Search

Contains fulltext : 153542.pdf (Publisher’s version ) (Open Access) For nearly a century, the diagnosis and grading of oligodendrogliomas and oligoastrocytomas has been based on histopathology alone. Roughly 20 years ago, the first glioma-associated molecular signature was found with complete chromosome 1p and 19q codeletion being particularly common in histologically classic oligodendrogliomas. Subsequently, this codeletion appeared to not only carry diagnostic, but also prognostic and predictive information, the latter aspect only recently resolved after carefully constructed clinical trials with very long follow-up times. More recently described biomarkers, including the non-balanced translocation leading to 1p/19q codeletion, promoter hypermethylation of the MGMT gene, mutations of the IDH1 or IDH2 gene, and mutations of FUBP1 (on 1p) or CIC (on 19q), have greatly enhanced our understanding of oligodendroglioma biology, although their diagnostic, prognostic, and predictive roles are less clear. It has therefore been suggested that complete 1p/19q codeletion be required for the diagnosis of 'canonical oligodendroglioma'. This transition to an integrated morphological and molecular diagnosis may result in the disappearance of oligoastrocytoma as an entity, but brings new challenges as well. For instance it needs to be sorted out how (histopathological) criteria for grading of 'canonical oligodendrogliomas' should be adapted, how pediatric oligodendrogliomas (known to lack codeletions) should be defined, which platforms and cut-off levels should ideally be used for demonstration of particular molecular aberrations, and how the diagnosis of oligodendroglioma should be made in centers/countries where molecular diagnostics is not available. Meanwhile, smart integration of morphological and molecular information will lead to recognition of biologically much more uniform groups within the spectrum of diffuse gliomas and thereby facilitate tailored treatments for individual patients.

Astrocytoma and oligodendroglioma are histologically and genetically well-defined entities. The majority of astrocytomas harbor concurrent TP53 and ATRX mutations, while most oligodendrogliomas carry the 1p/19q co-deletion. Both entities share high frequencies of IDH mutations. In contrast, oligoastrocytomas (OA) appear less clearly defined and, therefore, there is an ongoing debate whether these tumors indeed constitute an entity or whether they represent a mixed bag containing both astrocytomas and oligodendrogliomas. We investigated 43 OA diagnosed in different institutions employing histology, immunohistochemistry and in situ hybridization addressing surrogates for the molecular genetic markers IDH1R132H, TP53, ATRX and 1p/19q loss. In all but one OA the combination of nuclear p53 accumulation and ATRX loss was mutually exclusive with 1p/19q co-deletion. In 31/43 OA, only alterations typical for oligodendroglioma were observed, while in 11/43 OA, only indicators for mutations typical for astrocytomas were detected. A single case exhibited a distinct pattern, nuclear expression of p53, ATRX loss, IDH1 mutation and partial 1p/19q loss. However, this was the only patient undergoing radiotherapy prior to surgery, possibly contributing to the acquisition of this uncommon combination. In OA with oligodendroglioma typical alterations, the portions corresponding to astrocytic part were determined as reactive, while in OA with astrocytoma typical alterations the portions corresponding to oligodendroglial differentiation were neoplastic. These data provide strong evidence against the existence of an independent OA entity.

Isocitrate dehydrogenase 1 (IDH1) mutations are frequent in astrocytomas, oligoastrocytomas and oligodendrogliomas. We previously reported the generation of a mutation-specific antibody that specifically detects R132H mutated IDH1 protein (clone H09). Here, we investigate the feasibility of H09 immunohistochemistry to differentiate between oligodendrogliomas/oligoastrocytomas and other tumors with similar morphology. A total of 274 brain tumors presenting with focal or extensive clear cell morphology were investigated. High numbers of H09-positive cases were observed in adult grade II oligodendrogliomas (67 of 74, 91%), grade III oligodendrogliomas (65 of 69, 94%), grade II oligoastrocytomas (11 of 14, 79%) and grade III oligoastrocytomas (10 of 11, 91%). All cases of pediatric oligodendrogliomas (n = 7), neurocytomas (n = 41, 35 central, 4 extraventricular, 2 cerebellar liponeurocytomas), dysembryoplastic neuroepithelial tumors (n = 21), clear cell ependymomas (n = 8), clear cell meningiomas (n = 9) as well as 12 primary glioblastomas with oligodendroglial differentiation and 5 pilocytic astrocytomas with oligodendroglial-like differentiation were negative for H09 immunohistochemistry. Three oligodendrogliomas with neurocytic differentiation had evidence of IDH1/IDH2 mutations either by H09 immunohistochemistry or direct sequencing. We conclude that in tumors with an oligodendroglioma-like morphology, binding of H09 is highly specific for oligodendrogliomas or oligoastrocytomas and substantially helps in the discrimination from other clear cell tumors. Negative H09 immunohistochemistry of an adult oligodendroglioma or oligoastrocytoma should prompt the consideration of other clear cell neoplasms. Further, our observations firmly assign oligodendrogliomas with neurocytic differentiation to the group of oligodendrogliomas and demonstrate that H09 is especially helpful for the difficult discrimination of such lesions from extraventricular neurocytomas.

Although oligodendroglial neoplasms are traditionally considered purely glial, increasing evidence suggests that they are capable of neuronal or neurocytic differentiation. Nevertheless, ganglioglioma-like foci (GGLF) have not been previously described. Herein, we report seven examples where the primary differential diagnosis was a ganglioglioma with an oligodendroglial component. These five male and two female patients ranged in age from 29 to 63 (median 44) years at initial presentation and neuroimaging features were those of diffuse gliomas in general. At presentation, the glial component was oligodendroglioma in six and oligoastrocytoma in one; one was low-grade and six were anaplastic. A sharp demarcation from adjacent GGLF was common, although some intermingling was always present. The GGLF included enlarged dysmorphic and occasionally binucleate ganglion cells, Nissl substance, expression of neuronal antigens, GFAP-positive astrocytic elements, and low Ki-67 labeling indices. In contrast to classic ganglioglioma, however, cases lacked eosinophilic granular bodies and CD34-positive tumor cells. Scattered bizarre astrocytes were also common and one case had focal neurocytic differentiation. By FISH analysis, five cases showed 1p/19q codeletion. In the four cases with deletions and ample dysmorphic ganglion cells for analysis, the deletions were found in both components. At last follow-up, two patients suffered recurrences, one developed radiation necrosis mimicking recurrence, and one died of disease 7.5 years after initial surgery. We conclude that GGLF represents yet another form of neuronal differentiation in oligodendroglial neoplasms. Recognition of this pattern will prevent a misdiagnosis of ganglioglioma with its potential for under-treatment.

Doublecortin (DCX) is required for neuroblastic migration during the development of the cerebral cortex. DCX is a microtubule-associated protein that plays a role in cellular motility. These facts led us to hypothesize that DCX is increased in invasive brain tumors. DCX expression was assessed in 69 paraffin-embedded brain tumors of neuroepithelial origin. In addition, mouse brain sections of the subventricular zone and dentate gyrus were used as positive controls for immunostaining, and specificity of antibody staining was demonstrated by peptide neutralization. DCX was highly expressed in both high-grade invasive tumors (glioblastoma, n=11; anaplastic astrocytoma/oligoastrocytoma, n=7; and medulloblastoma/PNET, n=6) and low-grade invasive tumors (oligodendroglioma, n=3; and astrocytoma/oligoastrocytoma, n=5). However, DCX was less intensely expressed in the circumscribed group of tumors (pilocytic astrocytoma, n=6; ependymoma/subependymoma, n=7; dysembryoplastic neuroepithelial tumor, n=4; ganglioglioma, n=2; meningioma, n=9; and schwannoma, n=9). By the Cochran-Mantel-Haenszel statistical test, the circumscribed group was significantly different from both the high-grade invasive group (P

Pediatric oligodendrogliomas are rare neoplasms and have not been characterized extensively either pathologically or genetically. Given the recent interest in the significance of chromosomal losses in predicting the clinical course and in establishing uniform diagnoses of adult oligodendrogliomas, we reviewed the pathological and clinical features of a series of pediatric oligodendrogliomas and determined their 1p and 19q status using fluorescence in situ hybridization. Of 19 tumors originally diagnosed as oligodendroglioma, 7 were oligodendroglioma, 3 were anaplastic oligodendroglioma, 3 were oligoastrocytoma, and 6 were reclassified. Only one tumor, an anaplastic oligodendroglioma, had 1p loss; none had 19q loss. The single patient whose tumor had 1p loss did not have a particularly favorable clinical course. These results suggest that pediatric oligodendrogliomas arise by molecular alterations distinct from adult oligodendrogliomas, and such molecular alterations do not hold immediate promise as an adjunct to the diagnosis of pediatric oligodendrogliomas.

The blood-brain barrier (BBB), mediated by endothelial tight junctions, is defective in malignant gliomas such as glioblastoma, resulting in cerebral edema and contrast enhancement upon neuroradiological examination. The mechanisms underlying BBB breakdown are essentially unknown. Since non-neoplastic astrocytes are required to induce BBB features of cerebral endothelial cells, it is conceivable that malignant astrocytes have lost this ability due to dedifferentiation. Alternatively, glioma cells might actively degrade previously intact BBB tight junctions. To examine the latter hypothesis, we have employed a transepithelial electrical resistance breakdown assay using monolayers of the C7 subclone of Madin-Darby canine kidney (MDCK-C7) cells forming tight junctions similar to those of BBB endothelial cells. We found that glioblastoma primary cells co-cultured with the MDCK-C7 monolayer (without direct contact of the two cell types) resulted in marked breakdown of electrical resistance, whereas primary cultures derived from low-grade gliomas (fibrillary astrocytoma, oligoastrocytoma) showed delayed or no effects. These results suggest that malignant gliomas have acquired the ability to actively degrade tight junctions by secreting soluble factors, eventually leading to BBB disruption within invaded brain tissue.

IDH1 R132H mutations occur in approximately 70% of astrocytomas and oligodendroglial tumors. We developed a mouse monoclonal antibody targeting the IDH1 R132H mutation. Here, we show the high specificity and sensitivity of this antibody on Western blots and tissue sections from formalin fixed paraffin embedded tumor specimens. This antibody is highly useful for tumor classification, in detecting single infiltrating tumor cells and for the characterization of the cellular role of mutant IDH1 protein.

In normal rat and human brain tubulin, polymerization-promoting protein TPPP/p25 is predominantly localized in oligodendrocytes (Kovacs et al. in Neurobiol Dis 17:155–162, 2004; Lindersson et al. in J Biol Chem 280:5703–5715, 2005; Skjoerringe et al. in J Neurochem 99:333–342, 2006). TPPP/p25 promotes the polymerization of tubulin into double walled tubules and polymorphic aggregates and inhibits mitotic spindle assembly in vitro; furthermore, it is a common marker of -synuclein bearing pathological inclusions (summarized in detail: Kovacs et al., Acta Neuropath, this issue). Oligodendroglioma, oligoastrocytoma, clear cell ependymoma, and central neurocytoma share a common feature: the clear cell histopathology (Kleihues and Cavenee, IARC Press, Lyon, 2000). Therefore, diVerential diagnosis of clear cell primary brain tumours in the diagnostic setting often requires immunohistochemistry (Koperek et al. in Acta Neuropathol 108:24–30, 2004; Preusser et al. in Histopathology, in press). In this study, we evaluated the reliability of TPPP/p25 as an immunohistochemical marker of oligodendroglial neoplasms in the classiWcation and neuropathological diagnosis of primary clear cell brain tumours. Our cohort of human clear cell CNS neoplasms comprised 57 oligodendroglial neoplasms with chromosome 1p aberrations (1p deletion or imbalance status in 20 oligodendrogliomas, 20 anaplastic oligodendrogliomas, 17 oligoastrocytomas; 51/57 of these tumours additionally had aberrations of chromosome 19q), 3 oligodendroglial neoplasms without chromosome 1p aberration (2 oligodendrogliomas, 1 oligoastrocytoma), 10 central neurocytomas, 10 clear cell ependymomas, and 8 dysembryoplastic neuroepithelial tumours (DNTs). Further, we investigated 20 diVuse astrocytomas and glioblastomas, as well as 5 glioblastomas with clear cell components. Tumour typing was performed according to WHO criteria (Kleihues and Cavenee, IARC Press, Lyon, 2000). Aberrations of chromosome 1p and 19q were assessed by Xuorescence in situ hybridization (FISH) using the paracentromeric probe D1Z1 (1q12) and the subtelomeric probe D1Z2 (1p36.3; both Q-BIOgene, Heidelberg, Germany) for 1p and probes 19pter and 19q13.3 (both Vysis, BergishGladbach, Germany) for 19q as described in a previous study by our group (Gelpi et al. in Mod Pathol 16:708– 715, 2003). We also immunostained ten biopsy samples of non-neoplastic temporal cortex from patients diagnosed with chronic temporal lobe epilepsy. For immunohistochemistry, tissue specimens were Wxed in a phosphate-buVered 4.5% solution of formaldehyde, embedded in paraYn and cut at a thickness of 3–5 m. Sections were deparaVinized in xylene. Slides underwent heat induced epitope retrieval in 0.01 M citrate buVer (pH 6.0) in a steamer for 10 min. For antiTPPP/p25 immunostaining, we used two rat polyclonal M. Preusser · H. Budka · G. G. Kovacs Institute of Neurology, Medical University of Vienna, Vienna, Austria

The localization of cystatin C (CC) and transthyretin (TTR) synthesis was studied using Northern blot and immunohistochemical methods. Normal brain tissues from all sites studied contained CC mRNA. Immunoreactive CC was present in the choroid plexus epithelial cells, cerebral and cerebellar neurons, astrocytes, ependymal cells, macrophage-like cells of the arachnoid membrane and in neuroendocrine cells of the anterior pituitary lobe. TTR mRNA and TTR were restricted to the choroid plexus. In primary brain tumors, the transcript for CC was found in all 39 tumors examined, while the protein could only be demonstrated in 3/5 choroid plexus papillomas, 8/8 astrocytomas, 7/23 anaplastic astrocytomas and glioblastomas, 1/6 oligodendrogliomas, 1/1 oligoastrocytoma, 1/4 anaplastic oligodendrogliomas, 3/7 ependymomas, 0/1 anaplastic ependymoma, 0/5 primitive neuroectodermal tumors, 0/1 neuroblastoma, 3/11 meningiomas and 16/16 pituitary adenomas. CC cannot be used as a marker for any specific brain tumor type but the fact that the protein could be demonstrated more frequently in astrocytomas than in their more malignant counterparts suggests that the cellular production and secretion of CC changes with the malignant progression of these tumors. TTR mRNA and TTR were present only in the choroid plexus papillomas, indicating that TTR synthesis is mainly restricted to such brain neoplasms.

CIC and FUBP1 mutations have recently been detected in oligodendrogliomas but not in oligoastrocytomas. However, allelic losses in the regions on chromosomal arms 19q and 1p harboring CIC and FUBP1 are a common feature of both, oligodendrogliomas and oligoastrocytomas. To resolve this discrepancy, we analyzed CIC and FUBP1 mutations in a set of primary brain tumors including 18 oligodendrogliomas and 42 oligoastrocytomas. In addition, we analyzed 10 astrocytomas and 16 glioblastomas with allelic losses on 19q as well as a set of 12 medulloblastomas for CIC mutations. CIC mutations were found in 15/18 oligodendrogliomas, 14/42 oligoastrocytomas and 3/10 preselected astrocytomas. With the exception of a single case, all CIC mutations occurred in tumors with combined 1p/19q losses. In contrast to oligodendrogliomas where CIC mutations were always detected along with 1p/19q co-deletion, CIC mutations were only found in 52 % of the 1p/19q co-deleted oligoastrocytomas. FUBP1 mutations were detected in 7/61 tumors, all presenting with CIC mutations. FUBP1 mutations appear to cluster in the DNA binding domain spanning exons 5–14. CIC and FUBP1 mutations exclusively occurred in presence of either IDH1 or IDH2 mutations. Our data confirm CIC and FUBP1 mutations in oligodendrogliomas and demonstrate the presence of these mutations in oligoastrocytomas.

Oligoastrocytomas are glial tumours consisting of a mixture of neoplastic astrocytic and oligodendroglial cells. Genetic alterations of oligoastrocytomas include loss of heterozygosity of chromosomes 1p and/or 19q (LOH 1p/19q), typically occurring in oligodendrogliomas, and mutations of TP53, frequently occurring in astrocytomas. To investigate whether these neoplastic cell types in oligoastrocytomas have different genetic profiles, we examined the two different components of oligoastrocytomas in comparison with the histological diagnosis of the specific tumour area for LOH 1p/19q and TP53 mutations by using microdissection technique. We found a variety of lost markers for 1p and 19q, and the presence of two different TP53 mutations in the tumour samples. In the majority of cases (9/11), the oligodendroglial and astrocytic components of an individual oligoastrocytoma displayed the same genotype. We present two cases of biphasic oligoastrocytomas with aberrant findings, suggesting the coexistence of genetically and morphologically distinct tumour cell clones in these tumours. In one case, the oligodendroglial part of the tumour showed LOH19q, whereas the astrocytic part showed TP53 mutation (codon 273). In another case, we found LOH 1p/19q in the oligodendroglial component, but two retained areas on chromosome 1p in the astrocytic component of the tumour. No evidence was found for the coexistence of tumour cells with the two genotypical changes within the same morphological region of one individual tumour. The two cases of biphasic oligoastrocytomas in our sample that display a different genotype in the astrocytic and oligodendroglial part of the tumour show that different components of an oligoastrocytoma may be derived from different cell clones during neoplastic transformation.

The phrase “Let us now praise famous men, and our fathers that begat us” is from the second-century B.C.E. text entitled the Wisdom of Sirach or Ecclesiasticus, a book of ethics that played a role in early iterations of the Bible. The phrase was used more recently and most famously in an ironic context, by James Agee and Walker Evans in the title of their classic work of photojournalism, Let Us Now Praise Famous Men. This photoessay catalogued portraits of everyday sharecroppers in the southern United States in the 1930s—people who were the very least of “famous men.” The book poignantly drew attention to the problems faced by these impoverished sharecroppers, and drew praise over many decades for its innovative approach to journalism, particularly because of its striking use of the relatively new discipline of photography. Over the course of the twentieth century, the province of brain tumor classification (and tumor classification in general) was for the most part governed by famous men—individuals who penned influential texts and/or had been blessed with both outgoing personalities and strong convictions. (Of note, although the majority of the great brain tumor pathologists of the twentieth century were men, there were great women as well, perhaps most notably Dorothy Russell.) The aura of such individuals attracted consults, leading to greater experience, and in turn to greater abilities to present opinions as certainties. This experiential system led to major advances in the field and to our current classification systems. But, as the old aphorism states, “Good judgment is the result of experience; experience is the result of bad judgment.” In other words, the system that we currently use has arisen from many trials and errors, and from a good aliquot of subjectivity infused with the strength of convictions of our famous men. For the first time in history, however, we now see the inklings of a system that offers far more potential for objectivity, and hence less dependence on the vagaries of individual strong convictions. This early molecular approach now begs the question of whether we will see a shift from classification based on famous men to one based on famous molecules. Most neuropathologists can already name the famous molecules, genes and chromosomes involved in brain tumor classification: p53, 1p, 19q, IDH1, MGMT, EGFR, INI1, and so on. Newly implicated ones seem to appear at nearly every turn: witness the recent emergence of H3.3 and ATRX mutations. Clearly these molecules are not the complete cast of characters, but they for the first time constitute a sizeable quorum. And while not all neuropathologists currently have access to assays to assess the status of these molecules, the field is changing so rapidly that the next ten years should see such techniques widely—and perhaps universally—available. But the questions remain: if we can determine the status of these molecules, are we ready to use them? And if we do, what will happen to the old, tried and true classifications? I believe that we have reached the tipping point, the point at which we know enough about the significance of molecular events that we can begin to use them for classification and for transitioning away from the old systems. Nonetheless, such a transition cannot be abrupt and cannot put at risk the decades of clinicopathological correlation upon which our current approaches to therapy are based. Suggestions to change schemes radically in the past (e.g., the PNET vs medulloblastoma dichotomy) led to heated debates and stimulated further work in the field, but the strengths of the various convictions often dwarfed the facts regarding one side or the other. Given such history, I therefore recommend that the next brain tumor classification not replace the old, but rather be a transitional approach to classification that would include both the old and the proposed new systems. To do so, I suggest a system that incorporates three, rather than the current two, elements of classification and grading; in addition to histopathological name and WHO grade, one would add the molecular pathological name, preferably at the top of the diagnosis. Such a transitional system would allow the current clinically important clinicopathological system to remain in place while setting the future molecular pathological system into use. The molecular pathological name would be listed first and would include a generic histological designation followed by the molecular details. This generic histological designation would define the major categories of tumor, but would not be mired in the often subjective distinctions within histologies, e.g., astrocytoma vs oligoastrocytoma vs oligodendroglioma. The mutational details would follow a specific convention and would need to include a particular set of molecular alterations, which would need to be determined alongside of the next WHO classification modifications. The histopathological name would be next and would follow the WHO convention, presumably using the terms and criteria as updated in the next iteration of the WHO classification. Lastly, the WHO grade would follow, with the grade potentially being dependent on the molecular alteration. For example, a glioblastoma or an anaplastic astrocytoma with an IDH mutation might be WHO grade III, whereas a glioblastoma or an anaplastic astrocytoma without an IDH mutation would be WHO grade IV. Examples of such a transitional system would be: Diffuse glioma; IDH1 mutant, TP53 mutant, EGFR normal copy number Anaplastic astrocytoma WHO grade III Diffuse glioma; IDH wildtype, TP53 wildtype, EGFR amplified Anaplastic astrocytoma WHO grade IV Embryonal tumor; INI1 null Atypical teratoid/rhabdoid tumor WHO grade IV A system like this would allow flexibility in a number of ways. One, it would allow a listing of molecular alterations that could be used for diagnostic, prognostic or therapeutic purposes. Two, it would allow the rapid assimilation of the type of molecular change, e.g., the difference between BRAF duplication (“BRAF dup”) versus a BRAF mutation (“BRAF mut”), or between different IDH1 mutations (“IDH1 mut R132H” vs “IDH1 mut R132C”). Three, it would give the experienced pathologist the ability to override the molecular diagnosis with a clinicopathological diagnosis if he/she felt this was a better descriptor—or if the molecular diagnosis was not sufficiently specific to guide therapy. Four, such a change would align brain tumor grading systems with those of other tumors, allowing a single histological tumor type to have different grades rather than grade simply being linked to the tumor type (i.e., currently all anaplastic astrocytomas are grade III and all glioblastomas are grade IV, with no additional information being conveyed by the grade assignment). Lastly, it could decrease the confusion caused by subjective “gray area” diagnoses. For example, some neuropathologists opine that a glioma becomes a glioneuronal tumor in the presence of sometimes minor positivity for a neuronal marker (e.g., synaptophysin); instead, knowing objectively that such a lesion was IDH mutant would allow one to conclude that it would behave like a diffuse glioma (e.g., glioneuronal tumor with neuropil-like islands or a diffuse glioma invading cortex that was misdiagnosed as ganglioglioma). I envision that this system itself will be a transitional one. As we learn more about multiple genetic and epigenetic events converging on pathways that can be specifically targeted for therapy, it is likely that the mutational details described above would be replaced by pathway activation or pathway inactivation designations. In many places, we are already speaking of “Wnt pathway” medulloblastomas in this way. Moreover, while the system would be initially employed in the grading and classification of diffuse gliomas and embryonal tumors, I anticipate that it would be extended fairly soon to tumors such as ependymomas and meningiomas, as additional clinical-molecular correlations are made in these entities. We stand at a critical time in the evolution of diagnostic tumor neuropathology, with new objective techniques coming alongside comprehensive “–omic” analyses of tumors. The era of the famous men has been a vitally important part of our history; indeed, the famous men have been our mentors. But, as the phrase “Let us now praise famous men” states, the time has now come to take a big step forward and to allow the famous molecules (rather than more famous men) to be our legacy.

In 80 specimens of human glioma the production of glial fibrillary acidic protein (GFAP) by tumour cells invading meninges or connective tissue was studied immuno-cytochemically by the PAP technique. In 38 of 55 cases of astrocytoma, glioblastoma, gliosarcoma, and oligoastrocytoma, GFAP immunoreactivity was greater in the invading cells as compared with the main part of the neoplasm. Fifty-eight percent of the astroglial tumours invading the leptomeninges, all astroglial tumours invading connective tissue and all gliosarcomas showed enhanced GFAP immuno-reactivity of tumour cells getting in contact with collagenous tissue, whereas meningeal infiltrates of 25 non-astroglial tumours (oligodendroglioma, ependymoma, medulloblastoma) remained GFAP-negative like the main part of the respective tumours. In the majority of astroglial tumours an increase of GFAP immunoreactivity was found also in perivascular cells of the main part of the tumour. It is concluded that glioma cells are capable of adapting their cytoskeleton to their micro-environment. Contact with dense collagenous tissue appears as an important factor able to induce an increased production of GFAP by adjacent glial cells.

The glioneuronal tumor with neuropil-like islands (GTNI) is an uncommon histopathological entity composed of diffuse astrocytic or oligoastrocytic elements interspersed with sharply delimited, micronodular loci of neuronal differentiation. Despite its initial description over 10 years ago [10], the lineage classification of GTNI remains speculative. The clinical features of GTNI, its infiltrative growth characteristics, typically dominating astroglial components, and frequent p53 expression, are all consistent with a fundamental linkage to the diffuse astrocytic tumor group. Supporting this notion are reports documenting, by comparative genomic hybridization, an astrocytoma-like profile of 7q21.1-qter gain and 9p21-pter loss in one case [8] and the absence of oligodendroglioma-associated 1p/19q deletions in a series assessed by polymerase chain reaction-based loss of heterozygosity analysis [3]. Point mutations in the isocitrate dehydrogenase enzymes, IDH1 and IDH2, are found in the majority of WHO grade II and III diffuse gliomas—both astrocytic and oligodendroglial—as well as secondary glioblastomas (GBM), but they are almost entirely absent from primary GBM [7, 9, 11]. As such, they appear to represent a defining early genomic lesion in the pathogenesis of diffuse glioma variants exhibiting step-wise malignant evolution. Glioma-associated IDH mutations invariably involve active site arginine residues—R132 and R172 for IDH1 and IDH2, respectively—and, while their precise pathogenic mechanism remains unclear, they appear to promote the generation of the cancer-predisposing toxic metabolite, R(-)-2-hydroxyglutarate [6]. We assembled 12 cases of GTNI (Table 1) seen at the Memorial Hospital Department of Pathology, along with pertinent clinical information from the patients’ medical records. To evaluate whether IDH1 and IDH2 mutations were present in our sample set, we performed custom genotyping assays on the Sequenom platform [1, 2] and/or employed validated immunohistochemical staining (IHC) for the most common IDH mutation, IDH1 R132H [5]. We found that every tumor in our sample set (12/12) exhibited evidence of the IDH1 R132H mutation by either one or both of our assay methods (Fig. 1a–d), and concordant results were achieved for all cases screened by both Sequenom® and IHC (Table 1). Both glial and neuronal components exhibited immuno-expression of mutant IDH1 in all examined cases (Fig. 1d). Table 1 Cases included in GTNI study cohort listed with patient age at first diagnosis, sex, histopathological grade (Grade), Sequenom® genotyping result (Sequenom) and IHC for IDH1 R132H (R132H IHC) Fig. 1 GTNI is characterized by IDH1 R132H mutations. a A mutation-positive Sequenom® spectrum from a representative GTNI sample displays an additional extension product of different mass, due to the G to A point mutation. Photomicrographs of a representative ... The primary significance of our findings lies in strengthening the pathogenic link between GTNIs and the diffuse glioma tumor group. Studies to date indicate that within primary brain tumors, IDH mutations are largely restricted to diffuse gliomas evolving via step-wise malignant evolution (i.e., diffuse astrocytomas, oligodendrogliomas, and oligoastrocytomas). Combined with the frequent p53 immunoreactivity and general absence of 1p/19q chromosomal co-deletions previously documented in this setting, a high incidence of IDH1 mutations would fundamentally ally GTNIs with the diffuse astrocytoma and 1p/19q-intact oligoastrocytoma group. Indeed, the 100% incidence of IDH mutation in our sample set is even higher than that documented in diffuse astrocytoma so far (~80%) [7, 11]. As an aside, recent work has also identified IDH mutation as a prevalent feature of oligodendrogliomas with neurocytic differentiation [4], further emphasizing that focally prominent neuronal morphology falls well within the histopathological spectrum of diffuse gliomas in general. The invariable presence of the specific IDH1 R132H mutation in our GTNIs may represent an artifact of small sample number, but, nevertheless, it is striking. Whether other IDH1 or IDH2 mutational variants exist in GTNI and whether a genotypic restriction to IDH1 R132H alone would have any functional significance remain unclear.

Immuno-reactivity to glial fibrillary acidic protein (GFAP) is mainly regarded as a sign of astroglial histogenesis and/or differentiation. The presence of astrocytes in oligodendrogliomas is a well known phenomenon; in addition, GFAP-positive neoplastic oligodendrocytes have also been described but have not yet been studied systematically. Using an anti-GFAP serum in the peroxidase-antiperoxidase antiperoxidase (PAP) technique, 50 oligodendrogliomas and 16 mixed oligodendrocytomas were investigated; they had been diagnosed by routine histological stains. In half of all oligodendrogliomas, and only in a few (12%) of the mixed oligoastrocytomas, GFAP-positive oligodendrocytes were found in some areas of the classical honey-comb texture with a prominent vascular stroma. The term 'gliofibrillary oligodendrocyte' (gfoc) is proposed for these immuno-reactive cells. The existence of a tumour cell combining morphological characteristics of oligodendroglia with GFAP production in its cytoplasm may be considered analogous to transient GFAP expression by myelin-forming glia during normal development (Choi and Kim 1984), thus suggesting the return to a foetal behaviour by some neoplastic oligodendrocytes. Three tumours of the present series consisted largely of gfocs and, therefore, may be termed gliofibrillary oligodendrogliomas. In about 32% of all oligodendrogliomas, but only once in the mixed tumour group, a gradual morphological transition from gfocs to gemistocytic astrocytes was observed.(ABSTRACT TRUNCATED AT 250 WORDS)