Subscribe to eTOC


Neuroscientists have learned a great deal about the molecular biology of brain tumors in terms of why the cells are dividing, why they are invasive, and why they have different neoplastic properties. The challenge of the next decade, say neuroscientists interviewed for this Special Report, will be to match what has been learned about tumor molecular biology with agents that have potential for treatment.

To this end, researchers are exploring diverse approaches – from identifying distinct genetic alterations that promote or inhibit growth in various tumor types – using gene assays and antisense technology – to developing “designer” therapies that specifically target the unique properties of the molecular alterations in those tumors.

Neuroscientists are particularly interested in studying a family of structurally-related growth factor receptors – tyrosine kinases – which play a central role in regulating cell growth, and are often overexpressed or mutated in human brain tumors.


Neurosurgeon Herbert H. Engelhard, MD, PhD, Associate Professor of Neurosurgery, Bioengineering, and Molecular Genetics, and Director of the Multidisciplinary Neuro-Oncology Program at the University of Illinois at Chicago, welcomes this approach.

By focusing more on which malignant cells respond to particular drugs, he said, researchers may be able to better categorize patients with a particular diagnosis, identify which tumors will respond to particular therapeutic strategies, and then tailor the treatment to the biology of each patient's tumor.


Dr. Engelhard said the most clinically important application of gene expression profiling in neuro-oncology involves oligodendroglioma, tumors that usually arise in the cells that produce myelin. Allelic losses – referred to as loss of heterozygosity – on chromosome arms 1p and 19q are the most frequent genetic alterations that characterize oligodendrogliomas from other types of gliomas. These genetic mutations are also independent predictors of response to chemotherapy (Int J Radiat Oncol Bio Phys 2000; 48:825–830). But to date there have not been similar discoveries for other types of primary brain tumors.


Dr. Herbert H. Engelhard: “Just as we identify which bacteria are causing someones meningitis and then pick a drug that is going to kill that bacteria, in the case of glioblastoma we should characterize on a molecular level exactly which genes that particular tumor is using to continue to proliferate and invade.”


Antisense technology: To create antisense drugs, nucleotides are linked together in short chains called oligonucleotides. Their principle mode of action is to bind selectively to a nucleic acid target (a specific messenger RNA) and block the synthesis of the encoded protein. Antisense drugs bind to mRNA targets at multiple points of interaction within a single receptor site.Antisense technology illustration courtesy of Isis Pharmaceuticals, Inc.


In other gene profiling studies, researchers at the Mayo Clinic and elsewhere are developing models of, and therapies against, epidermal growth factor receptors (EGFRs). Neuroscientists are interested in these proteins because they are found at abnormally high levels on the surface of many types of cancer cells, causing the cells to divide. By targeting EGFR, researchers believe they may be able to inhibit the growth of cancerous cells.

“EGFR is amplified or overexpressed in approximately one-third of glioblastomas,” said C. David James, PhD, Professor of Laboratory Medicine at the Mayo Medical School in Rochester, MN.

“If you have a reasonable model for EGFR inhibitors, you could determine whether there was a differential response dependent on the presence or absence of the gene alteration. Then, presumably, as many as one-third of glioblastoma patients would be good candidates for applying EGFR inhibitor therapy.”


Dr. James said the preferred model for analyzing EGRF inhibitors uses tumors derived from patient surgical specimens and grown in nude mice. He explained that it is difficult to study EGFR in cellular models alone, because when the tumors that have these gene amplifications are cultured, the amplification is lost in culture.

Dr. James's lab plans to test an EGFR-inhibitor called “Iressa” (also known as ZD 1839), on brain tumors. In animal and cell studies, Iressa has slowed the growth and shrunk tumors in lung and prostate cancer patients.


In other gene profiling studies, neuroscientists are exploring rapamycin – a peptide isolated in 1975 and used as an immunosuppressant drug to prevent organ transplant rejection reactions – as a potential treatment for glioblastoma.

Rapamycin inhibits cell growth by interfering with the function of a novel kinase termed mammalian target of rapamycin (mTOR) – inhibiting pathways that may be activated during malignant cell transformation and tumor progression (Cancer Metastasis Rev 2001:20(1–2):69–78).

Dr. James said that research published in 2001 suggests that glioblastomas and other tumors with alterations in the tumor suppressor gene PTEN (phosphatase and tensin homolog deleted on chromosome 10) appear to be sensitive to rapamycin, and to its synthetic pharmaceutical derivatives such as CCI779 (Proc Natl Acad Sci USA 2001; 98(18):10320–10325).


The PTEN tumor suppressor gene acts as a phospholipid phosphatase. (A) Under normal growth conditions, stimulatory signals from the insulin receptor activate the enzyme phosphoinositide kinase (PI3-kinase), which phosphorylates phosphatidylinositol 4,5-bisphosphate (PIP2) to generate phosphatidylinositol 3,4,5-trisphosphate (PIP3), a lipid signaling molecule. Downstream, PIP3 activates several effectors, including the proto-oncogene product PKB/Akt. The role of PTEN is to dephosphorylate PIP3, acting as a negative control on PKB/Akt activation. (B) If a mutation in PTEN renders it unable to carry out its phosphatase function, PIP3 can no longer be deactivated, and so continues to propagate its signal downstream. This may result in the continued activation of PKB/Akt, which, in combination with other factors, could lead to increased cell growth and possible tumor development.Image reproduced from the “Coffee Break,” National Center for Biotechnology Information,

In in vivo mice models, the loss of PTEN function in glioblastoma cells resulted in hyperactivity of AKT1 kinase, an enzyme that acts on a pathway that includes the molecule mTOR, Dr. James said. “The data suggest that tumors that don't have PTEN function will respond favorably to rapamycin.”


There are three classes of gene mutations implicated in cancer – oncogenes, tumor suppressor genes, and DNA repair genes.


Oncogenes cause a cells growth-signaling pathway to become hyperactive by producing abnormal versions or quantities of cellular growth-control proteins. A cancer cell may contain one or more oncogenes, which means that one or more components in this pathway will be abnormal.Source: National Cancer Institute


Tumor Suppressor Genes are a family of normal genes that instruct cells to produce proteins that restrain cell growth and division. Since tumor suppressor genes code for proteins that slow down cell growth and division, the loss of these proteins allows a cell to grow and divide in an uncontrolled fashion.Source: National Cancer Institute


DNA Repair Genes code for proteins whose normal function is to correct errors that arise when cells duplicate their DNA prior to cell division. Mutations in DNA repair genes can lead to a failure in DNA repair, which allows subsequent mutations in tumor suppressor genes to accumulate.Source: National Cancer Institute


In other efforts, Dr. Engelhard's work is focused on characterizing genes that are expressed in an individual patient's tumor, and then attempting to use antisense technology to block that gene expression.

To create antisense drugs, nucleotides are linked together in short chains called oligonucleotides. Their principle mode of action is to bind selectively to a nucleic acid target – a specific messenger RNA (mRNA) and thereby block the synthesis of a specific protein. Antisense drugs bind to mRNA targets at multiple points of interaction within a single receptor site.

“Just as we identify which bacteria are causing someone's meningitis and then pick a drug that is going to kill that bacteria, in the case of glioblastoma we should characterize on a molecular level exactly which genes that particular tumor is using to continue to proliferate and invade,” he said.

But this treatment strategy is complicated by the fact that malignant brain tumors are a heterogeneous lot, Dr. Engelhard explained. Not only do these tumors differ from one another, but individual areas within a single patient's tumor may vary in the molecular strategies they use and their response to therapy.


With that in mind, Dr. Engelhard and colleagues are also studying drugs that have multiple effects on different genes and that inhibit several pathways – in particular, histone deacetylase inhibitors such as phenyl-butyrate. Treatment of a cell with phenyl-butyrate results in hyperacetylation of histones, changing the pattern of gene expression in that cell, he said.

Histones are proteins that bind to DNA, packaging it into compact structures; they help roll several feet of DNA into the microscopic span of a single nucleus only six micrometers in diameter – to form chromosomes. The compact packaging of DNA must be relaxed somewhat for DNA replication and transcription to occur. The attachment of acetyl or methyl groups appears to affect the structure of histones.

“The process changes the accessibility to the genes, affecting their transcription,” Dr. Engelhard explained. “I think it is something that needs further study, too.”

Dr. Engelhard added: “We don't require that cells die, only that they revert to the more normal phenotype and so lose the ability to invade and proliferate,” he said. “In this approach, known as differentiation therapy, if a tumor is using several different pathways in different areas of a tumor to proliferate and invade, it still may be susceptible to the drug.”


Is there one particular signaling pathway that will effectively target all tumors? According to Dr. James, the theory has yet to be proven, despite research suggesting that there are invariant molecular features common in all cells for proliferation or growth, irrespective of cell or tumor type.

“The idea of a magic bullet that works in all cases is still out there,” Dr. James said. Nevertheless, Dr. James believes a “cocktail” approach to treatment is more likely to be successful treating brain tumors and other cancers, because of cellular heterogeneity within a cancer.

Dr. James said he believes future research will focus on combining molecular profiling of brain tumors and in vivo and animal laboratory testing to find relationships between the profile and a tumor's response to specific therapies.

But profile-versus-response testing cannot be done blindly, he said. Researchers must know what an agent is inhibiting, why that gene alteration has occurred, and how the alteration affects signaling pathways in order to select chemical inhibitors against the diagnostic subset of a tumor.


Hassan Fathallah-Shaykh, MD, agrees that a broader understanding of the genetics of brain tumors is important. Knowledge is quite limited when it comes to understanding the genes that play a role in the pathogenesis and progression of brain tumors, said Dr. Fathallah-Shaykh, MD, who is Associate Professor of Neurological Sciences at Rush-Presbyterian Medical Center in Chicago.

“What we've been doing is throwing rocks at these tumors, and I don't think we are going to hit the target unless we advance our understanding of the complete genetic picture.”

He said one problem, for example, with the identification of allelic losses on chromosome arms 1p and 19q – the genetic alterations that characterize oligodendrogliomas – is that no one knows what the genes are or what they do.

Dr. Fathallah-Shaykh would rather see researchers focus on understanding all the genetic changes occurring in tumors. He believes therapies aimed at individual genes or tumor mechanisms will not be effective in the long run because of the genetic heterogeneity of tumors and patients.

He contends, however, that the completion of the Human Genome Project offers potential for promising breakthroughs in research and treatment. “With a genome-wide view of the genetic changes in a tumor, discovering all the pathways and more importantly how they interact is possible; the new information may help us succeed in devising a treatment that works.”

Though there are promising developments, he and others interviewed here say that treatment of brain tumors remains an uphill battle.


Tumors are graded from I to IV; cells from higher-grade tumors are more abnormal looking and generally grow faster than cells from lower-grade tumors.

The most common primary brain tumors are gliomas, which begin in the glial tissue. There are several types of gliomas:

  • Astrocytomas arise from small, star-shaped cells called astrocytes. They may grow anywhere in the brain or spinal cord. In adults, astrocytomas most often arise in the cerebrum. In children, they occur in the brain stem, the cerebrum, and the cerebellum. A grade III astrocytoma is sometimes called anaplastic astrocytoma. A grade IV astrocytoma is usually called glioblastoma multiforme.
  • Ependymomas usually develop in the lining of the ventricles. They may also occur in the spinal cord. Although these tumors can develop at any age, they are most common in childhood and adolescence.
  • Oligodendrogliomas arise in the cells that produce myelin. These tumors usually arise in the cerebrum. They grow slowly and usually do not spread into surrounding brain tissue. Oligodendrogliomas are rare. They occur most often in middle-aged adults but have been found in people of all ages.

Source: National Cancer Institute


Investigators involved in brain tumor research are encouraged by molecular advances that have led to the development of therapies for other cancers. C. David James, PhD, Professor of Laboratory Medicine at the Mayo Medical School in Rochester, MN, said recent trials of trastuzumab (trade name, Herceptin) for breast cancer and imatinib mesylate (trade name, Gleevac) for chronic myeloid leukemia (CML), a cancer of white blood cells, support the viability of pursuing molecular targets for brain tumors.

Clinical trials show trastuzumab is effective in women with breast tumors that amplify the erbB-2 gene and overexpress the protein, Dr. James said. Some breast cancer cells can be stimulated to divide and grow when epidermal growth factor attaches itself to another protein, erbB-2, found on the surface of breast cancer cells. Trastuzumab blocks this action by attaching itself to the erbB-2 protein so that the epidermal growth factor cannot reach the breast cancer cells – thereby, stopping the cells from dividing and growing.

Dr. James noted that EGFR is highly amplified in about one-third of glioblastomas and is also a member of the erbB-2 family.

Imatinib mesylate is promising for treating CML patients who have bcr-abl translocations. Bcr-abl abnormal proteins are present in nearly all who have the disease, he added. By blocking the bcr-abl proteins, imatinib mesylate kills the leukemia cells. Dr. James said that other laboratories are testing imatinib mesylate in malignant gliomas.


  • Loss of heterozygosity (LOH) is the loss of a normal allele through somatic mutation or chromosomal rearrangement at a tumor suppressor locus that removes constraints on growth-promoting genes in the tissue.
  • Gene expression: In any type of tissue, many – but not all – genes are switched on at any given time. Also, from one tissue type to another, the limited set of genes involved will vary. Thus, each tissue can be identified by its unique pattern of gene expression. This pattern is characterized by microarray analysis and is called an “expression profile” or a “molecular signature.”
  • Translocation is a process in which a bit of genetic material from one chromosome swaps places with a bit from another chromosome. In CML, a piece (called “abl”) from chromosome 9 is translocated onto a segment called “bcr” on chromosome 22, creating the bcr-abl oncogene.
  • Tyrosine kinases are enzymes involved in communication within cells, or signaling pathways.
  • PTEN (phosphatase and tensin homolog deleted on chromosome 10) – a tumor-suppressor gene on chromosome 10q23 – is missing or mutated in a variety of human cancers, including glioblastoma. PTEN acts as a phosphatase – that is, it counteracts the action of kinases, enzymes that have a stimulatory effect in cell signaling pathways. PTEN is similar (homologous) to tensin, a protein that interacts with actin filaments at sites of intense signaling activity on the inner surface of cells known as focal adhesions.


  • “Coactivators and Corepressors in Gene Expression,” a report on the complex activators involved in gene expression, online at
  • Horn PJ, Peterson CL. Chromatin higher order folding: wrapping up transcription. Science 2002; 297:1824–1827.
  • Armandola, EA. Methylation and acetylation: manipulating phenotypes in cancer cells. Online at
  • Porgo BG, Allfrey VG, Mirsky AE. RNA synthesis and histone acetylation during the course of gene activation in lymphocytes. Proc Natl Acad Sci USA 1996;55:805–812.
  • Bello MJ, Leone PE, Vaquero J, Allelic loss at lp and 19q frequently occurs in association and may represent early oncogenic events in oligodendroglial tumors. Int J Cancer 1995;204:207–210.
  • Mendelsohn J, Baselga J. The EGF receptor family as targets for cancer therapy. Oncogene 2000;19:6550–6565.
  • Hildago M, Rowinsky EK. The rapamycin-sensitive signal tranduction pathway as a target for cancer therapy. Oncogene 2000;19:6680–6686.