Lehigh University
Lehigh University


Attacking from all angles

Despite impressive medical advances, cancer continues to frustrate clinicians—and overcome its victims. At Lehigh, a diverse group of researchers are employing varied techniques to try and defeat cancer once and for all.
A century is but a small speck on the timeline of the history of science. But the last century has yielded more scientific breakthroughs in the area of cancer research than anyone could have possibly foreseen 100 years ago: A greater collective understanding of what causes cancer, how to prevent it and how to treat it has saved countless lives.

Though for all the astonishing advances made by physicians and scientists, there is much we still don’t know about how cancer cells develop, the mechanisms they use to spread, and how to stop them. At Lehigh, a cadre of cancer researchers is working to answer some of these very questions. University faculty, supported by federal research grants, are leading investigations that are enhancing our knowledge of the biological and chemical properties of disease. Their findings may lead not only to safer drugs to fight cancer, but just as importantly, more accurate and cost-effective screening tools to catch the disease before it has a chance to ever develop.

The latter goal is the focus of a study currently being led by the Xiaolei Huang, an associate professor in Lehigh’s Computer Science & Engineering Department who is currently working to defeat cancer through two different research projects. Huang is a computer scientist who uses imaging analysis to study the functioning of human cells, the brain, the eye and the body as a whole. More recently, she’s turned her attention to cervical cancer—which, despite great advances in cancer treatment, remains one of the most common causes of cancer death among women worldwide.

Early detection through widespread use of the Pap screening test and diagnostic procedures has significantly improved survival rates, yet each year, 12,000 women are still diagnosed with cervical cancer, and 4,000 women will die of the disease in America, according to the Centers for Disease Control and Prevention.

In poorer countries, cervical cancer remains one of the leading causes of cancer death among middle-aged women, Huang says. More than 275,000 women died from cervical cancer worldwide in 2008, and nearly 90 percent of the deaths occurred in developing parts of the world, according to the American Cancer Society. Cervical cancer can be a silent killer; it’s a slow-growing cancer and symptoms often do not start until the cancer has spread to nearby areas.
In many poor nations, early detection through Pap testing isn’t available due to a lack of laboratories and trained personnel for conducting screening, diagnostic, and follow-up tests.

An alternative test is the cervigram, which involves taking a digital photo of the cervix. While the technology is low cost and is more widely accessible in resource-poor regions, experts differ on the effectiveness of cervigram test results as a conclusive diagnostic tool.

Through a grant from the National Institutes of Health, Huang was granted access by the National Library of Medicine and the National Cancer Institute to databases containing some 100,000 anonymous images of cervical lesions and the accompanying notes made by physicians and imaging professionals.

Huang’s an expert in image object segmentation, and her group developed a computer system that categorized each image according to color, texture, size and shape. With access to comparison images, clinicians can more accurately grade the severity of new lesions. “The advantage of computers is their ability to do quantitative things,” she says. “To make a precise diagnosis, we need quantitative information.”

After the databases were indexed, Huang next created a program capable of correctly classifying the severity of disease in a new, undiagnosed patient by comparing that patient’s image against the known patient outcome data from other records in the database. “My interest was, ‘Can I look at these (unlabeled) images and … develop some computer-assisted interpretation of these images so we can make this diagnosis more accurately?” Huang says.

The system could prove to be a powerful tool to help doctors differentiate low-grade cervical lesions from high-grade lesions and invasive cancer, possibly negating the need for expensive and invasive follow-up tests.

Other similar methods only perform processing or segmentation of cervigrams without patient classification. Huang’s cervigram image interpretation algorithm, by contrast, can produce a cervical dysplasia diagnosis with high accuracy. In fact, in a recent trial using 280 patient cases from women in Costa Rica, Huang found that adding image study to traditional Pap and HPV tests significantly improved the accuracy of diagnosis related to high-grade lesions. 
Stop the growth, stop the disease 
In a separate five-year, $1.3 million project supported by the National Institutes of Health, Huang is collaborating with Dimitrios Vavylonis, an associate professor of physics in Lehigh’s College of Arts and Sciences. They’ve teamed up to study how actin proteins come together to form microfilaments, a key component of the cytoskeleton—or structural framework—of yeast cells. Through their work, the duo hopes to gain a fuller understanding how cancer grows—and, by extension, how that growth can be halted. “Actin is a very important protein in the cells,” Huang says. “It plays an important role in cell division and cell movement. As it relates to cancer research, given this fundamental understanding of how cells divide and what triggers abnormal cell division, we can aid in the search for drugs that block cancer spread.”

When a cell is about to divide, Huang explains, the actin filaments in the cytoskeleton start to condense into center of the cell, forming a dense ring-like structure that constricts until it snaps. The cell then divides into two identical daughter cells. Biologists have many theories on where actin filaments come from, how they mesh into this cohesive network and how they coordinate to achieve these tasks. What they don’t have is any firm answers.

Actin filaments can be seen on three-dimensional images of cytoskeletons taken over time with confocal microscopy. These images offer clues on how the network changes over time. Huang and Vavylonis proposed developing computer software capable of tracking individual filaments and segmenting them by pinpointing their centerline, but the task was complicated by the fact that the intensity (or brightness) varied along each individual filament. Additionally, the contrast between the filaments and the background created photographic noise that needed to be cancelled out. “For a computer this was not trivial at all,” Huang says. “The challenge was to develop a quantitative image analysis algorithm to be able to deal with all these issues. If we could extract out all the filaments and identify the junctions, we would have their topology.”

Recently, Huang and Vavylonis took a big step toward achieving that goal: A student in Huang’s lab developed new software that can automatically find the centerline for filaments comprising an actin meshwork in yeast cells, which are commonly used to study biological pathways and in drug research. The results of their study were published in the journal Cytoskeleton.
Flipping the switch
In the Chemistry department, another Lehigh researcher is looking to devise new methods for inhibiting the growth of a particularly virulent type of cancer cells. Assistant Professor Marcos Pires is studying a protein called PAD4—short for peptidyl arginine deiminase type 4—that plays a key role in regulating immune suppression and gene expression. The enzyme is believed to help certain cancer cells become more aggressive and more resistant to drug therapies. Pires first became interested in PAD4 after a conversation one day last year with his sister, also a researcher. The lab where she worked was exploring why PAD4 showed up more in cancer cells than in healthy cells in breast cancer patients. Researchers have found no good way to measure PAD4 activity, which is the only way to verify whether potential drug agents would be effective at neutralizing it.

Pires’ interest lies in manufacturing synthetic molecules that can be used to probe biological systems and responses. He set out to develop a test capable of screening as many drugs as possible to find what works and what doesn’t. “Our focus isn’t just making molecules for the hell of it, but to try to answer something about a living organism,” he says. “It occurred to me we could design a small molecule that would look like the protein that PAD4 acts on.”

The work involved creating a “light bulb test” of sorts. First, the researchers synthesized PAD4 and attached a chemical signal to it that would illuminate if the protein is working. This switching on/off effect can only be observed under fluorescence. The trial worked, and the results of the study were published in the research journal ChemBioChem. One of many next steps is seeking NIH assistance in using the test to screen thousands of available molecules for the purpose of finding the ones that could potentially work as a drug. A Penn State research team published a study last year identifying a drug that inhibits the PAD4 protein, and results showed it capable of reducing tumors in mice by 70 percent. But the drug was highly toxic, making it a poor candidate for use in humans. Pires is collaborating with that team and others to try and develop more effective agents using the assay he developed.

He is also interested in replicating the test on PAD4 samples from actual human cancer cells. Such work could yield greater knowledge of the protein’s structure and potentially lead to new drugs that more safely target cancer cells without producing side effects in normal tissues. “We’ve never had a way to measure this before,” he says. “There could be surprises in how PAD4 gets turned on/off in live human cells.”
New approaches, new hope
Treatment of cancer patients has stalled in recent history despite advances in the biomedical field. The death rate for breast cancer has remained unchanged for the past three decades, driving the need the more highly targeted therapies. In yet another project that aims to tackle that challenge, Pires is collaborating with fellow Lehigh chemistry Assistant Professor Damien Thévenin to develop a novel methodology for delivering anti-cancer agents to tumors without damaging healthy cells. The peptide “homing device” technique they’ve pioneered releases the unmodified drug only when it reaches the inside of targeted cancer cells.

Most current targeting strategies take aim at specific cancer surface proteins. However, they have had limited success against solid tumors in part because cancer cells are almost identical to healthy non-cancerous cells. Drugs inadvertently accumulate in healthy tissues, limiting their efficacy and causing toxic side effects. For this reason, less than 10 percent of new drugs move past Phase III clinical trials.
Cancer cells do have important differences than other cells, however, and Thévenin has focused specifically on one element: The fact that malignant tumors create an acidic micro-environment around themselves. With that in mind, Thévenin’s created a homing device—a pH(Low) Insertion Peptide (pHLIP)—that selectively targets tumors in mice solely based on their acidity. In their research, Thévenin’s peptide is paired with a unique and stable chemical link Pires created to latch a cancer-fighting drug with an antibody capable of bonding with specific proteins inside the cancerous cells. Once the bond is made, the “traceless linker” they designed switches “off” and the unmodified, FDA-approved cancer chemotherapy drug is delivered. The project was supported by a Lehigh University Faculty Innovation Grant; those awards support members of the faculty in establishing new research project or expanding existing projects, with the hope to creating a new line of research that can continue to grow for years to come.

In the realm of cancer, at least, Lehigh seems to be moving in precisely that direction. Because in its own unique way—interdisciplinary, practical, highly focused—Lehigh is taking the challenge of beating cancer head-on—and yielding real results, too. 
Illustration by Simon Pemberton 

Story by Brian M. Schleter

Posted on Tuesday, October 01, 2013

share this story: