The Continuing Evolution

In keeping with the theme of the Congress of Neurological Surgeons (CNS) annual meeting, this article briefly discusses how the evolution of our understanding of the biology of brain tumors has influenced the development of treatments. The first modern brain tumor operation was performed in London in 1884 by London surgeon Rickman Godlee on a 25-year-old patient who had suffered from focal motor epilepsy and progressive hemiparesis.1 This operation was made possible by 3 critical discoveries of that century: the introduction of anesthesia, with the first operation being undertaken under anesthesia on October 16, 1846, at what is now called the Ether dome at the Massachusetts General Hospital; the introduction of antiseptic techniques by Lord Lister in 1867, with his concept being based on the important work of Pasteur, Koch, Semmelweiss, and others; and the development of the hypothesis of cerebral localization of neurological function.1 An understanding of the pathology of brain tumors was essential for the evolution of intracranial surgery for tumors. The grinding of improved lenses by Amici in 1827 led directly to the development of the compound microscope that made possible the recognition of the cell as the basic unit of living matter. Shortly after, Schleiden and Schwann developed a cell theory.1 Rudolph Virchow, known as the “pope of medicine” during his time, was the first to describe neuroglia and to classify brain tumors with gliomas as a separate entity. By 1900, the initial enthusiasm of the pioneering operations had waned as a result of the generally poor results of tumor surgery. It was Harvey Cushing who introduced the methodical and meticulous techniques of Halsted to the neurosurgical operations and reignited the enthusiasm for brain tumor surgery.1 Cushing, together with Percival Bailey, studied the histological appearance of gliomas and classified them on a histogenetic basis.2 By the late 1970s, the median survival for these tumors after surgery, radiation, and chemotherapy was reported as being <1 year.3,4 Even with present gold standard treatment, >3 decades later, the median survival is just over 14 months.5 This relative lack of progress has led to numerous development efforts to optimize our present standard therapies. The place of surgery in the management of gliomas is still debated in this diffuse disease. Debulking the tumor is a rapid and effective method of relieving the features of raised intracranial pressure. Although we believe that cytoreduction has an oncological benefit, it is not feasible to do a randomized controlled study to absolutely prove the benefits of resection. Many techniques have been introduced to optimize surgery by maximizing the extent of resection and reducing the morbidity. These include using stereotaxy, awake surgery with physiological mapping, functional magnetic resonance imaging, metabolic imaging, tractography, intraoperative magnetic resonance imaging, and 5-aminolevulinic acid photodetection. These techniques have been especially successful in improving the safety of resections in eloquent regions and when used to help control seizures. Numerous publications have argued the oncological benefits of resection, although none of them are prospective, randomized controlled trials.6 Perhaps the best evidence comes from the report by Stummer et al,7 evaluating a subgroup of the randomized ALA-Glioma study, which minimizes selection bias, although this is still only Class 2b evidence. The same debate concerning the oncological value of resection continues, perhaps with even more intensity, for low-grade gliomas, with conflicting reports on the value of resection in prognosis,8 but again, the recent Cochrane collaboration report noted that there were no randomized controlled studies.9 Although surgery remains the initial and crucial component of treatment of gliomas generally, with increasing technology to assist surgery, we must be very careful that we do not make the same mistakes for the treatment of gliomas that were inflicted on women with breast cancer during the greater portion of the last century, ignoring the basic biology of the tumors and making the disease fit into an entirely surgical paradigm. Adjuvant therapies have been directed against both the local nature of these tumors and their diffuse spread through the brain. It is to be noted that although 80% of the tumors recur locally after our standard treatments, indicating a failure of local control,10 any treatments that attack the tumor entirely locally are doomed to failure as a result of the widespread invasion of the tumor into the brain. Although there have been considerably sophisticated advances in the delivery of radiation therapy, most have been aimed at focusing the treatment more accurately, ignoring the biology of the tumor in that it is a diffuse disease and any entirely local treatment is inevitably able to achieve only a very limited benefit. In general, conventional chemotherapy has been very disappointing despite an ever-increasing number of chemotherapy agents and methods of delivery, including those that involve maximal delivery to a wide area such as using blood-brain barrier breakdown,11,12 or more focused techniques such as implanted wafers impregnated with chemotherapy agents13 or microcatheters for convection-enhanced delivery.14 The concept of immunotherapy for brain tumors has evolved from the 1960s from active immunotherapy to adoptive immunotherapy and more recently to dendritic cell treatments.15 As noted in an editorial several years ago, it is truly a long and winding road that has yet to come to fruition, despite many promises.16 The current gold standard for treatment (the Stupp protocol) provides only a marginal increase in the median survival. What has been particularly interesting is the tail on the survival curve, which extends out to 5 years with 10% survival at this time compared with only 2% with radiotherapy alone.17 However, from a tumor biologist's point of view, what was particularly significant is related to the MGMT effect18 in that it recognizes for the first time that not all tumors are the same and that there are critically important differences in their genetic signatures, emphasizing the importance of the unique molecular properties of each tumor for the relative effectiveness of therapies. Of course, epigenetic activation of genes by promoter methylation has been recognized as an important mechanism by which tumor suppressor genes are shut down during the development of tumors. It represents one of the best-studied mechanisms of aberrant epigenetic modification in cancer, although in itself it emphasizes the complexities of tumor biology because silencing of repair genes such as MGMT sensitizes the tumors to some therapies whereas promoter methylation of the MMR gene confers resistance.19 When new therapies are being designed, it is impossible to separate the rational or scientific treatment of glioma from an understanding of the biology of the tumor. Of course, the evolution here has been understanding the molecular genetic composition of both the tumor cell and its environment, both of which are potential targets for therapies. We discuss here just some of the biological characteristics that have considerable implication for the development of therapies for these tumors. Arguably, the “new biology” began with the discovery by Stanley Cohen and Rita Levi-Montalcini of growth factors, for which they were jointly awarded the Nobel Prize in 1986. By 1988, growth factors were recognized as being important in the biology of brain tumors in an article by Manfred Westphal and colleagues20 and have remained so, being featured on the July 2012 cover of Cell Cancer. Over the years, it has been recognized that gliomas secrete numerous growth factors and cytokines, which both affect growth enhancement by various methods and have in themselves a powerful immunomodulatory effect on not only the patient’s immune status but also the immune competence of the tumor. The complexities of these growth factors, and especially their effect on cell pathways, are best exemplified by the numerous genes and proteins regulated by just transforming growth factor-β1 (Table 1).21TABLE 1: Proteins Regulated by Transforming Growth Factor-β1aFor a cell to become a cancer, there are 4 hallmarks of change: It must be able to undergo endless proliferation, evade apoptosis, induce angiogenesis, and migrate and metastasize.22 Each of these is dependent on ligand-receptor interactions on the cell surface, leading to a cascade of cytoplasmic events. Many of the important growth factors exert their numerous effects by activating receptor tyrosine kinases, which are transmembrane receptors. Although we are used to identifying these growth factors and their receptors as simple cartoons, they are extremely complex structures (Figure). Antony Burgess in Melbourne and Alex Levitzki at the Hebrew University in Jerusalem have spent a lifetime studying the structure of the endothelial growth factor (EGF) and its receptor, especially the mechanism by which the naturally occurring ligand will change the shape of the receptor dimer, thereby either allowing or preventing antibodies to bind and activate the receptor.23FIGURE: The extracellular domain of the epidermal growth factor (EGF) receptor (EGFR) dimer complexes to 2 EGF ligands (space-filled atoms). Domains I and II of the ribbon diagram of the EGFR-extracellular domain (ECD) are depicted in light orange, with the epitope for mAb806 (residues 273-293) shown in yellow toward the end of domain II. Cetuximab binds to the top surface of the EGFR ligand-binding domain III (dark blue). The 2 domain IVs of the EGFR-ECD dimer are shown in dark orange and red at the bottom of the model. The C-termini of these domains lead to the transmembrane domain of the EGFR (not shown).The Ras-mitogen-activated protein kinase pathway is the central mitogenic pathway upregulated in tumors. Ras is a membrane-bound G protein that initiates mitogen-activated protein kinase signal transduction. It is activated by a number of growth factors of the tyrosine kinase superfamily, and a number of tyrosine kinase inhibitors and monoclonal antibodies have already been used to target these pathways. There are a number of positive and negative signaling influences on these intracellular pathways. These are also potential targets for therapies. P53 mutations in human cancer were first described by Bert Vogelstein’s group24 in the early 1990s. P53 is a transcription factor that increases the transcription of P21 and P27, and although it is thought to be a possible important target for therapies, as noted in an editorial in 2010, “20 years later there is still much to learn” about its structure and function.25 Somatic mutations of PTEN genes occur in numerous tumor types and possibly in up to half of glioblastomas,26 as well as being especially important in prostate and endometrial tumors. PTEN is a tumor suppressor located on chromosome 10 that acts by removing phosphate groups from tyrosine moieties and the lipid PIP3, and it is particularly important in proliferation and apoptosis. It is important in controlling the P13 kinase-AKT pathway, which transmits antiapoptotic survival signals. This pathway is also important in treatment resistance, and its upregulation is associated with chemotherapy and radiotherapy resistance.27 So far, treatments aimed at targeting particular intracellular signaling pathways with tyrosine kinase inhibitors and monoclonal antibodies have had disappointing results.28 There seems to be considerable redundancy in the activities of the pathways, and we know that there is considerable cross-talk between the various pathways, making targeting of any of them particularly somewhat problematic in terms of producing a therapeutic effect. MicroRNAs are small noncoding RNA molecules (22 nucleotides long) that function in transcriptional and posttranscriptional regulation of gene expression.29 Discovered in 1993 by Victor Ambros and Rhonda Feinbaum, they have an important role in cancer, particularly in glioma.30 Interestingly, Tali Siegal’s group31 at the Hadassah Medical Organization showed that gliomas display a microRNA expression profile similar to that of the neuronal precursor cells. Of course, both the pathways and multiple mechanisms32,33 that subserve both tumor invasion and angiogenesis34 remain prominent targets for therapies, but the difficulty has been in translating the very effective laboratory results to humans. There are numerous small molecules and antibodies that have been tested in trials to regulate these pathways.28 As yet, none of these have been shown to be effective overall in prognosis, but perhaps we are ignoring the unique genetic profiles that may be displayed by different tumors, thereby choosing the wrong drug for a particular tumor. The future must be to match the genetic malfunction with the correct antidote. The genetic events in the process of tumor development from normal tissue are not as clearly understood for brain tumors as in the bowel for which, in the 1990s, Vogelstein’s group35 described the process of sequential chromosomal instability, which is important in the oncogenic evolution. However, some important genetic alterations have been identified (Table 2).36 EGF receptor amplification occurs in 40% of tumors; about half of these display the delta 2-7 mutation (EGFRvIII), which is constitutively activated and of most interest as a potential important target.37TABLE 2: Differences Between Primary and Secondary Glioblastoma Multiforme in Terms of Clinical and Genetic Factors36,aNext-generation sequencing, combined with the detailed knowledge of the clinical background, is the key to truly understanding the biology of these cancers in the future. In this, we can all be clinical scientists because we as neurosurgeons are in the unique position of being able to provide both the specimen and the clinical provenance. It is the marriage of the technology and the annotated specimen that will provide the revelation of the complex genomic landscape that underlies cancer.38 The results of the initial Cancer Genome Atlas Study39 confirmed the core pathways that are already known, underscoring the importance of the RTK and phosphoinositide 3-kinase pathways, MGMT, PTEN and P53, but also discovered previously unknown frequent mutations such as in IDH1. Transcriptional studies have revealed 4 major patterns of tumors: proneural, neural, classic, and mesenchymal, with the mesenchymal group having the poorest prognosis.40,41 These types of genetic studies, with comprehensive genomic characterization allied to clinical details, have already had considerable influence on our understanding of medulloblastoma and treatment options.42 A major limitation of the assessment of any new therapy is our present coarse evaluation techniques limited to radiological assessment. The development of serum markers for glioma would be a considerable advance. Tali Siegal's group43 at the Hebrew University in Jerusalem and the Hadassah Medical Organization have already shown that serum DNA markers can define the biological activity of the tumor. In a collaborative study, we are currently further evaluating this, particularly with microRNA. Other groups have shown the possibility of studying endothelial progenitors, and a group in our laboratory has shown that there is a differential in circulating endothelial cells responding to brain tumor surgery.44 The identification of a glioma stem cell by Peter Dirks’ group45 in Toronto was perhaps 1 explanation of why our present therapies fail and that these cells are much more resistant to treatments. Although there has been some debate regarding the role of these CD133-positive cells,46 like others, our present program has been involved with establishing a panel of glioma stem cell lines as models to test novel targeting drugs and to use large throughput sequencing to discover possible new gene targets. We have particularly been interested in looking at differentially expressed genes in the glioma stem cells and especially in investigating pathways that modulate stem cell expression. We have been able to study the effectiveness of novel agents using stem cell lines tagged with luciferase in an orthotopic xenograft murine model. We have shown that the glioma sphere cultures have high activation of the phosphoinositide 3-AKT pathway, which is similar to that previously known for glioblastoma, although it seems that stem cells retain the EGFRvIII mutation better than traditional cell lines. Focal adhesion kinase expression is upregulated in these cells, and inhibiting focal adhesion kinase prevents glioma sphere formation. Katherine Holland’s studies have shown that the conventional receptor tyrosine kinase pathways are upregulated in glioma stem cells and that developmental signaling genes such as the Salvador-Warts-Hippo47 and Wnt-β-catenin pathways.48 Yes-associated protein,49 the effector protein of the Salvador-Warts-Hippo pathway, is upregulated in glioma stem cells and is associated with worse survival. Currently, we are undertaking a project sequencing multiple transcriptomes of the glioma cell lines and looking for differential expression between the long- and short-term survivors in primary and recurrent tumors. Our view of the future is that these glioma stem cell models could provide new answers as to why brain cancers inevitably recur and could identify new pathways and molecules related to adhesion, invasion, and migration that are important. Targeting these pathways might provide a more rational approach to glioma therapy in the future. All of our work can be undertaken by only standing on the shoulders of giants, many of whom I have mentioned in this article. However, I particularly want to pay my respects to 2 personal mentors, the late John Curtis from Melbourne, a master surgeon and clinical scientist, and Chris Adams in Oxford. Colleagues, ladies, and gentlemen, I want to thank the CNS and its president, Ali Rezai, for the honor of the invitation to attend and present at the CNS annual meeting. Disclosure The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. Acknowledgments We would like to thank Professor Antony Burgess for providing the EGFR structure figure and Kate Lagerewskij for editorial assistance.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.