Epigenetic modification of DNA by cytosine methylation to produce 5-methylcytosine (5mC) has become well-recognized as an important epigenetic process in human health and disease. Recently, further modification of 5mC by the ten eleven translocated (TET) family of enzymes to produce 5-hydroxymethylcytosine (5hmC) has been described. Brent Orr and Charles Eberhart, together with collaborators in the Cancer Center, used immunohistochemistry to evaluate the distribution of 5hmC in human brain during different periods of development and in a large series of 225 gliomas. They found that during development, 5hmC levels are high in more differentiated compartments like the fetal cortex, but low in the periventricular progenitor cell regions. In brain tumors, 5hmC levels were high in low grade tumors and reduced in malignant glioma. Additionally, they identified a significant relationship between low levels of 5hmC and reduced survival in malignant glioma. This observation was further supported by in silico analysis showing differential expression of genes involved in 5hmC homeostasis in aggressive subsets of glioblastoma. Finally, they found that several genes involved in regulating the levels of 5hmC are also prognostic in malignant glioma. These findings suggest that 5hmC regulation in malignant glioma may represent an important determinant of tumor differentiation and aggressive behavior, as well as a potential therapeutic target. The study was published in PLoS One (http://www.ncbi.nlm.nih.gov/pubmed/22829908).

His research interests focus on understanding the pathogenesis of neurodegenerative disorders like Alzheimer’s disease and Parkinson’s disease. He will use Drosophila as a model to study the mechanisms underlying central nervous system malfunction in humans. By expressing pathological human genes in the fly, he can generate abnormal phenotypes, such as slowed motor activity or degeneration of the retina. These phenotypes can then be used in conjunction with the rich genetic toolbox that Drosophila researchers have developed to identify pathways that contribute to the degeneration. Their small size, rapid generation time, and low costs for maintenance make fruit flies ideal for studying neurodegenerative disease. The Drosophila models provide a platform for genome-wide screens and unbiased genetic screens to identify components of pathological pathways.

“We don’t question whether brain cancer cells have the potential to express blood vessel markers and may occasionally find their way into blood vessels, but we do question the extent to which this happens,” says Charles Eberhart, M.D., Ph.D., chief of neuropathology at the Johns Hopkins University School of Medicine.  “In general, we find no evidence in our study that these vessels contain substantial amounts of cancer cells.”

Eberhart, professor of pathology, ophthalmology and oncology at Johns Hopkins, said he first encountered claims about the cancerous nature of tumor blood vessels about a year ago when he was invited to join students at a journal club meeting, a forum for discussing studies published in medical journals.  “My first reaction to this research was ‘How could this be true?’” says Eberhart. “Our clinical experience examining tissue from brain cancers does not support it.”

Studies have long demonstrated that malignant brain tumors contain large numbers of blood vessels to feed their growing demand for nutrients. The blood vessels are formed when tumors pump out growth factors that increase vessel production. Such studies opened the door to treatment strategies that specifically targeted blood-vessel growth and the vessel cells themselves.

More recently, scientists in Italy and the Memorial Sloan Kettering Cancer Center in New York published results of studies suggesting that these tumor blood vessels are made by primitive types of brain cancer cells that are a form of stem cells.  In their studies, they found tumor markers on blood vessel cells in 20 to 90 percent of their brain cancer samples.  The U.S./Italian research teams said their findings also suggested that the cancerlike blood vessels were more prone to drug resistance, potentially explaining why drugs like bevacizumab yield tumor-shrinking responses, but only for short periods. Bevacizumab is currently approved by the U.S. Food and Drug Administration for use in patients with colorectal, lung, kidney and brain cancers.

Eberhart said pathologists, including those who work on brain tissue, use certain tissue-based techniques to distinguish cancer cells from normal ones.  When evaluating specimens of brain tissue removed during surgery for suspected cancer, he said, most pathologists agree that blood vessel cells in these specimens consistently lack the molecular changes associated with cancer cells, according to Eberhart.  In fact, they often use these blood vessel cells as “normal controls” to compare with potentially cancerous ones.

After the journal club experience, Eberhart teamed up with fellow neuropathologist Fausto Rodriguez, M.D., and colleagues at the Dana Farber Cancer Institute and Harvard Medical School in Boston to look more closely at the molecular features of blood vessel cells in brain cancer samples.  They tested more than 100 samples from patients at Johns Hopkins and Dana Farber for EGFR and IDH1 markers, two common genes altered in brain cancer.

“We also used a marker called CD34 to differentiate vascular [blood vessel] cells from other types of cells,” says Rodriguez, assistant professor of pathology at Johns Hopkins.  The research teams found no more than 10 percent of their samples contained vascular cells with EGFR or IDH1 cancer markers, and in those rare tumor samples, only a few cells exhibited those markers.  The Johns Hopkins-Dana Farber-Harvard team tested all parts of the vessel walls for presence of the cancer markers.

Results of the team’s laboratory experiments were published in the online journal Oncotarget in January.

Although the two groups used different markers to identify vessel cells, Rodriguez says “there is no marker that is absolute for each cell.”

Eberhart and Rodriguez noted that the U.S./Italian research teams focused mainly on cell-by-cell research techniques in dissociated specimens to evaluate cancer markers, losing associations that can be made by looking at a cell’s shape and physical relationship within clusters of cells.  The Johns Hopkins and Dana Farber researchers conducted studies examining cells in intact tissue.

“Pathologists with extensive experience in examining cells become accustomed to quickly identifying a blood vessel cell from a normal cell, and we can gain a lot of information when we look at how cells connect with other cells in real-life examples,” notes Rodriguez, who says that his team’s findings could potentially apply to any cancer thought to contain stem cells.

In addition to Eberhart and Rodriguez, the research team included Brent Orr from Johns Hopkins and Keith Ligon from the Dana Farber Cancer Institute/Harvard Medical School.

Funding for the study was provided by a National Institutes of Health postdoctoral fellowship (T32CA067751) to Orr and a grant (5R01NS055089) to Eberhart.

http://www.alzforum.org/new/detail.asp?id=3000

 Like the caretaker of a condemned building shutting off the lights for the last time, DNA methyltransferases shut off motor neuron genes during apoptosis. This is according to a paper in the November 16 Journal of Neuroscience by Lee Martin of the Johns Hopkins University School of Medicine in Baltimore, Maryland. The researchers report that the action of methyltransferases was crucial for apoptosis in cell culture and mice. Their work implies that epigenetic methylation during cell death is not static, but malleable. Martin suggested that treatments modulating methyltransferase activity might temporarily maintain neurons during injury or disease, such as amyotrophic lateral sclerosis (ALS).

Epigenetic tags control processing of DNA without altering its sequence. They tell the transcriptional machinery which genes to turn on and which to switch off. Histone tags, primarily acetylations, regulate many processes, including learning and memory (see ARF related news story on Peleg et al., 2010). They have been linked to schizophrenia (see Schizophrenia Research Forum story) and longevity (see ARF related news story on Greer et al., 2011). DNA methyltransferases (Dnmts) are enzymes that silence genes by adding methyl groups to cytosines in cytosine- and guanine-rich areas known as CpG islands. These islands are common in promoters. DNA methylation has mostly been studied in cancer, but aberrant methylation patterns are known to be associated with Alzheimer’s disease (see ARF related news story on Mastroeni et al., 2009), Rett syndrome (Kondo et al., 2000), and synaptic plasticity (Levenson et al., 2006).

Martin suspected a direct role for DNA methylation in neurodegeneration. He hypothesized that when motor neurons die in ALS, the cell uses methylation to turn off survival factors. Martin recruited graduate student and first author Barry Chestnut—who has since moved on to a postdoc at the National Cancer Institute in Bethesda, Maryland—from a cancer lab to take a look at methylation in ALS.

Using methylation-sensitive restriction enzymes to chop up the genome, Chestnut deduced that DNA methylation patterns in people who had ALS were very different from those in control subjects. To study the link between methylation and motor neuron death, he turned to cultured NSC34 cells—a fusion of spinal cord motor neurons with neuroblastoma cells—. Chestnut focused on Dnmt1, a maintenance enzyme that adds methyl groups to new DNA strands during replication, and Dnmt3a, which catalyzes de-novo methylation when post-mitotic cells turn off genes. He quickly ran into a problem overexpressing Dnmt3a—the cells just started dying. “They were very apoptosis-like,” said Martin, with caspase expression typical of the process. In contrast, overexpressing Dnmt1 did not kill the cells.

How does excess Dnmt3a kill neurons? Surprisingly, Chestnut and colleagues discovered that the methyltransferase tagged with green-fluorescent protein not only turned up in the nucleus, as one would expect, but also in mitochondria and synaptic terminals. Scientists didn’t know previously that Dnmts appeared outside the nucleus in neurons, said Courtney Miller of the Scripps Research Institute, Jupiter, Florida, who was not involved in this work. The question, of course, is what the enzymes could be doing in mitochondria and synapses. Methylating DNA is a possibility that would be a “striking” result, said Bernard Futscher of the University of Arizona in Tucson, noting that the mitochondrial genome is normally unmethylated. Futscher, who was not involved in the study, said he would like to see direct evidence for methyl groups on mitochondrial DNA. NSC34 cells are a cancerous, immortalized line, he added, so their biology with respect to cell death could be abnormal.

Rajiv Ratan of the Burke Medical Research Institute in New York pointed out that histone deacetylases were once thought to only act on histones, but are now known to acetylate cytoplasmic proteins such as tubulin. It is possible that the enzymes are methylating something other than DNA, said Ratan, who was not an author on the study. Another member of the Martin team is now looking into what mitochondrial genes Dnmt might target, Martin said. He also suggested that DNA methylation at this location could be a mechanism of synaptic plasticity.

To study the role of Dnmts in cell death, Chestnut induced apoptosis in NSC34 cells. Within hours, expression of both Dnmt1 and Dnmt3a increased. The researchers then blocked Dnmt activity or expression in apoptotic cells. This was not particularly effective for Dnmt1, but was for Dnmt3a, so the team concluded that Dnmt3a was more important for the cells to complete apoptosis. Methylated cytosine accumulated in the nuclei of apoptotic cells, but this methyl-cytosine disappeared if Dnmt3a was inhibited.

To explore the physiological relevance, the researchers examined a mouse model of motor nerve damage. Chestnut cut the sciatic nerve where it exits the spinal cord, which causes apoptosis in 80 percent of nerves (Martin and Liu, 2002; Martin et al., 2005). Dnmt1, and particularly Dnmt3a, rose in the dying cells less than a week after lesioning. Dnmt3a accumulated in mitochondria as well as in synaptic terminals, and within two days of the lesion, methylated cytosine accumulated in the dendrites, nucleus, and cell body, including mitochondria.

In time, methylation increases and the cells get smaller and smaller, Martin said. That shrinkage was preventable. When the researchers infused the mice with RG108, a general Dnmt inhibitor, it not only prevented cell shrinkage and apoptosis, but the motor neurons grew even larger than normal. “Apparently, motor neurons like to have their genomes hypo-methylated,” Martin said. He suggested this kind of treatment could perhaps sustain motor neurons after axon injury.

Finally, the researchers returned to human ALS tissue. ALS motor neurons contained more Dnmt1 and Dnmt3a than other motor neurons. “It is rather striking,” Martin said. “There is no doubt in my mind that in the ALS motor cortex, motor neuron DNA is getting extensively hyper-methylated.” Drugs like RG108 might help. Dnmt inhibitors are under trial as cancer treatments, noted Jian Feng, at the Mount Sinai School of Medicine in New York. Any anti-methylation treatment would have to be specific for motor neurons to avoid side effects that could result from activating silenced genes across the body, noted Mark Mattson of the National Institute on Aging in Bethesda, Maryland. For example, inhibiting Dnmt can block memory formation, said Yitshak Francis of Columbia University in New York (Miller and Sweatt, 2007). In addition, methyltransferases would probably only block ALS symptoms. It would just be keeping the lights on until doctors fix the root cause of the problem.—Amber Dance.

Reference:
Chestnut BA, Chang Q, Price A, Lesuisse C, wong M, Martin LJ. Epigenetic regulation of motor neuron cell death through DNA methylation. J Neurosci. 2011 Nov 16; 31(46):16619-16636. Abstract

“These tumors are slow-growing to start with, and sometimes stop growing, and now we have a pretty good idea of why that happens,” says Charles G. Eberhart, M.D., Ph.D., associate professor of Pathology, Ophthalmology and Oncology at Johns Hopkins. “These tumors also can suddenly become more aggressive, which we now think represents an inactivation of this tumor-suppressor gene, and this inactivity could be used as a marker to determine which patients need more therapy.”

Pilocytic astrocytoma arises in brain cells known as astrocytes, which, among many functions in the brain, help support neurons. These cancerous astrocytes have DNA mutations that force a growth-related gene, BRAF, into an abnormal, always-on state. Biologists call such cancer-driving genes oncogenes.

Eberhart and his team used a viral gene-transfer technique to deliver an oncogenic, always-on version of BRAF, to fetal brain cells in a lab dish. The idea was to create a cell model of pilocytic astrocytoma, to enable easier study of its growth patterns. As the researchers expected, the cells quickly formed tumorlike colonies — but the growth of these colonies soon sputtered out.

The same phenomenon, sparked by an oncogene, was first described six years ago in a study of the biology of skin moles. Moles typically begin in skin cells whose inherited or spontaneous mutations — often affecting BRAF — drive the cells’ growth beyond normal limits. “The oncogene drives the excessive growth of skin cells, which forms a mole. This overgrowth triggers the downstream activation of tumor-suppressor genes, which stops the mole from growing further,” says Eberhart.

In the current study, Eberhart and his colleagues found evidence that this same process, which is called oncogene-induced senescence, also occurs in pilocytic astrocytoma and minimizes its spread. As their tumor-model cells became senescent, the activity of p16, a well-known tumor-suppressor gene, increased and acted as a brake to stop further tumor growth.

Next, the researchers checked pilocytic astrocytoma samples from 66 patients, using a tissue registry at the Johns Hopkins Department of Pathology. Most (57 of 66) showed signs of p16 tumor-suppressor activity, and the remaining nine samples had no signs of p16 activity. Of the p16-active tumors, only two samples (3.6 percent) were from patients who had died of their cancer; however, three of the nine samples with inactive p16 (33 percent) were from patients who had died.

“Our hypothesis now is that these tumors become fast-growing and aggressive again when they can somehow find a way to shut off p16 and escape senescence,” says Eric Raabe, M.D., Ph.D., fellow in pediatric oncology at Johns Hopkins. “In many cases, a single tumor may contain some cells that are senescent plus others that have escaped senescence and started proliferating again,” he added.

In future work, Eberhart says, he and his colleagues will examine whether a new class of BRAF-inhibiting cancer drugs has the unintended side effect of shutting down p16. “Clinical trials of these BRAF inhibitors are now just starting in the U.S. and Europe,” he says. “We think it’s important to determine whether these drugs end up affecting the process of oncogene-induced senescence.”

The study was supported by the PLGA Foundation, Children’s Cancer Foundation, the Pilocytic/Pilomyxoid Astrocytoma Research Fund at Johns Hopkins Medicine, Lauren’s First and Goal, St. Baldrick’s Foundation Fellowship, and the Comprehensive Cancer Center, Freiburg, Germany.

Other researchers involved in the study were Kah Suan Lim, Alan Meeker, Xing Gang Mao, Deepali Jain, Eli Bar, Julia M. Kim, and Kenneth J. Cohen from Johns Hopkins; and Guido Nikkhah, Jarek Maciaczyk and Ulf Kahlert of the University Hospital, Freiburg.

“Protecting the body is absolutely essential to protecting the brain,” says senior study investigator and Johns Hopkins neuropathologist Vassilis Koliatsos, M.D. “Blast-related injuries, including what we call blast-induced neurotrauma, are the signature medical events of current wars, and improvements to body armor in addition to helmet- wearing are likely going to be needed if we want to minimize their threat to our soldiers’ health,” says Koliatsos, a professor at the Johns Hopkins University School of Medicine.

In a report to be published in the May edition of the Journal of Neuropathology and Experimental Neurology, Koliatsos and his team used a metal shock tube specially designed at Hopkins’ Applied Physics Laboratory to isolate the effects of an explosion’s primary blast wave on mice.

Researchers found that a plastic glass covering around the torso of shocked mice fully protected them from any axonal nerve cell damage in critical parts of the brain responsible for body movement, including the cerebellum and the corticospinal tract, which links nerves in the brain to those in the spinal cord. Body armor also shielded mice from over 80 percent of the axonal damage observed in the brain’s visual pathways when compared to mice wearing no body armor.