Childhood Sarcoma Classification by Gene Expression Profiles

IPAM #3: Childhood Sarcoma Classification by Gene Expression Profiles Timothy J. Triche CHLA/USC Clinical Classification of Childhood Cancer • Historical: Morphologic diagnosis + clinical data => risk group, protocol eligibility, treatment (eg, group-based treatment) immunophenotype, genomic defect) => patient-specific group-based treatment on multi-genic phenotype? • Current: Combined (morphology, • Future: Patient-specific therapy, based Osteosarcoma • Five histologic types, no prognostic value • Weak prognostic features: site, size, age • No specific, predictive genetic abnormality (RB, p53) • Clinical stage only significant prognostic indicator at presentation Osteosarcoma Prognosis • Pre-resection chemotherapy => major increase in survival • Improved survival limited to patients with ≥95% tumor kill • Patients w/ metastases can be salvaged But, many exceptions occur: – Responders who metastasize & die – Non-responders who survive – Metastatic patients who survive after resection of mets Thus, predicting outcome & tailoring therapy remains a major problem Osteosarcoma: Response to Chemo Before After Osteosarcoma Survival • Surgery only: <10% • Metastases, no surgery: 0% • Metastases, surgery: ~20% • Single-agent chemotherapy: <20% • Conventional chemotherapy: ~44% • Up-front chemotherapy: ~65% • Responders: ~80% • Non-responders: <40% Multi-gene Analysis by Microarrays • Single gene abnormalities, even when present, are inadequate alone to: – – – – – Establish a diagnosis Identify individual patients risk profile Predict clinical course Predict response to therapy Predict outcome • Increasing evidence suggests gene expression profiles may favorably address these issues Gene Expression Analyses • Scatter Analyses – 1X1 – Groups • Outlier Gene Analyses – Up & down regulated from mean – Identity • Cluster Analyses – All genes – Various methods Specimen Handling A) Cut pilot section of OCT embedded frozen tissue B) Cut ~12 frozen sections C) Extract RNA (<5ug total RNA) D) Synthesis of double-stranded cDNA E) In-vitro transcription w/ biotinylated nucleotides F) Size confirmation of cRNA transcripts tumor non-tumor dissection of tumor tissue when possible pure tumor 500 bp G) Fragmentation of cRNA Osteosarcoma: Gross Appearance Histopathology of Osteosarcoma Gene Expression: Osteosarcoma Pilot data 2 1 6= primary tumor, 1993 11= first metastasis, 1996 9= second metastasis, 1998 (died 1999) Met 1 vs. met 2: little similarity Primary vs. 1st Metastasis Primary 1st Pulmonary Met Primary vs. Metastatic Osteosarcoma Ribosomal Protein L30 Oste one ctin Ribosomal Protein L37A TF SL1 Thymosin IMP E16 Pinch Prote in CPT1 Cyclin A Tat-SF1 CAMP PK RII subunit NGF be ta Tyrosine Phosphatase PIGA, A Uncoupling Protein 3 PSA Ribosomal Protein L32 PSG11 Differential Gene Expression: Osteonectin lost in metastasis Primary Metastasis 0 50000 100000 150000 200000 250000 -50000 Primary vs. “Metastasis” Primary, 1993, pre-Rx Tibia lesion, 1998, pre-Rx Gene Expression Data Clustering Multiple methods work Pattern Recognition No process knowledge Data Postulated Patterns Millions of possible patterns Generate possible patterns: Scenario analysis Non-numeric simulations Computational linguistics Neural networks Linear/non-linear optimization methodology Discovery of patterns buried in massive dataset Iterative Process Pattern Tested Neural net uses data to optimize pattern New Postulated Patterns Pattern recognition New rules developed Optimized Set of Patterns Limited set of probable patterns Agglomerative vs. Optimizing Hierarchical Clustering • Both build a tree of clusters, with data points as leaves, & “nearby” data points as siblings. • Agglomerative method repeatedly finds closest pair and irreversibly groups them. Bottom-up. Binary tree. • Optimization methods reconsider assignments based on other assignments and their effects on cluster means & variances. • Minimize sum of squared distances. – Distance measure matters. – Relate to statistical noise models, co-regulation models & likelihood of fit. Agglomerative vs. Optimizing Hierarchical Clustering, cont. • Optimize means, variances, and cluster memberships. • Currently we optimize top-down, by levels • Expectation Maximization: soft memberships. K-means: hard. • Optimize tree topology (fanout) by CV • SOM also optimizes at one level, and requires low-dimensional grid embedding of cluster means. • Alternative to data-cluster distances: cliques of low data-data distances. Also has EM-like stat mech algorithms. Mimir User Interface Courtesy of Eric Mjolsness, JPL Data Flow forSarcoma Analysis gene clustering sample clustering classifiers data labels scoring Pilot Study of Sarcomas 17 cases of osteosarcoma and rhabdomyosarcoma 6800 GeneChip analysis 6800 genes yield 14 gene clusters Reduced mean space yields 4 sample clusters OS OS OS, OS ERMS OS, ARMS 1ª ERMS X 4 OS x 3 ARMS met X 3 Expandable Tree of Variables Characterizing a Tissue Sample All variables Subject Conditions Genes Outcomes Clinical Demographics Clinical response Metastasis Survival Treatment Pathology Age, Sex, etc.… EM (Expectation Maximization) Gene Clustering Sarcoma Dataset: 45 cases of RMS (Alv + Emb) & Osteosarcoma (R + NR) A B C F G D J K = POOR = INTERMEDIATE = FAVORABLE Working hypothesis: Gene expression profiling can detect prognostic distinctions among sarcomas independently of conventional clinical or diagnostic criteria Future Directions • Analyze larger data set (institutional, COG) to test hypothesis • Expand to all sarcomas (RMS, non-RMS, OS, ESFTs) • Identify biologically important genes • Creation of custom “sarcoma” arrays using oligomers representing these genes • Long term studies of COG sarcoma patients using arrays in context with current clinical & biology studies All osteosarcomas Osteosarcoma vs RMS Genes Log (0steo) 5.68 194.82 6.50 1952.82 7.31 2240.65 6.01 162.35 6.66 5.90 5.75 1225.65 6.43 5.64 8.63 12771.76 8.25 12913.76 6.18 8.77 19340.53 7.09 6.36 6.62 732.18 1659.82 907.00 8.58 10916.54 7.72 4836.68 8.00 6595.57 1.83 1.84 1124.24 8.01 7048.57 6.29 1756.00 1.90 2.25 5.63 2011.11 2.33 1166.59 890.82 8.40 9229.32 7.48 3556.00 10.75 77271.57 2.33 2.36 2.39 2058.88 629.94 8.93 13677.36 7.93 9770.93 7.71 3923.54 2.44 2.49 2.56 8.28 8099.00 2.71 9.77 31139.89 2.79 8.87 16832.71 2.83 Mean (Osteo) Log (Rhabdo) 7.96 Mean (Rhabdo) 6357.93 Mean log Rhabdo 2.83 2.83 GENE DESCRIPTION TNNT1 Troponin T1, skeletal, s low MEST Mesoderm specific trans cript (mouse) homolog IGF2 Insulin-like growth factor 2 (somatom edin A) FGFR4 Fibroblast growth factor receptor 4 IGF2 Insulin-like growth factor 2 (somatom edin A) Adrenal-Specific Protein Pg2 Steroid receptor coactivator (SRC-1) m RNA GB DEF = DNA for cellular retinol binding protein (CRBP) exons 3 and 4 RBP1 Cellular retinol-binding protein Insulin-Like Growth Factor 2 PTN Pleiotrophin (heparin binding growth factor 8, neurite growthpromoting factor 1) Muscle acetylcholine receptor alpha-s ubunit MMP2 Matrix m etalloproteinase 2 (gelatinase A; collagenase type IV) CCND2 Cyclin D2 TNNI1 Troponin I, s keletal, s low MYL1 Myos in light chain (alkali) Proposed COG Study of All Sarcomas Acknowledgements • CHLA: – Deb Schofield – Jingsong Zhang • Caltech: – Barbara Wold – Chris Hart • USC: – Jonathan Buckley – Kim Siegmund • JPL: – – – – Eric Mjolsness Tobias Mann Joe Roden Ben Bornstein • NCCF: – Mark Krailo • UBC: – Poul Sorensen