Machine-learning system determines the fewest, smallest doses that could still shrink brain tumors

MIT researchers are employing novel machine-learning techniques to improve the quality of life for patients by reducing toxic chemotherapy and radiotherapy dosing for glioblastoma, the most aggressive form of brain cancer.

Glioblastoma is a malignant tumor that appears in the brain or spinal cord, and prognosis for adults is no more than five years. Patients must endure a combination of radiation therapy and multiple drugs taken every month. Medical professionals generally administer maximum safe drug doses to shrink the tumor as much as possible. But these strong pharmaceuticals still cause debilitating side effects in patients.

In a paper being presented next week at the 2018 Machine Learning for Healthcare conference at Stanford University, MIT Media Lab researchers detail a model that could make dosing regimens less toxic but still effective. Powered by a “self-learning” machine-learning technique, the model looks at treatment regimens currently in use, and iteratively adjusts the doses. Eventually, it finds an optimal treatment plan, with the lowest possible potency and frequency of doses that should still reduce tumor sizes to a degree comparable to that of traditional regimens.

In simulated trials of 50 patients, the machine-learning model designed treatment cycles that reduced the potency to a quarter or half of nearly all the doses while maintaining the same tumor-shrinking potential. Many times, it skipped doses altogether, scheduling administrations only twice a year instead of monthly.

“We kept the goal, where we have to help patients by reducing tumor sizes but, at the same time, we want to make sure the quality of life—the dosing toxicity—doesn’t lead to overwhelming sickness and harmful side effects,” says Pratik Shah, a principal investigator at the Media Lab who supervised this research.

The paper’s first author is Media Lab researcher Gregory Yauney.

Rewarding good choices

The researchers’ model uses a technique called reinforced learning (RL), a method inspired by behavioral psychology, in which a model learns to favor certain behavior that leads to a desired outcome.

The technique comprises artificially intelligent “agents” that complete “actions” in an unpredictable, complex environment to reach a desired “outcome.” Whenever it completes an action, the agent receives a “reward” or “penalty,” depending on whether the action works toward the outcome. Then, the agent adjusts its actions accordingly to achieve that outcome.

Rewards and penalties are basically positive and negative numbers, say +1 or -1. Their values vary by the action taken, calculated by probability of succeeding or failing at the outcome, among other factors. The agent is essentially trying to numerically optimize all actions, based on reward and penalty values, to get to a maximum outcome score for a given task.

The approach was used to train the computer program DeepMind that in 2016 made headlines for beating one of the world’s best human players in the game “Go.” It’s also used to train driverless cars in maneuvers, such as merging into traffic or parking, where the vehicle will practice over and over, adjusting its course, until it gets it right.

The researchers adapted an RL model for glioblastoma treatments that use a combination of the drugs temozolomide (TMZ) and procarbazine, lomustine, and vincristine (PVC), administered over weeks or months.

The model’s agent combs through traditionally administered regimens. These regimens are based on protocols that have been used clinically for decades and are based on animal testing and various clinical trials. Oncologists use these established protocols to predict how much doses to give patients based on weight.

As the model explores the regimen, at each planned dosing interval—say, once a month—it decides on one of several actions. It can, first, either initiate or withhold a dose. If it does administer, it then decides if the entire dose, or only a portion, is necessary. At each action, it pings another clinical model—often used to predict a tumor’s change in size in response to treatments—to see if the action shrinks the mean tumor diameter. If it does, the model receives a reward.

However, the researchers also had to make sure the model doesn’t just dish out a maximum number and potency of doses. Whenever the model chooses to administer all full doses, therefore, it gets penalized, so instead chooses fewer, smaller doses. “If all we want to do is reduce the mean tumor diameter, and let it take whatever actions it wants, it will administer drugs irresponsibly,” Shah says. “Instead, we said, ‘We need to reduce the harmful actions it takes to get to that outcome.'”

This represents an “unorthodox RL model, described in the paper for the first time,” Shah says, that weighs potential negative consequences of actions (doses) against an outcome (tumor reduction). Traditional RL models work toward a single outcome, such as winning a game, and take any and all actions that maximize that outcome. On the other hand, the researchers’ model, at each action, has flexibility to find a dose that doesn’t necessarily solely maximize tumor reduction, but that strikes a perfect balance between maximum tumor reduction and low toxicity. This technique, he adds, has various medical and clinical trial applications, where actions for treating patients must be regulated to prevent harmful side effects.

Optimal regimens

The researchers trained the model on 50 simulated patients, randomly selected from a large database of glioblastoma patients who had previously undergone traditional treatments. For each patient, the model conducted about 20,000 trial-and-error test runs. Once training was complete, the model learned parameters for optimal regimens. When given new patients, the model used those parameters to formulate new regimens based on various constraints the researchers provided.

The researchers then tested the model on 50 new simulated patients and compared the results to those of a conventional regimen using both TMZ and PVC. When given no dosage penalty, the model designed nearly identical regimens to human experts. Given small and large dosing penalties, however, it substantially cut the doses’ frequency and potency, while reducing tumor sizes.

The researchers also designed the model to treat each patient individually, as well as in a single cohort, and achieved similar results (medical data for each patient was available to the researchers). Traditionally, a same dosing regimen is applied to groups of patients, but differences in tumor size, medical histories, genetic profiles, and biomarkers can all change how a patient is treated. These variables are not considered during traditional clinical trial designs and other treatments, often leading to poor responses to therapy in large populations, Shah says.

Source: Read Full Article

Researchers look to worms for a new model of a peripheral nervous system disease

Studying transthyretin amyloidoses-a group of progressive nerve and cardiac degenerative diseases caused by the buildup of misfolded transthyretin (TTR) proteins in the body-has long been hampered by the lack of animal models of the disease. Mice, for instance, don’t show the same symptoms as humans, even when misfolded TTR accumulates in their organs.

Now, scientists at Scripps Research have discovered that Caenorhabditis elegans, a nematode, or microscopic roundworm, develops similar nerve damage to human patients when their muscle cells are genetically engineered to produce TTR.

“This is really the first model that recapitulates what we see in humans both with regards to the molecular and cellular signatures of the disease, and the symptoms,” says Sandra Encalada, Ph.D., Arlene and Arnold Goldstein Assistant Professor of Molecular Medicine at Scripps Research.

The new C. elegans model, which Encalada and her team described recently in the journal Proceedings of the National Academy of Sciences, has already let Scripps Research scientists make inroads into understanding how TTR proteins become misfolded and aggregate to cause disease in neurons.

In humans, TTR is produced and secreted by the liver, where sets of four copies of the protein assemble together into tetrameric TTR that’s sent out into the bloodstream. In the blood, TTR normally binds and transports the thyroid hormone thyroxine, as well as vitamin A bound to retinol binding protein to deliver them throughout the body.

But there’s a ticking clock: the four TTR copies also fall apart over time, and then, in some cases, change their conformation or shape and regroup or misassemble into larger aggregates that deposit in tissues. There is genetic and pharmacologic evidence that this process causes neurodegeneration.

People can suffer from a variety of diseases based on the sequence of TTR that misfolds and misassembles and depending on where misfolded TTR aggregates accumulate. In the two most common forms of TTR amyloid disease, the protein accumulates in the heart-causing cardiac symptoms-or in the nerves of the legs and arms-causing a peripheral neuropathy. While some people who develop these diseases have mutations in their TTR protein, making them more prone to aggregate, others have normal TTR that can also misfold and misassemble.

“We know quite a bit about the molecular dynamics of how TTR comes apart and how it creates aggregates,” says Encalada. “But until now we didn’t have any mechanism at the cellular level. How do heart or nerve cells degenerate when TTR aggregates?” Scientists working with dozens of rodent and fruit fly models have failed to replicate what is seen in humans with these conditions.

In an attempt to answer these questions, Encalada and her collaborators engineered C. elegans to produce TTR in their muscle cells. They then tested the bodies of the nematodes for the presence of TTR. The protein, they showed, was secreted out of the muscle cells and into the worms’ body cavity. And just as in humans, the TTR broke down from tetramers and converted into misfolded and aggregated TTR molecules over the course of about a week.

When the researchers gave the nematodes a mutated version of TTR known to cause progressive peripheral neuropathy in humans, the worms showed abnormal growth of sensory nerve cells, and lost the ability to feel pain and temperature -the same impairments that are seen in humans. Moreover, when the worms were treated with drugs that ameliorate TTR peripheral neuropathy in humans, the worms showed dramatic improvement of the aforementioned degenerative phenotypes.

Encalada’s team tracked where the TTR was going in the worms’ bodies, and they found that tetramers of the protein secreted from the muscle, accumulated in the cells responsible for breaking down the body’s waste. These cells, the researchers showed, were degrading TTR and preventing the production of toxic aggregates. Deleting these cells enhanced the aggregation of TTR and increased the percentage of animals that had signs of nerve disease, including loss of pain sensation, as observed in humans.

“The big picture is that we were able to modulate levels of TTR degradation without touching neurons or the muscle cells producing TTR,” says Encalada. “In humans, being able to tweak levels of TTR degradation could act as a means of stopping TTR toxicity.”

Source: Read Full Article

Doxorubicin disrupts the immune system to cause heart toxicity

Doxorubicin is a chemotherapy drug widely used in ovarian, bladder, lung, thyroid and stomach cancers, but it carries a harmful side effect. The drug causes a dose-dependent heart toxicity that can lead to congestive heart failure.

University of Alabama at Birmingham researchers now describe an important contributor to that heart pathology—disruption of the metabolism that controls immune responses in the spleen and heart. These immune responses are vital for heart maintenance, repair and control of inflammation. This dysregulated immunometabolism impairs resolution of inflammation, and chronic, non-resolving inflammation leads to advanced heart failure.

Immunometabolism is the study of how metabolism regulates immune cell function, and it is a recent and growing aspect of immunology. Two key players in immunometabolism are immune-responsive enzymes called lipoxygenases and cyclooxygenases. These immune-sensitive enzymes create a variety of bioactive lipid mediators that regulate immune cell responses.

The UAB researchers, led by Ganesh Halade, Ph.D., an assistant professor in the UAB Department of Medicine’s Division of Cardiovascular Disease, used a mouse model to study the effect of doxorubicin on immunometabolism. In the mice, doxorubicin induced fibrosis in the heart, increased the programmed cell death called apoptosis and impaired the pumping of the heart. The drug also caused a wasting syndrome in the heart and the spleen.

Mounting research has shown that the spleen—which acts as a reservoir of immune cells that speed to the site of heart injury to begin clearance of damaged tissue—plays a leading role in the initiation of immune response after a heart attack. Now, Halade and colleagues have found that the doxorubicin is also involved in the deleterious response to the spleen.

First, the UAB researchers found that doxorubicin induced irreversible dysregulation that lowered levels of lipoxygenases and cyclooxygenases in the left ventricle of the heart. This reduced the levels of bioactive lipids mediators produced by these enzymes, mediators that usually would help resolve inflammation.

Second, in the spleen, doxorubicin also poisoned a special group of marginal zone immune cells called CD169+ macrophages, causing the spleen to diminish in size. This loss of specialized macrophages means an impaired host defense system because these unique macrophages usually coordinate the first-responders monocyte deployment plan to sites of injury or infection in order to synthesize bioactive lipids to activate the resolution of inflammation.

Third, doxorubicin caused an imbalance of the cell-signaling molecules called chemokines and cytokines, and this imbalance suggests suppressed defense capacity of spleen-leukocyte immune cells. Specifically, the researchers found decreased levels of tumor necrosis factor-alpha in the spleen, and they found decreased levels of the immune-cells reparative marker MRC-1, also known as CD206, in the heart.

Source: Read Full Article

Microfluidic system incorporates neuroinflammation into ‘Alzheimer’s in a dish’ model

Building on their development of the first culture system to replicate fully the pathology behind Alzheimer’s disease, a Massachusetts General Hospital (MGH) research team has now produced a system that includes neuroinflammation, the key biological response that leads to the death of brain cells. The investigators describe their system, which incorporates the glial cells that that not only surround and support neurons but also provide some immune system functions, in a paper published in Nature Neuroscience.

“Our original ‘Alzheimer’s in a dish ‘ system recapitulated the plaques and tangles typically seen the brains of patients with Alzheimer’s disease, but did not induce neuroinflammation,” says Rudolph Tanzi, Ph.D., director of the Genetics and Aging Research Unit in the MassGeneral Institute for Neurodegenerative Disease (MIND) and co-senior author of the current paper. “Studies have shown that we can have many plaques and tangles in our brains with no symptoms, but when neuroinflammation kicks in, exponentially more neurons die and cognitive impairment leading to dementia is induced. A complete model of Alzheimer’s pathology needs to incorporate that ‘third leg of the stool’.”

In their 2014 Nature paper, the MGH team described using a gel-based, three-dimensional culture system—developed by Doo Yeon Kim, Ph.D., of the Genetics and Aging Unit, also a co-senior author of the current study—to induce the formation of both amyloid-beta plaques and neurofibrillary tangles in human neural cells carrying gene variants associated with early-onset, familial Alzheimer’s disease (FAD). That study confirmed that amyloid deposition was the essential first step leading to the formation of tangles containing the pathogenic form of the protein tau

The updated system also brings in technology developed by co-senior author Hansang Cho, Ph.D., now of the University of North Carolina at Charlotte, when he was a postdoctoral fellow in the MGH BioMEMS Resource Center . In a 2013 Scientific Reports paper, Cho and his co-authors reported using a microfluidic device consisting of two circular chambers, one inside the other, to measure the migration of microglia—glial cells that function as nervous system immune cells—from the outer chamber into the amyloid-loaded inner chamber by means of connecting channels.

For the current study Cho and lead author Joseph Park, Ph.D., of the MGH Genetics and Aging Unit, used the ‘Alzheimer’s in a dish’ system to culture neural stem cells with FAD variants in the central chamber of Cho’s device. Several weeks later, the neurons and astrocytes, glial cells that support and insulate neurons, that had differentiated were found to contain elevated levels of amyloid-beta and tau, as well as inflammatory factors known to contribute to the neuroinflammation seen in Alzheimer’s disease.

When human microglia were added to the outer chamber of the device, they soon began to show structural changes signifying their activation and migrate through the channels into the inner chamber. Once the microglia arrived in the inner chamber, they directly attacked neurons, causing visible damage to key structures, while levels of inflammatory factors like TNF-alpha, IL-6 and IL-8 rose significantly. Six days later the central chambers had lost 20 percent of both their neurons and their astrocytes.

Source: Read Full Article

How your immune system promotes friendly gut bacteria

They found that immunoglobulin A (IgA) antibodies released by the gut can alter how bacteria express their genes.

This encourages the microorganisms to form communities that work together to defend against disease and safeguard the health of their “host.”

Antibodies are involved in the immune response and have long been recognized as fighters of harmful agents. More recently, it has also emerged that they play an important role in regulating good bacteria in the gut.

But until the new study — now published in the Journal of Experimental Medicine — it was not clear how they did this.

Senior study author Dr. Keiichiro Suzuki, of the RIKEN Center for Integrative Medical Sciences in Japan, says that they already “knew that [IgA] contributed in some way to gut health.”

But they were excited, he adds, to find that the “new mechanism” that they uncovered “actually promotes symbiosis among the bacteria that inhabit the mucus membrane of the gut.”

Gut microbiota and IgA

Our guts contain “complex and dynamic” communities of bacteria and other microorganisms that play an important role in health and disease.

Collectively known as the gut microbiota, these tiny creatures have evolved in partnership with us over millennia to mutual benefit.

In their millions, they strengthen the hundreds of square feet of our guts, shape their lining, regulate metabolism, collect energy, defend against pathogens, and help develop our immune systems.

In previous work, the team had shown that IgA helps control the mix and location of bacteria in the gut, and that its stabilizing influence seems to comes from an ability to “coat” bacteria.

They found that a common species of human gut bacteria called Bacteroides
thetaiotaomicron
“was particularly susceptible to coating by IgA.”

IgA alters gene expression

In the new study, the scientists probed the molecular underpinning of this activity.
They discovered that IgA alters gene expression in B. theta.

Dr. Suzuki and team called these proteins “mucus-associated functional factors (MAFFs),” and they discovered that they seemed to be doing two things to promote friendly gut bacteria.

First, the MAFFs seemed to help B. theta grow in the mucus-secreting lining of the gut. And, second, they stimulated B. theta to make molecules that encouraged the growth of Clostridiales and other friendly bacteria.

The researchers confirmed this beneficial influence of MAFFs in mice. They injected the mice with B. theta that did not produce an abundance of MAFFs. The mice’s gut bacteria changed and the animals become prone to colitis, or inflamed gut.

The team hopes that the findings will eventually lead to new treatments for inflammatory bowel disease.

The MAFF system is also present in humans so it is an interesting target of research, but there is still much to be investigated.”

Dr. Keiichiro Suzuki

Source: Read Full Article