Talented bacteria detect cancer, diabetes

E. coli has come a long way from sickening people at picnics. The authors of two separate studies have reengineered the humble bacterium, shown here, to detect cancerous tumors in the liver and the spilling of sugar into urine--both without so much as a pinprick.

E. coli has come a long way from sickening people at picnics. The authors of two separate studies have reengineered the humble bacterium, shown here, to detect cancerous tumors in the liver and the spilling of sugar into urine--both without so much as a pinprick.

(Chris Bickel / Science Translational Medicine)

You might want to park at the door any assumptions about bacteria being primitive, inert or yucky. This is a story about genetically engineered forms of bacteria that may someday diagnose cancer by analyzing a urine sample, and that have already shown a knack for recognizing the hallmarks of diabetes in humans without so much as a pinprick.

In its unaltered form, these bacteria belong to the much-maligned Escherichia coli family. Like the picnic-spoiling strain that induces vomiting and diarrhea and casts suspicion on Aunt Rita’s potato salad, this E. coli enters the gut and makes its way throughout the body. But there the likeness ends.

In two articles published Wednesday in the journal Science Translational Medicine, researchers reported they had equipped E. coli bacteria with genetically encoded digital amplifying genetic switches. Those engineering tweaks transformed the bacteria into living sensors, capable of staying in a mouse’s body for as long as a month.

In one of studies, conducted on mice, the engineered bacteria provided warning--in the form of a visible change in the color of the animal’s urine--when a tumor had established itself in the liver.


Essentially, scientists improved the E. coli bacteria’s ability to pass through the gut’s walls and enter the liver of a mouse, gravitating directly toward cancer cells, for which these bacteria have an affinity. Once the organisms find and colonize a tumor there, they have been engineered to begin producing an enzyme that is visible as a detectable change in the color of the affected mouse’s urine.

Liver cancer is a difficult malignancy to spot early, since tumors don’t show up well on imaging scans. Obesity and hepatitis infections are driving up the number of patients at risk for the disease in the U.S. and worldwide. And liver cancers frequently metastasize to the colon, lungs, ovaries or pancreas before they are detected. For all those reasons, finding a reliable way to detect such cancers early is crucial.

The bioengineering efforts were conducted by researchers hailing from the Massachusetts Institute of Technology and UC San Diego (liver cancer) and University of Montpellier in France and Stanford University (diabetes). Their work is a key component of broader efforts to make the diagnosis and treatment of diseases such as cancer increasingly precise and targeted. Using the emerging techniques of synthetic biology, scientists are engineering living cells so they can perform specific functions in medicine and in environmental toxicology.

“The field of synthetic biology aims to design and engineer biological components and systems for specific purposes,” said Jessica Tucker, director of the National Institute of Biomedical Imaging and Bioengineering’s program in synthetic biology for technology development.

These papers “elegantly merge the promise of synthetic biology with very practical and clinical applications in biosensors,” she added.

“Our platform architecture is highly modular and could be repurposed for various applications,” wrote the researchers. As scientists discern the genetic signature of certain cancers, bioengineering bacteria could, for instance, be programmed to recognize those signatures, rendering earlier diagnoses, monitoring a patient’s response to treatment, and delivering treatment.


In mice, the visible evidence that “bactosensors” had latched onto cancer cells came within 24 hours. If the bactosensor proves safe and effective in humans--whose gut microbiomes are a very different environment than that of mice--it could provide early warning of the presence of tumors or perhaps even circulating metastases. That, in turn, could give physicians and their patients treatment options at a stage when they are more likely to be effective.

With some changes, the enzyme-producing bacteria might also prove effective in the detection of other cancers of the gastrointestinal tract, such as colorectal cancer.

In urine samples from humans tested in the second study, researchers tried their bactosensors to probe for glycosuria, the presence of sugar in urine that is a telltale sign of uncontrolled diabetes.

Suspended in hydrogel beads, the engineered E. coli sensors turned urine samples fluorescent red in almost 89% of cases where glycosuria was present. And they rarely sent up a false alarm, suggesting diabetes where it was not present in just over 3% of cases. Those measures of “sensitivity and specificity”--the ability to detect disease without creating a dragnet of false positives--made the bactosensors almost as reliable as urine dipsticks currently used in physicians’ offices.

Such living sensors could one day allow the kind of monitoring done in a physician’s office to follow patients home. And they might allow physicians in remote or field clinics to diagnose disease and monitor treatment response. For E. coli, that would be a big step up from spoiling picnics.

Follow me on Twitter @LATMelissaHealy and “like” Los Angeles Times Science & Health on Facebook.