Reconstruc*ng Human Tumor Histories By Comparing Genomes From Different Parts of the Same Cancer

Transcription

1 Reconstruc*ng Human Tumor Histories By Comparing Genomes From Different Parts of the Same Cancer Darryl Shibata Professor of Pathology University of Southern California Keck School of Medicine

2 Soma*c Evolu*on in the Clinic Widely Accepted That Tumors Evolve Many Different Possible Pathways Serial Observa*ons Imprac*cal A Pa*ent Suddenly Has Cancer Clinical Ques*ons (Pa*ent Specific): How Did This Tumor Evolve? Do Different Evolu*onary Histories MaWer?

3 An Approach To Pa*ent Specific Tumor Histories Coalescent Theory Molecular Clocks Topography LeY Right tumor genomes Present Number of Differences ( *me or PWD) First Transformed Cell Time Past

4 2011 Basic Cancer Ancestry Reconstruc2on Cancer in a 70 YO Back In Time first transformed cell tumor mito2c age total mito2c age 1941 (birth) zygote

5 A Problem: The Evolu*on of Any Individual Human Cancer is Unknown st transformed cell Sequen*al or Clonal Evolu*on Big Bang (full malignant poten*al at transforma*on) Older Parts More Diverse Uniform Diversity

6 Colorectal Cancer: Adenocarcinoma (small glands or popula2ons or neighborhoods of adjacent cells)

7 How To Sample Tumor Diversity? EDTA Washout: Single Cancer Glands Isolate DNA PCR clock locus clone PCR products into bacteria sequence individual clones calculate PWD (pairwise distance)

8 How To Sample Tumor Diversity? Common Gland Ancestor young cancer older cancer start First Transformed Cell gland age tumor age gland age Common Gland Ancestor gland age tumor age gland age Time Or Diversity

9 Somatic Cell Molecular Clock Problems: --- Somatic Cell DNA Replication Fidelity Too High! Potential Solution: Epigenetic Molecular Clock 5 3 CGATCTGCATCGACTGCCGCG GCTAGACGTAGCTGACGGCGC Substitute the 5 to 3 Order of Bases With the 5 to 3 Order of CpG DNA Methylation

10 Replica2on Clock Molecular Clock: Informa*on Passed From Cell to Cell Epigene*c Fidelity is less than Gene*c Fidelity 10-9 Genome Replica2on versus 10-5

11 Human Colorectal Cancer ley side 5 3 six cancer glands six cancer glands right side

12 PWDs Between Cancer Sides Differ Between Cancers (N=12) BGN pairwise distance between sides es*mated divisions (mito*c age) ley tumor side Different human cancers have different ages first transformed cell right tumor side

13 Cancer Glands From LeX and Right Sides Are Similar For The 12 Human CRCs ley side intragland PWD A p=0.0007, r=0.84 BGN right side intragland PWD ley side intragland PWD p= , r=0.94 LOC right side intragland PWD Glands Within A Cancer Have Similar Ages

14 ley tumor side right tumor side first transformed cell Pairwise Distance Older Cancers Have older Glands younger cancer within gland left right between sides within gland left right between sides older cancer 0

15 Simple Models of Tumor Growth 1 intragland PWD 4 2 BGN 3 4 1st transformed cell Cancer 2 Cancer 12 glands differ in age or PWDs glands have similar ages or PWDs

16 Molecular Clocks: Different Speeds 1) DNA CpG Methyla*on (Epigene*c ) 2) Chromosomal Copy Number ( CIN ) 3) Loss of Heterozygosity (LOH) 4) DNA Sequence Muta*ons Challenge Is to Gather the Data And Then Interpret the Data Fast Slow Illumina 660 SNP Microarray

17 Chromosomal Changes 5 3 LOH 5 LeY Right..AGCTCGCA TCTTCAAGCCT ACCATTAAT...AGCTCGCA TCTTCAAGCCT ACCATTAAT AGCTCGCA TCTTCAAGCCT ACCATTAAT...AGCTCGCA TCTTCAAGCCT ACCATTAAT. 5 3

18 An Approach To Pa*ent Specific Tumor Histories Coalescent Theory Molecular Clocks Topography LeY Right Present Number of Differences ( *me ) Time First Transformed Cell Past

19 Genomes Are Historical Documents (almost perfect copies of copies) zygote (start) Acknowledgements Yasushi Yatabe Kyoung-Mee Kim Jung Yeon Kim Peter Calabrese Kim Siegmund Paul Marjoram Simon Tavare cancer cell (present day)