VHL and clear cell Renal Cell Carcinoma Gene expression profiles in - PowerPoint PPT Presentation

11/7/2014 VHL and clear cell Renal Cell Carcinoma Gene expression profiles in renal cell • VHL syndrome hallmark cancer: carcinoma – Clear cell renal cell carcinoma (ccRCC) October 23, 2014 • VHL mutations also a hallmark of sporadic ccRCC • VHL mutation  HIF stabilization and gene expression changes associated with the hypoxia response. Title W. Kimryn Rathmell, MD, PhD Renal Cell Carcinoma (RCC) Expression profiling reveals cell of origin • Originates in the renal cortex • Most common solid lesion occurring in the kidney (80 ‐ 85% of all primary renal neoplasms) Diseased Kidney 3 Cancer Genome Atlas Consortium, Nature, 2013 1

11/7/2014 VHL/HIF Regulatory Pathway Outline HIF2 α HIF1 α HIF2 α HIF1 α • Gene expression profiles in clear cell renal cell carcinomas. HIF2 α HIF β HIF1 α HIF β • Validating a prognostic signature in RCC. • Exploring gene expression profiles in VHL GLUT1 CCD1 PDGF MMP2 PDK syndrome RCC tumors. FLT1 TGF α CXCR4 OCT4 LDHA VEGF BNIP3 Proliferation De- Invasion Angiogenesis Glycolysis differentiation Metastasis Apoptosis Gene expression patterns Sporadic VHL mutant tumors express show HIF ‐ specific variability HIF1 and HIF2, or HIF2 alone. • HIF1 drives glycolytic genes, mTOR pathway • HIF2 drives genes involved in cell cycle and DNA damage. • Overlap in angiogenesis and motility targets. Gordan et al, Cancer Cell, 2008 Gordan et al, Cancer Cell, 2008 2

11/7/2014 Using gene expression to identify K=3, K=4 still fall into two groups tumor subgroups. ccA and ccB Brannon, et al, Genes and Cancer, 2010 Brannon, et al, Genes and Cancer, 2010 ccA (classical angiogenic), ccB (bad) Validation in a historical dataset Brannon, et al, Genes and Cancer, 2010 Brannon, et al, Genes and Cancer, 2010 3

11/7/2014 Extent of Disease at Diagnosis Determining Prognosis: Anatomic Extent of Disease • Most cancers of the kidney and renal pelvis are • Most consistent factor used to determine RCC prognosis diagnosed when the disease is still localized to the 5-year Cancer-specific Survival primary site Based on TNM Stage Unknown 5% 1.00 Stage I (N=185) Probability of Survival TNM 5-year Cancer- Metastatic Stage II (N=57) Stage specific Survival .75 Spread 20% Stage I 91 ± 2.5% .50 Stage II 74 ± 6.9% Stage III Localized (N=83) Stage III 67 ± 6.1% Disease 56% .25 Stage IV 32 ± 3.2% Log Rank P V alue<.001 Stage IV (N=318) 0 Loco-regional 12 24 36 48 60 72 84 96 108 120 132 Spread 19% Months of Postsurgery National Cancer Institute. SEER Stat Fact Sheets. Available at: 13 14 http://seer.cancer.gov/statfacts/html/kidrp.html. Accessed August 28, 2008. Reprinted with permission from Tsui KH, et al. J Urol . 2000;163:1090-1095. ccA/ccB predicts for cancer specific and Validation dataset reveals a small overall survival outcomes distinct tumor set Brannon, et al, Genes and Cancer, 2010 Brannon, et al, Eur Urol, 2012 4

11/7/2014 Cluster 3 are highly divergent from ccA VHL mutants in both ccA and ccB and ccB in metabolic genes. Brannon, et al, Eur Urol, 2012 Brannon, et al, Eur Urol, 2012 Developing a clinic tool: ClearCode34 Prognostic value of ClearCode34 evaluated in TCGA 95 clear cell tumors LAD and ConsensusCluster analysis Set aside arrays with non-concordant assignments 0.0 0.2 0.4 0.6 0.8 1.0 .0 ccA .0 ccA 0.0 0.2 0.4 0.6 0.8 1.0 Recurrence-free Survival 1 1 Cancer-Specific Survival ccB ccB 72 arrays (microarray standard set) .8 .8 0 0 (69 tumors, 3 replicates) (probability) (probability) .6 .6 0 0 43 ccA arrays 29 ccB arrays .4 .4 0 ccA ccB ccA ccB 0 (42 tumors, 1 replicate) (27 tumors, 2 replicates) (n=205) (n=175) (n=205) (n=175) .2 No. of events 14 29 0 No. of events 50 75 .2 0 Median CSS, months -- -- Median RFS, months 91 53 HR, 2.9; 95% CI, 1.6 to 5.6; P =0.0005 .0 HR, 2.3; 95% CI, 1.6 to 3.3; P =4.3e-06 Prediction Analysis for Microarrays (PAM) 0 .0 0 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 Time (months) Time (months) Time (months) Time (months) Predictive biomarkers: ClearCode34 Brooks, et al, Eur Urol, 2014 Brooks, et al, Eur Urol, 2014 5

11/7/2014 Prognostic value of ClearCode34 Prognostic value of ClearCode34 validated in TCGA validated in UNC cohort 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 ccA ccA ccB .0 .0 1 Recurrence-Free Survival 1 Cancer-Specific Survival (n=69) (n=88) ccB No. of events 26 59 .8 .8 0 Median RFS, months 88 52 0 HR, 2.1; 95% CI, 1.3 to 3.4; P =.001 (probability) (probability) .6 .6 0 0 .4 .4 0 0 ccA ccB (n=69) (n=88) .2 .2 0 0 No. of events 7 25 ccA Median CSS, months -- 151 .0 HR, 3.0; 95% CI, 1.3 to 7.0; P =.005 Abbreviation: HR, hazard ratio ccB .0 0 0 0 50 100 150 200 0 50 100 150 200 Subtype ccA was used as reference in univariate and multivariate analysis. 0 50 100 150 200 0 50 100 150 200 Time (months) Time (months) Time (months) $ Stage I was used as reference in univariate and multivariate analysis. Stage was encoded Time (months) as an ordinal variable with three levels. || Grade 1 and 2 were combined and used as reference in univariate and multivariate analysis. Grade was encoded as an ordinal variable with three levels. Brooks, et al, Eur Urol, 2014 Brooks, et al, Eur Urol, 2014 RCC Algorithms for cancer-specific survival Integrated prognostic models can evaluate risk outcomes Group Risk Score Low 0-0.5 Intermediate 0.5-1.5 High >1.5 0.0 0.2 0.4 0.6 0.8 1.0 .0 Low Risk (N=98) 0.0 0.2 0.4 0.6 0.8 1.0 .0 1 1 UCLA Integrated Staging System (UISS) Intermediate Risk (N=140) Cancer-Specific Survival Recurrence-free Survival High Risk (N=28) .8 .8 0 0 Log-rank P= 3.04e-09 .6 (probability) .6 (probability) 0 0 .4 .4 0 0 Low Risk .2 .2 0 0 Intermediate Risk High Risk .0 Log-rank P= 2.03e-08 .0 0 0 0 50 100 150 200 0 50 100 150 200 0 50 100 150 200 0 50 100 150 200 Time (months) Time (months) Time (months) Time (months) Brooks, et al, Eur Urol, 2014 6

11/7/2014 ClearCode34 Model outperforms ClearCode34 Summary established algorithms • ClearCode34 can accurately classify ccRCC tumors .55 .60 .65 0 2 4 6 8 10 .55 .60 .65 .70 0 . 6 5 1 0 0 5 10 15 20 25 30 35 0 .7 0 Change in chi-squared statistic 3 5 Change in chi-squared statistic • Prognostic value of ccA/ccB classification validated 3 0 1.9 6.7 * 8 14.2 * 3.2 in a TCGA and UNC cohort 0 . 6 5 C h a n g e in C h i-s q u a r e d S t a tis tic 0 .6 0 C h a n g e in C h i- s q u a r e d S t a t is t ic 2 5 C-index C-index 6 C - in d e x • Integrated model for recurrence-free and cancer- C - in d e x 2 0 0 . 6 0 specific survival constructed using ccRCC subtypes 4 1 5 0 .5 5 and traditional clinical variables stage and grade. 1 0 0 . 5 5 2 .50 .50 5 * • This classifier adds value to predicting cancer- 2.5 7.3 * * 0 .5 0 26.9 15.9 UISS ClearCode34 0 0 . 5 0 Model->UI UI->Model SSIGN ClearCode34 UISS ClearCode34 Model 0 Model->SS SS->Model CC34Model->UISS Model specific survival above and beyond established SSIGN ClearCode34 Model CC34Model->SSIGN Model algorithms. Brooks, et al, Eur Urol, 2014 How do we know this is meaningful for VHL cases in the TCGA ‐ A mixture of VHL patients? ccA and ccB TCGA includes 14 tumors from VHL Size patients: TCGA Sample ID ClearCode34 Status Grade (cm) T Stage N Stage M StatusStage KIRC TCGA ‐ A3 ‐ xxxx ccA G2 3 T1a NX M0 Stage I 2 ‐ class comparison KIRC TCGA ‐ A3 ‐ xxxx ccA G3 4.9 T1b N0 M0 Stage I reveals NO significantly KIRC TCGA ‐ A3 ‐ xxxx ccB G2 4 T1a N0 M0 Stage I KIRC TCGA ‐ AK ‐ xxxx G2 5.5 T1 N0 M0 Stage I different upregulated KIRC TCGA ‐ AK ‐ xxxx ccA G2 12 T3b N0 M0 Stage III genes, 209 significantly KIRC TCGA ‐ AK ‐ xxxx ccA G2 7.5 T2 N0 M0 Stage II downregulated. KIRC TCGA ‐ AK ‐ xxxx G3 6.5 T3b N1 M0 Stage III KIRC TCGA ‐ AK ‐ xxxx ccB G3 11 T2 NX M0 Stage II KIRC TCGA ‐ AK ‐ xxxx ccA G2 7 T1b NX M0 Stage I KIRC TCGA ‐ AK ‐ xxxx G2 9 T2 N0 M1 Stage IV KIRC TCGA ‐ AK ‐ xxxx G2 6 T1b NX M0 Stage I KIRC TCGA ‐ AK ‐ xxxx ccB G3 5.5 T3a NX M0 Stage III *VHL disease tumors and sporadic ccRCCs (the majority VHL KIRC TCGA ‐ AK ‐ xxxx ccA G2 4 T1a NX M0 Stage I KIRC TCGA ‐ AK ‐ xxxx ccA G3 9 T2 N0 M0 Stage II mutated) are not distinguishable by gene expression. 7