Info

Failures

Survivors tumor expression profiling: glioma

Failures

Survivors

J02611 APOD Apolipoprotein D

D86974 KIAA0220 gene

L06419 PLOD Lysyl hydroxylase

U37673 Beta-NAP

Markers of Treatment Failure

X69150 Ribosomal protein S18

X13293 MYBL2 V-myb

M36072 RPL7A Ribosomal protein L7a

U14972 Ribosomal protein S10

a = standard deviation from mean figure 5.4 Signal to noise rankings of genes comparing tumors from surviving patients to those that experienced treatment failure. (J Clin Oncol, Fernandez-Teijeiro et al. (2004) Copyright American Society of Clinical Oncology. Reproduced with permission.) See Plate 5.4 in Color Plate Section.

Markers of Survival

J02611 APOD Apolipoprotein D

D86974 KIAA0220 gene

L06419 PLOD Lysyl hydroxylase

U37673 Beta-NAP

Markers of Treatment Failure

X69150 Ribosomal protein S18

X13293 MYBL2 V-myb

M36072 RPL7A Ribosomal protein L7a

U14972 Ribosomal protein S10

a = standard deviation from mean figure 5.4 Signal to noise rankings of genes comparing tumors from surviving patients to those that experienced treatment failure. (J Clin Oncol, Fernandez-Teijeiro et al. (2004) Copyright American Society of Clinical Oncology. Reproduced with permission.) See Plate 5.4 in Color Plate Section.

There is also new evidence that overexpression of ERBB2, which encodes for class I receptor tyrosine kinases, predicts poor survival [62]. Many of these studies, however, were performed with small sample sizes and still need to be validated in large-scale institutional studies. For these reasons, it is now clear that all future diagnostic tools for risk stratification must incorporate clinical features and histologic subtype as well as the expression of specific genes and proteins.

Microarray analysis can also be used to identify gene expression profiles that indicate an activated regulatory pathway or multiple interacting processes that lead to a known cellular response. Medulloblas-tomas are known to express specific markers which indicate that they usually arise from granule cell progenitors [47,63-65]. For example, mutations of the Sonic hedgehog (Shh) receptor patched (PTCH) have been linked both to proliferation of granule cell progenitors in the developing cerebellum and to medulloblastoma oncogenesis [66-68]. Mice heterozygous for mutation of Ptc develop medulloblastomas in a multi-step process, first manifested as persistent proliferation of granule cell progenitors that then progress to overt tumor growth [54,69,70]. Children heterozygous for germline PTCH mutations (Gorlin syndrome) account for less than 1 per cent of medulloblastomas, but mutations of PTCH and related molecules are linked by gene expression profiling to the desmoplastic subclass of medulloblas-tomas (approximately 25 per cent of tumors) [42].

Mutations of PTCH and other molecular signaling pathways provide insight into the molecular patho-genesis of medulloblastomas as well as targets for the development of small molecule inhibitors. In a mouse model of spontaneous medulloblastoma, small molecule inhibitors of the Shh pathway eliminated tumor and reduced cell proliferation [71]. The mechanism of action is thought to be through downstream target inhibition and could be monitored by measuring dose-dependent gene downregulation resulting from inhibition of downstream effectors.

For several childhood cancers, such as neuroblas-toma and leukemia, therapy is adapted according to a combined assessment of disease status and molecular profile [72,73]. In medulloblastomas, genomic methods that look beyond the expression and mutation of single genes have begun to reveal a biological subclassification based on gene-expression profiles that define invasive growth, response to therapy, and linkage to the Sonic hedgehog pathway [42,43]. Moreover, gene expression profiles were found to predict clinical outcome and survival with greater accuracy than current clinical risk criteria and outcome predictors or by single gene outcome predictors [42]. Gene expression differences not only were predictive of survival, but also successfully classified classic and desmoplastic medulloblastomas (Fig. 5.4) [65].

TUMOR EXPRESSION PROFILING: GLIOMA

Currently, patients with the most common form of glioma in adults, glioblastoma, have a median survival of twelve months after diagnosis. These dismal statistics remain remarkably stable, despite aggressive surgical resection, radiation therapy, and current forms of chemotherapy [74]. The present classification schema for gliomas is based on pathologic and microscopic criteria; this is based on the assumption that tumor cells share similarity to a presumed neural or glial precursor (Fig. 5.5). According to this schema, tumors that are considered less aggressive share more similarity with their cells of origin and normal tissue counterparts; by contrast, tumors that are more malignant, such as glioblastoma, share features of less differentiated precursor cells.

Who Brain Tumors Clasification

figure 5.5 Classification scheme for brain tumors. (A) The present classification scheme for brain tumors. This classic model is based on the assumption that tumor cells of a specific lineage share microscopic similarity to a presumed neural or glial precursor. The black arrows indicate the hypothesized normal development and the red arrows indicate the hypothesized cell of origin of CNS tumors. (B) According to this scheme, less malignant tumors resemble their normal tissue counterparts; more malignant tumors resemble less differentiated precursor cells. Tumors are graded on the basis of the extent of anaplasia and other microscopic features that connote aggressive behavior — such as mitotic activity, tumor necrosis and angiogenesis. The white arrow points to a mitotic figure. (Nature Rev Neurosci, Mischel et al. (2004). Copyright Nature Publishing Group. Reproduced with permission). See Plate 5.5 in Color Plate Section.

figure 5.5 Classification scheme for brain tumors. (A) The present classification scheme for brain tumors. This classic model is based on the assumption that tumor cells of a specific lineage share microscopic similarity to a presumed neural or glial precursor. The black arrows indicate the hypothesized normal development and the red arrows indicate the hypothesized cell of origin of CNS tumors. (B) According to this scheme, less malignant tumors resemble their normal tissue counterparts; more malignant tumors resemble less differentiated precursor cells. Tumors are graded on the basis of the extent of anaplasia and other microscopic features that connote aggressive behavior — such as mitotic activity, tumor necrosis and angiogenesis. The white arrow points to a mitotic figure. (Nature Rev Neurosci, Mischel et al. (2004). Copyright Nature Publishing Group. Reproduced with permission). See Plate 5.5 in Color Plate Section.

Therefore, tumors are graded according to features such as the number of mitoses, presence or absence of neovascularization and necrosis, nuclear atypia, and mitotic activity. The intra- and inter-patient tumor heterogeneity has made this pathologic classification difficult. As with the prior example for medulloblas-toma, this system is useful for stratifying patients according to histopathological diagnosis, but is less useful for identifying clinically relevant subsets of patients that might differ both in their clinical course and in their responsiveness to therapies.

Aberrations of signaling pathways have also been implicated in the pathogenesis and progression of glial tumors. Constitutive activation of the phospha-tidylinositol 3-kinase and the Ras-MAPK (mitogen-activated protein kinase) signaling pathways have been shown to promote tumor formation and progression in mouse genetic models [75]. Patient-derived tumor samples have confirmed chronic activation of these pathways in correlative experiments [76,77]. As with medulloblastoma, it is hoped that subclassification of molecular subtypes of gliomas will then permit more specific targeting by biologic agents that target these abnormal pathways. The biggest challenge in the utility of such small molecule inhibitors is that brain tumors are both histologically and genetically heterogeneous. For example, the model for use of biologic agents as inhibitors of specific pathways is chronic myelogenous leukemia (CML), in which the target is a constitutively activated BCR-ABL breakpoint cluster region. The ABL-kinase inhibitor imati-nib mesylate (Gleevac, STI-571, Novartis) promotes remission in up to 95 per cent of patients with CML [78,79]. For brain tumors, the molecular target may ultimately be present in only a small subset of pathologically identical tumors, which would complicate clinical trials substantially [80]. Moreover, if multiple molecular abnormalities are present, specific targeting of one step in the pathway may not take into account downstream modifiers or downstream mutations involving the same pathway [81].

0 0

Post a comment