Cancer Biology I Computer-Aided Discovery Methods

1
Cancer Biology IComputer-Aided Discovery Methods

• Ching C. Lau, MD PhD
• Cancer Genomics Program
• Texas Childrens Cancer Center
• Baylor College of Medicine

2
What is Cancer?

• Not a single disease
• Carcinoma vs sarcoma
• Organ and tissue specific
• Characterized by unregulated cell growth and the
  invasion and spread of cells from the site of
  origin
• Genetic disease at the cellular level (clonal)

3
CANCER ARISES FROM THE ACCUMULATION OF GENETIC
CHANGES



Hormones, Diet, Host Factors
genetic
first mutation

chemical
second mutation


physical
third mutation

virus
fourth or later mutation
4
The Hallmarks of Cancer
Hanahan and Weinberg, Cell, 2000
5
Integrated Circuit of the Cell
Hanahan and Weinberg, Cell, 2000
6
Integrated Circuit of the Cell
Hanahan and Weinberg, Cell, 2000
7
Integrated Circuit of the Cell
Hanahan and Weinberg, Cell, 2000
8
Cancer as Complex Tissues
Hanahan and Weinberg, Cell, 2000
9
Parallel Pathways of Tumorigenesis
Hanahan and Weinberg, Cell, 2000
10
Regulation of Cell Number
Pecorino, Mol Biol of Cancer
11
Cell Cycle Regulation
Pecorino, Mol Biol of Cancer
12
Classes of Cancer Genes
Gain of Function
Loss of Function
Pecorino, Mol Biol of Cancer
13
Integrated Circuit of the Cell
Hanahan and Weinberg, Cell, 2000
14
(No Transcript)
15
(No Transcript)
16
TWO HIT HYPOTHESIS IN RETINOBLASTOMA
17
(No Transcript)
18
Types of Mutations
Pecorino, Mol Biol of Cancer
19
Components of a Gene
Pecorino, Mol Biol of Cancer
20
Chemical Carcinogens
21
Response to DNA Damage
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
Subway Map of Cancer Pathways
Sci Am July, 2003
26
(No Transcript)
27
Evolution of Cancer Theory
Scientific American July, 2003
28
Four Theories of Oncogenesis
Sci Am July, 2003
29
Epigenetics

• Heritable changes in gene expression that are not
  accompanied by changes in DNA sequence.
• Gene silencing at the level of chromatin
• DNA methylation
• Histone acetylation
• Nucleosome remodeling
• Required for differentiation, imprinting, and
  silencing of large chromosomal domains (X
  chromosome)

30
Implications of Cancer Stem Cells
Nature 442743, 2006
31
Concept of Cancer Stem Cell
32
MicroRNA (miRNA) Encoded in genome Transcribed as
primary miRNA (Pri-miRNA) Cleaved to precursor
miRNA (Pre-miRNA) by dsRNA endonuclease
Drosha Exported to cytoplasm by Exportin
5 Further processed by RNase III Dicer to mature
miRNA (21nt) Suppress translation in RISC
33
Predominant Types of Cancer in Different
Populations
Pecorino, Mol Biol of Cancer
34
In Vitro Studies

• Cell proliferation
• Focus formation
• Anchorage independence
• Migration/invasion
• Angiogenesis

35
In Vivo Studies

• Tumorigenicity
• Subcutaneous xenografts
• Orthotopic xenografts
• Metastasis
• Genetically engineered mice (GEM)
• Cancer stem cells

36
Cancer Genomics

• Examine the structural integrity and activity of
  all genes in the cancer cells
• Characterized by high throughput methodologies
  combined with statistical and computational
  analysis of the results.

37
Human Genome Project

• Initiated in 1990
• International cooperative effort to sequence the
  entire human genome
• Initially projected to complete by 2005
• Celera announced completion by 2001
• Draft copy of human genome released in February
  2001

38
Applications of Cancer Genomics

• Improve diagnosis of disease
• Detect genetic predispositions to disease
• Create drugs based on molecular information
• Use gene therapy and control systems as drugs
• Design custom drugs based on individual genetic
  profiles

39
(No Transcript)
40
Cancer Genome Anatomay Project (CGAP) - National
Cancer Institute

• Goals -
• to achieve the comprehensive molecular
  characterization of normal, precancerous, and
  malignant cells
• to identify all the genes responsible for the
  establishment and growth of cancer

41
The Genetic Code
42
Screening Targets

• Genome 3 billion base pairs, only 1 coding
  sequences
• Transcriptome 25,000 to 30,000 genes, only a
  subset of which are expressed in any one cell
  type
• Proteome 500,000 proteins
• Post-translational modification
• Subcellular localization
• Metabolome - metabolites and intermediates in all
  biochemical pathways

43
(No Transcript)
44
(No Transcript)
45
Case 176 Diagnosis PNET (recurrent tumor)
46,XX,t(1215)(p13q25-26)
15
15
12
12
46
cDNA Microarray vs. Oligonucleotide Microarray
47
CGH Method

• Detects gains, losses, amplifications in a single
  hybridization
• Maps lesions to chromosomal region
• Can not detect balanced translocations or small
  alterations

48
CGH - Limitations

• Limited resolution (10-20 Mb)
• Dependent on chromosome morphology
• Labor intensive
• Requires confirmation by FISH using
  locus-specific probes (YAC, BAC, PAC, etc.)
• Subsequent cloning of candidate genes
  time-consuming and laborious

49
DNA Microarray for Genome Scanning
Array BACs on a glass surface
BACs
Labeled Normal DNA
Competitive hybridization with
Fluorescence Ratio
Labeled Tumor DNA
Analyze fluorescence intensity ratio
Gain
Loss
Loss
Gain
Loss
Gain
50
BAC Array CGH

• Improved reliability reproducibility
• Higher resolution over chromosomal CGH
• Confirmation by FISH is streamlined
• Subsequent identification of candidate genes much
  more efficient
• High throughput potential for automation

51
1Mb Human array

• 2600 BAC clones spotted in duplicates
• Average distance between BACs is 1Mb
• Each experiment is done in duplicate with reverse
  labeling

52
Correlation in gains between profiles from frozen
and fixed tissues
ch1
ch22
Frozen
Fixed
53
Whole Genome Amplification Does Not Alter Array
CGH Results
unamplified
amplified
LCM -amplified
Chr 1
whole genome
54
HapMap Project

• Total of about 10 million SNPs
• Haplotype - A set of associated SNP alleles in a
  region of a chromosome
• Tagging SNPs A few SNPs that can provide most
  of the information on the pattern of genetic
  variation in the region

55
HapMap Project

• The HapMap will describe the common patterns of
  genetic variation in humans. It will include the
  chromosome regions with sets of strongly
  associated SNPs, the haplotypes in those regions,
  and the SNPs that tag them. It will also note the
  chromosome regions where associations among SNPs
  are weak.

http//www.hapmap.org
56
Single nucleotide polymorphic allele array SNP
Chip
57
(No Transcript)
58
Glioma International Consortium GLIOGENE

• US
• MDACC
• BCM/TCCC
• Harvard
• UIC/Duke
• Mayo
• UCSF
• CWU
• MSKCC
• Columbia

• Sweden
• Israel
• UK
• Denmark
• China

59
Study Aims

• Screen all glioma cases (N15,000 ) for family
  history of gliomas
• Collect families with at least two gliomas -
  about 2-4 will have significant family history
  for linkage (N400 high-risk famiies)
• Identify regions of the genome where a gene
  candidate linked to familial brain tumors resides
  (Illumina)
• Further interrogate these gene candidate regions
  established in linkage aim by genotyping closely
  spaced genetic markers

60
Genetic Map and Markers

• Genetic maps serve to guide a scientist toward a
  gene, just like an interstate map guides a driver
  from city to city.
• Genetic maps use landmarks called genetic markers
  to guide researchers on their gene hunt.
• Restriction fragment length polymorphisms (RFLP)
• Microsatellite polymorphisms
• Single nucleotide polymorphisms (SNP)

61
Mapping Disease Genes using DNA Markers
Grandparents G1 G2 G3 G4 chromosomes 111111111
222222222 333333333 444444444 111111111 22222222
2 333333333 444444444 Mom Dad mom
dad 111111111 333333333 222222222 4444
44444 Children P1 P2 P3 P4 111221122 2211211
11 111112222 222222111 333344443 444433344 44333
3444 334444444
62
Genome-wide Association Studies (GWAS)
63
(No Transcript)
64
Genetic Changes in Glioblastoma
TCGA, Nature 2008
65
A Paradigm for Molecular Profiling

• Biorepository for tumor and normal tissues
• Cytogenetic profiling by CGH and SKY
• Detailed alleotyping by genotyping
• Expression profiling by cDNA/oligo microarrays
• Epigenetic profiling by CpG island microarrays
• Proteomic profiling by Protein Chip
• Mutation screening of target genes
• Functional studies of differentially expressed
  genes
• Clinical validation of molecular profiles

66
Goals of Genomics Profiling

• Develop genome-based diagnostics
• Molecular markers for diagnosis
• Molecular markers for prognosis
• Surrogate markers for therapeutic and disease
  monitoring
• Risk markers identified by genetic epidemiology
  studies
• Identify novel therapeutic targets

67
Hypothesis

• By performing comprehensive, genome-wide
  characterization of genetic alterations in
  pediatric brain tumors, we will be able to
• Improve the classification of these tumors for
  the purpose of prognostication stratification
• Identify new therapeutic targets

68
Histopathology of Embryonal Brain Tumors
69
Classification of embryonal brain tumors by gene
expression profiling
Pomeroy et al., Nature, 2002
70
Molecular Genomics of Medulloblastomas

• Can gene expression patterns predict outcome of
  medulloblastomas?

71
Treatment Failures
Survivors
Markers of Outcome for Medulloblastoma
Pomeroy et al., Nature, 2002
72
Molecular Signature of Chemoresistance
Poor Responders
Good Responders
Man et al, Can Res 2005
73
Signature of Chemoresistance

• 45 predictor genes were selected in the
  definitive surgery samples of osteosarcoma
  patients that can predict response to
  neo-adjuvant chemotherapy (LOO CV 70)
• These predictor genes can also be used to predict
  good and poor responders at the time of diagnosis
  using initial biopsy samples (overall correct
  prediction rate 93)
• Some of the predictor genes also show significant
  correlation with overall survival (p lt 0.05).

Man et al, Cancer Res, 2005
74
Prediction of Initial Biopsies
Predicted Predicted
ME NM Total
Actual ME 22 0 22
Actual NM 1 6 7
Overall accuracy 97 Sensitivity
100 Specificity 86
11 LD to ME 11 correctly predicted 11 ME at Dx
10 correctly predicted
Man et al, 2008