. Situated behind mammals’ eyes and ears, the lobes are also involved in Alzheimer’s and dementias and appear somewhat more vulnerable than other brain structures to head traumas, said UW Medicine neurosurgeon Jeff Ojemann.

Recording digital signatures of images in real time

In the experiment, signals from electrocorticographic (ECoG) electrodes from multiple temporal-lobe locations were processed powerful computational software that extracted two characteristic properties of the brain signals: “event-related potentials” (voltages from hundreds of thousands of neurons activated by an image) and “broadband spectral changes” (processing of power measurements  across a wide range of frequencies).

Averaged broadband power at two multi-electrode locations (1 and 4) following presentation of different images; note that responses to people are stronger than to houses. (credit: K.J. Miller et al./PLoS Comput Biol)

Target image (credit: K.J. Miller et al./PLoS Comput Biol)

The subjects, watching a computer monitor, were shown a random sequence of pictures: brief (400 millisecond) flashes of images of human faces and houses, interspersed with blank gray screens. Their task was to watch for an image of an upside-down house and verbally report this target, which appeared once during each of 3 runs (3 of 300 stimuli). Patients identified the target with less than 3 percent errors across all 21 experimental runs.

The computational software sampled and digitized the brain signals 1,000 times per second to extract their characteristics. The software also analyzed the data to determine which combination of electrode locations and signal types correlated best with what each subject actually saw.

By training an algorithm on the subjects’ responses to the (known) first two-thirds of the images, the researchers could examine the brain signals representing the final third of the images, whose labels were unknown to them, and predict with 96 percent accuracy whether and when (within 20 milliseconds) the subjects were seeing a house, a face or a gray screen, with only ~20 milliseconds timing error.

This accuracy was attained only when event-related potentials and broadband changes were combined for prediction, which suggests they carry complementary information.

Steppingstone to real-time brain mapping

“Traditionally scientists have looked at single neurons,” Rao said. “Our study gives a more global picture, at the level of very large networks of neurons, of how a person who is awake and paying attention perceives a complex visual object.”

The scientists’ technique, he said, is a steppingstone for brain mapping, in that it could be used to identify in real time which locations of the brain are sensitive to particular types of information.

“The computational tools that we developed can be applied to studies of motor function, studies of epilepsy, studies of memory. The math behind it, as applied to the biological, is fundamental to learning,” Ojemann added.

Lead author of the study is Kai Miller, a neurosurgery resident and physicist at Stanford University who obtained his M.D. and Ph.D. at the UW. Other collaborators were Dora Hermes, a Stanford postdoctoral fellow in neuroscience, and Gerwin Schalk, a neuroscientist at the Wadsworth Institute in New York.

This work was supported by National Aeronautics and Space Administration Graduate Student Research Program, the National Institutes of Health, the National Science Foundation, and the U.S. Army.

* In previous studies, such as these three covered on KurzweilAI, brain images were reconstructed after they were viewed, not in real time: Study matches brain scans with topics of thoughts, Neuroscape Lab visualizes live brain functions using dramatic images, How to make movies of what the brain sees.

Abstract of Spontaneous Decoding of the Timing and Content of Human Object Perception from Cortical Surface Recordings Reveals Complementary Information in the Event-Related Potential and Broadband Spectral Change

The link between object perception and neural activity in visual cortical areas is a problem of fundamental importance in neuroscience. Here we show that electrical potentials from the ventral temporal cortical surface in humans contain sufficient information for spontaneous and near-instantaneous identification of a subject’s perceptual state. Electrocorticographic (ECoG) arrays were placed on the subtemporal cortical surface of seven epilepsy patients. Grayscale images of faces and houses were displayed rapidly in random sequence. We developed a template projection approach to decode the continuous ECoG data stream spontaneously, predicting the occurrence, timing and type of visual stimulus. In this setting, we evaluated the independent and joint use of two well-studied features of brain signals, broadband changes in the frequency power spectrum of the potential and deflections in the raw potential trace (event-related potential; ERP). Our ability to predict both the timing of stimulus onset and the type of image was best when we used a combination of both the broadband response and ERP, suggesting that they capture different and complementary aspects of the subject’s perceptual state. Specifically, we were able to predict the timing and type of 96% of all stimuli, with less than 5% false positive rate and a ~20ms error in timing.

Posted on

Machine-learning technique uncovers unknown features of multi-drug-resistant pathogen

According to the CDC, Pseudomonas aeruginosa is a common cause of healthcare-associated infections, including pneumonia, bloodstream infections, urinary tract infections, and surgical site infections. Some strains of P. aeruginosa have been found to be resistant to nearly all or all antibiotics. (illustration credit: CDC)

A new machine-learning technique can uncover previously unknown features of organisms and their genes in large datasets, according to researchers from the Perelman School of Medicine at the University of Pennsylvania and the Geisel School of Medicine at Dartmouth University.

For example, the technique learned to identify the characteristic gene-expression patterns that appear when a bacterium is exposed in different conditions, such as low oxygen and the presence of antibiotics.

The technique, called “ADAGE” (Analysis using Denoising Autoencoders of Gene Expression), uses a “denoising autoencoder” algorithm, which learns to identify recurring features or patterns in large datasets — without being told what specific features to look for (that is, “unsupervised.”)*

Last year,  Casey Greene, PhD, an assistant professor of Systems Pharmacology and Translational Therapeutics at Penn, and his team published, in an open-access paper in the American Society for Microbiology’s mSystems, the first demonstration of ADAGE in a biological context: an analysis of two gene-expression datasets of breast cancers.

Tracking down gene patterns of a multi-drug-resistant bacterium

The new study, published Jan. 19 in an open-access paper in mSystems, was more ambitious. It applied ADAGE to a dataset of 950 gene-expression arrays publicly available at the time for the multi-drug-resistant bacterium Pseudomonas aeruginosa. This bacterium is a notorious pathogen in the hospital and in individuals with cystic fibrosis and other chronic lung conditions; it’s often difficult to treat due to its high resistance to standard antibiotic therapies.

The data included only the identities of the roughly 5,000 P. aeruginosa genes and their measured expression levels in each published experiment. The goal was to see if this “unsupervised” learning system could uncover important patterns in P. aeruginosa gene expression and clarify how those patterns change when the bacterium’s environment changes — for example, when in the presence of an antibiotic.

Even though the model built with ADAGE was relatively simple — roughly equivalent to a brain with only a few dozen neurons — it had no trouble learning which sets of P. aeruginosa genes tend to work together or in opposition. To the researchers’ surprise, the ADAGE system also detected differences between the main laboratory strain of P. aeruginosa and strains isolated from infected patients. “That turned out to be one of the strongest features of the data,” Greene said.

“We expect that this approach will be particularly useful to microbiologists researching bacterial species that lack a decades-long history of study in the lab,” said Greene. “Microbiologists can use these models to identify where the data agree with their own knowledge and where the data seem to be pointing in a different direction … and to find completely new things in biology that we didn’t even know to look for.”

Support for the research came from the Gordon and Betty Moore Foundation, the William H. Neukom Institute for Computational Science, the National Institutes of Health, and the Cystic Fibrosis Foundation.

* In 2012, Google-sponsored researchers applied a similar method to randomly selected YouTube images; their system learned to recognize major recurring features of those images — including cats of course.

Abstract of ADAGE-Based Integration of Publicly Available Pseudomonas aeruginosa Gene Expression Data with Denoising Autoencoders Illuminates Microbe-Host Interactions

The increasing number of genome-wide assays of gene expression available from public databases presents opportunities for computational methods that facilitate hypothesis generation and biological interpretation of these data. We present an unsupervised machine learning approach, ADAGE (analysis using denoising autoencoders of gene expression), and apply it to the publicly available gene expression data compendium for Pseudomonas aeruginosa. In this approach, the machine-learned ADAGE model contained 50 nodes which we predicted would correspond to gene expression patterns across the gene expression compendium. While no biological knowledge was used during model construction, cooperonic genes had similar weights across nodes, and genes with similar weights across nodes were significantly more likely to share KEGG pathways. By analyzing newly generated and previously published microarray and transcriptome sequencing data, the ADAGE model identified differences between strains, modeled the cellular response to low oxygen, and predicted the involvement of biological processes based on low-level gene expression differences. ADAGE compared favorably with traditional principal component analysis and independent component analysis approaches in its ability to extract validated patterns, and based on our analyses, we propose that these approaches differ in the types of patterns they preferentially identify. We provide the ADAGE model with analysis of all publicly available P. aeruginosa GeneChip experiments and open source code for use with other species and settings. Extraction of consistent patterns across large-scale collections of genomic data using methods like ADAGE provides the opportunity to identify general principles and biologically important patterns in microbial biology. This approach will be particularly useful in less-well-studied microbial species.

Abstract of Unsupervised feature construction and knowledge extraction from genome-wide assays of breast cancer with denoising autoencoders

Big data bring new opportunities for methods that efficiently summarize and automatically extract knowledge from such compendia. While both supervised learning algorithms and unsupervised clustering algorithms have been successfully applied to biological data, they are either dependent on known biology or limited to discerning the most significant signals in the data. Here we present denoising autoencoders (DAs), which employ a data-defined learning objective independent of known biology, as a method to identify and extract complex patterns from genomic data. We evaluate the performance of DAs by applying them to a large collection of breast cancer gene expression data. Results show that DAs successfully construct features that contain both clinical and molecular information. There are features that represent tumor or normal samples, estrogen receptor (ER) status, and molecular subtypes. Features constructed by the autoencoder generalize to an independent dataset collected using a distinct experimental platform. By integrating data from ENCODE for feature interpretation, we discover a feature representing ER status through association with key transcription factors in breast cancer. We also identify a feature highly predictive of patient survival and it is enriched by FOXM1 signaling pathway. The features constructed by DAs are often bimodally distributed with one peak near zero and another near one, which facilitates discretization. In summary, we demonstrate that DAs effectively extract key biological principles from gene expression data and summarize them into constructed features with convenient properties.