Biomedicine

Nano light: Nanoparticles that emit bright infrared light and target melanoma tumors are seen here in mice.
Credit: Ulrich Wiesner

Researchers are testing the safety of a nanoparticle that targets cancer cells

MIT Technology Review, February 22, 2011, by Katherine Bourzac  —   A nanoparticle that targets melanoma and highlights cancerous tissue is entering an early-stage clinical trial. Researchers testing the nanotherapeutic agent, which has been under development for over a decade, hope it provides a way to target melanoma and map its spread throughout the body. Researchers have tested the drug in animals and found no toxicity. Safety tests in five melanoma patients should be completed by the end of the year.

Drugs that help doctors image, characterize, and treat diseases could result in treatments that are better targeted to an individual patient’s disease. “With cancer genome programs, we’re learning more and more about differences between individuals’ diseases,” says Jerry Lee, director of the Office of Physical Sciences-Oncology at the National Cancer Institute. That information will tell doctors what drugs will work best for a patient, and how they might best be delivered. “Multifunctional, tailored nanoplatforms will bridge with that biological information,” enabling doctors to act on it to improve patient care, Lee says.

The new melanoma-targeting nanoparticles were developed by Ulrich Wiesner, professor of materials science at Cornell. He’s worked with a group led by Michelle Bradbury, a radiologist at the Memorial Sloan-Kettering Institute in New York City, to test the nanoparticles in animals. Bradbury is also leading the clinical trial.

The researchers hope to use the nanoparticles to address two major clinical needs. First, they want to use it to develop a therapy that seeks out melanoma tumors. “There’s never been a targeted therapy for melanoma,” says Bradbury. Melanoma starts on the skin, but when it spreads to other parts of the body, it is invisible and deadly. A targeted therapy would seek melanoma out no matter where it has spread.

“Another gap in the field is the lack of an optical imaging agent to visualize lymph nodes,” says Bradbury. Today, surgeons use radioactive labels and a handheld gamma detector to find cancer-carrying lymph nodes in the head and neck during surgery. But this is a tricky process. Bradbury hopes the nano-imaging agent can be used to light up cancer-carrying lymph nodes during surgery, providing a map that helps doctors remove the cancer while avoiding unnecessary cutting that can lead to painful side effects.

The core of the nanoparticles is a silica sphere, about eight nanometers in diameter, surrounding an organic dye molecule that emits infrared light. This is then coated with a biocompatible polymer that helps the nanoparticles stick around in the body. Wiesner and a former student first developed the nanoparticles over 10 years ago. The nanoparticles are made by a company called Hybrid Silica Technologies. The coated nanoparticles can be modified to serve many different purposes. “Through simple biochemistry, you can attach peptides to target tumors, drugs, and radioactive imaging labels,” says Wiesner.

For the initial patient trial, Wiesner and Hybrid Silica Technologies provided the clinical researchers with the nanoparticles. The nanoparticles were treated with radioactive iodine in order to make them visible on PET scans. The advantage of PET scans is their incredible sensitivity, says Bradbury. If an MRI label were added to the particle and that imaging technique were used instead, a much higher dose would be necessary. “PET enables you to do microdosing,” she says. PET scans help provide a very detailed map of where the nanoparticles travel inside the body.

Bradbury hopes that oncologists will eventually use this type of imaging to better understand a patient’s disease. PET imaging is sensitive enough to allow researchers to estimate how many of different types of receptors are present on an individual tumor’s cells, information that should help doctors determine how aggressive a tumor is, where it might spread and when, and how it should be treated.

However, this type of agent must strike the right balance between remaining in the body long enough to do its work but not overstaying its welcome. “It remains in the blood for enough time to target the tumor, yet clears through the kidneys efficiently,” says Bradbury. Drugs that move through the liver stay in the body longer and can get broken down into potentially toxic side-products. In mice, the silica particles are excreted in about 24 hours. Ten years of tests in animals have shown no toxicity.

“If we can get these into the clinic, this is a platform that could really expand what we can do for patients,” says Bradbury.

University of Groningen, ScienceDaily.com  —  The Groningen professors Bauke Dijkstra and Lubbert Dijkhuizen have deciphered the structure and functional mechanism of the glucansucrase enzyme that is responsible for dental plaque sticking to teeth. This knowledge will stimulate the identification of substances that inhibit the enzyme. Just add that substance to toothpaste, or even sweets, and caries will be a thing of the past.

The results of the research have been published this week in the journal Proceedings of the National Academy of Sciences (PNAS).

The University of Groningen researchers analysed glucansucrase from the lactic acid bacterium Lactobacillus reuteri, which is present in the human mouth and digestive tract. The bacteria use the glucansucrase enzyme to convert sugar from food into long, sticky sugar chains. They use this glue to attach themselves to tooth enamel. The main cause of tooth decay, the bacterium Streptococcus mutans, also uses this enzyme. Once attached to tooth enamel, these bacteria ferment sugars releasing acids that dissolve the calcium in teeth. This is how caries develops.

Three dimensional structure

Using protein crystallography, the researchers were able to elucidate the three dimensional (3D) structure of the enzyme. The Groningen researchers are the first to succeed in crystallizing glucansucrase. The crystal structure has revealed that the folding mechanism of the protein is unique. The various domains of the enzyme are not formed from a single, linear amino acid chain but from two parts that assemble via a U-shaped structure of the chain; this is the first report on such a folding mechanism in the literature.

Functional mechanism

The unravelling of the 3D structure provided the researchers with detailed insight into the functional mechanism of the enzyme. The enzyme splits sucrose into fructose and glucose and then adds the glucose molecule to a growing sugar chain. Thus far the scientific community assumed that both processes were performed by different parts of the enzyme. However, the model created by the Groningen researchers has revealed that both activities occur in the same active site of the enzyme.

Inhibitors

Dijkhuizen expects that specific inhibitors for the glucansucrase enzyme may help to prevent attachment of the bacteria to the tooth enamel. Information about the structure and functional mechanism of the enzyme is crucial for developing such inhibitors. Thus far, such research has not been successful, states Dijkhuizen: ‘The various inhibitors studied not only blocked the glucansucrase, but also the digestive enzyme amylase in our saliva, which is needed to degrade starch.’

Evolution

The crystal structure also provides an explanation for this double inhibition. The data published by the Groningen scientists shows that glucansucrase proteins most likely evolved from amylase enzymes that degrade starch. ‘We already knew that the two enzymes were similar’, says Dijkhuizen, ‘but the crystal structure revealed that the active sites are virtually identical. Future inhibitors thus need to be directed towards very specific targets because both enzymes are evolutionary closely related.’

Toothpaste and sweets

Dijkhuizen points out that in future glucansucrase inhibitors may be added to toothpaste and mouthwash. ‘But it may even be possible to add them to sweets’, he suggests. ‘An inhibitor might prevent that sugars released in the mouth cause damage.’ However, Dijkhuizen doesn’t expect that toothbrushes have had their day: ‘it will always be necessary to clean your teeth.’

Credit: Technology Review

Providing developers with machine learning on tap could unleash a flood of smarter apps

MIT Technology Review, by Tom Simonite  —  From Amazon’s product recommendations to Pandora’s ability to find us new songs we like, the smartest Web services around rely on machine learning–algorithms that enable software to learn how to respond with a degree of intelligence to new information or events.

Now Google has launched a service that could bring such smarts to many more apps. Google Prediction API provides a simple way for developers to create software that learns how to handle incoming data. For example, the Google-hosted algorithms could be trained to sort e-mails into categories for “complaints” and “praise” using a dataset that provides many examples of both kinds. Future e-mails could then be screened by software using that API, and handled accordingly.

Currently just “hundreds” of developers have access to the service, says Travis Green, Google’s product manager for Prediction API, “but already we can see people doing some amazing things.” Users range from developers of mobile and Web apps to oil companies, he says. “Many want to do product recommendation, and there are also interesting NGO use cases with ideas such as extracting emergency information from Twitter or other sources online.”

Machine learning is not an easy feature to build into software. Different algorithms and mathematical techniques work best for different kinds of data. Specialized knowledge of machine learning is typically needed to consider using it in a product, says Green.

Google’s service provides a kind of machine-learning black box–data goes in one end, and predictions come out the other. There are three basic commands: one to upload a collection of data, another telling the service to learn what it can from it, and a third to submit new data for the system to react to based on what it learned.

“Developers can deploy it on their site or app within 20 minutes,” says Green. “We’re trying to provide a really easy service that doesn’t require them to spend month after month trying different algorithms.” Google’s black box actually contains a whole suite of different algorithms. When data is uploaded, all of the algorithms are automatically applied to find out which works best for a particular job, and the best algorithm is then used to handle any new information submitted.

“Getting machine learning to a Google scale is significant,” says Joel Confino, a software developer in Philadelphia who builds large-scale Web apps for banks and pharmaceutical companies, and a member of the preview program. He used Prediction API to quickly develop a simple yet effective spam e-mail filter, and he says the service has clear commercial potential.

For example, a bank or credit-card company wanting to use machine learning to build systems that make decisions based on historical transactions is unlikely to have the specialized staff and necessary infrastructure for what is a computationally intensive approach. “This API could be a way to get a capability cheaply that would cost a huge amount through a traditional route.”

Google’s new service may also be more palatable to businesses wary of handing over their data to cloud providers, says Confino. “The data can be completely obfuscated, and you can still use this service. Google doesn’t have to know if those numbers you are sending it are stock prices or housing prices.”

Google does, however, get some information that it can use to improve its machine-learning algorithms. “We don’t look at users’ data, but we do see the same metrics on prediction quality that they do, to help us improve the service,” says Green. The engineers running Prediction API will know if a particular algorithm is rarely used, or if a new one needs to be added to the mix to better process certain types of data.

Prediction API has the potential to be a leveler between established companies and smaller startups, says Pete Warden, an ex-Apple engineer now working on his own startup OpenHeatMap.com. “That’s been a competitive advantage for large companies like Amazon, whose product recommendation is built on machine learning,” he explains. “Now you still have to have a decent set of training data, but you don’t have to have the same level of expertise.”

Warden has yet to gain access to Prediction API, but has plans to use it to improve a service he built that shows where people using a particular word or phrase on Twitter are located. “It would be really interesting to also see where they are saying positive and negative things on a subject,” says Warden. Prediction API could be trained to distinguish between positive and negative tweets to do that, he says.

Chris Bates, a data scientist with online music service Grooveshark and a member of the preview program, agrees that Google’s black box will enable wider use of machine learning, but he contends that the service needs to mature. “Today it is good at predicting which language text is in and also sentiment analysis, for example to pick out positive and negative reviews,” he says.

Ultimately, though, being unable to inspect the inner workings of the algorithms and fine-tune them for a specific use may have its limits. “It’s good for cases that are not mission-critical, where you can afford a few false positives,” Bates says. For example, a spam filter that occasionally lets through the occasional junk message could still be usable, but a credit-card company might be less able to accept any errors.