Skip to Content

Kenneth Aldape, M.D.

University of Texas MD Anderson Cancer Center
1515 Holcombe Blvd, Unit 0085, G1.3595c
Houston, TX 77030
United States of America 
(713) 792-0634
(713) 404-0692 pager

Our laboratory works on the development of molecular markers in brain tumors. Specifically, we identify molecular subtypes of gliomas in order to help personalize therapy for patients with these tumors. In addition we identify specific biomarkers of outcome that can be used as an adjunct to traditional histopathologic diagnosis.  These biomarkers are used so that clinicians can have a better sense of patient outcome prior to the optimizing therapy for these patients. Our labatory is involved in several areas of research for the identification of molecular subtypes and prognostic markers in primary brain tumors. A summary of our recent findings is below.

  1. Identification of molecular subtypes of gliomas including identification of the proneural and mesenchymal subtypes.
  2. Identification of a multi-gene prognostic markers in glioma that can be used for stratification of patient therapy.
  3. Identification of a new methylator phenotype of glioma with implications for targeted therapy for patients whose tumors have this contact.

Identification of molecular subtypes of glioma. Using expression profiling and unsupervised analyses we have identified several robust sub subtypes of glioma (Phillips et al Cancer Cell 2006). One subtype we named proneural because genes that are over expressed are associated with neural development. The second molecular subtype that we have identified is the mesenchymal subtype. This subtype is so named because overexpressed genes are associated with mesenchyme.  The acquisition of mesenchymal characteristics in glioma is potentially analogous to the epithelial to mesenchymal in transition that is found in epithelial carcinoma. Using these molecular subtypes we are currently examining whether specific therapeutic agents are differentially effective based on subtype.  In addition, we are identifying the transcriptional network (Caro et al Nature 2010) responsible for mechanistic regulation of the mesenchymal phenotype.  Finally, we are learning that interactions of the tumor cell with its microenvironment are key in the acquisition of mesenchymal characteristics.  This is a very exciting new direction in our laboratory. 

Identification of a multi-gene prognostic marker panel. Using a variety of data sets, both publicly available and from our own laboratory, we have identified a set of of genes whose expression is associated with patient outcome. We find that a multi-marker panel composed of specific prognostic genes is a more reliable and robust measure of patient outcome as compared to any single gene. This is to perhaps related to the heterogeneity of single gene measurements which is balanced out when averaging multiple expression measurements. This multigene panel has been developed into a clinical test which is in use to estimate prognosis of patients. Patients whose tumors show in unfavorable genetic signature are giving more aggressive therapy as compared to patients whose tumors show a more favorable molecular signature. We plan to validate the signatures on large clinical trial samples using uniformly treated patients.

Identification of a CpG methylator phenotype. Using  recently developed methods that performs unbiased screens of CpG island methylation, we have identified a unique pattern of methylation that occurs in a subset of gliomas (Noushmeher et al Cancer Cell 2010). This methylation pattern effects specific set of genes and it's analogous to the CpG islands methylator phenotype (CIMP) previously described in other solid tumors. Using matched primary and recurrent samples from the same patients, we find that this epigenetic change is stable, implying that is associated with tumor development and recurrence. Patients whose tumors show the center also and show somatic mutations in the  recently characterized gene IDH1. The coexistence of the methylation can you type in dictation's in the IDH1 gene suggest a mechanistic link between these two genetic lesions. In addition the existence of anti-methylation agents in the clinic suggest a potential therapeutic for patients whose tumors showed this phenotype. Our ongoing work will more fully characterize the relationship of the methylator phenotype with vacation in the IDH1 gene to identify mechanistic principles which regulate these epigenetic changes.

© 2014 The University of Texas MD Anderson Cancer Center