Clinical Impact of Molecular Profiling in Cancer and Other Human Diseases

1
Clinical Impact of Molecular Profiling in Cancer
and Other Human Diseases
Susan Done
2
Translational Research
Basic Science
Clinical Practice
3
Translational Researcher

• Credible to basic scientists
• Credible to clinicians
• Persuasive
• Pioneering
• Tireless

4
Disease

• Diagnosis Classification - Prognostic markers
• Treatment Predictive markers

5
Diagnosis

• Signs and symptoms
• Tissue
• Histology
• Microbiology
• Clinical Chemistry
• DNA/RNA
• Single genes
• Microarrays

6
What are microarrays?
Hedenfalk et al. NEJM 2001, 344 539-48
7
Microarrays - Platforms

• Planar
• cDNA
• BAC
• Oligo
• Immobilised bead randomly arranged optically
  encoded beads
• Liquid bead arrays optically encoded by colour
  or size, can be simultaneously decoded and
  analysed by flow cytometry
• Barcoded nanoparticles or quantum dots

8
Suddenly able to look at whole transcriptome
9
Subsets of breast cancer

• Luminal A and luminal B, basal, ErbB2 and normal
• Tumours cluster into distinct groups reflected in
  different patterns of cytokeratin expression
• Associated with different outcome
• Further divides groups that histologically appear
  homogeneous
• May help in identification of pathways that lead
  to breast cancer

Sørlie et al. Proc. Natl. Acad. Sci. USA 2001
98, 10869-74
10
Basal and Triple negative breast cancer

• Lack of expression of ER, PR, Her2
• Approximately 15, higher in African and African
  American women
• Overlap with basal subgroup (CK5/6, EGFR, P53
  mtn)
• Similarities to BRCA1 associated breast cancers

11
Significance of gene expression patterns

• Clusters of genes associated with
• ER
• (Bertucci et al. Hum Mol Gen 2002, 11863-72)
• Tumour stage
• (van de Vijer et al. NEJM 2002, 3471999-2006)
• Tumour size
• (Martin et al. Cancer Res 2000, 602232-8)
• BRCA1 or BRCA2 mutations
• (Hedenfalk et al. NEJM 2001, 344539-48)
• Basal vs. Luminal cancers
• (Sørlie et al. Proc. Natl. Acad. Sci. USA 2001
  9810869-74)

12
Gene profile predicts outcomes
117 small cancers from Netherlands Cancer
Institute 70-gene profile independently predicted
distant spread and overall survival. Better than
standard predictive models and nodal status.
van de Vijer et al. NEJM 2002, 3471999-2006
13
70 gene prognosis profile

• Further tested in 295 patients (lt 53 years, stage
  I or II, 144 LN positive)
• Found to be an independent prognostic indicator

van de Vijer et al. NEJM 2002, 3471999-2006
14
Gene expression pattern predicts for prognosis

• 117 young patients (lt 55 yrs)
• 25,000 genes on microarray
• 70 genes predict for prognosis
• Signature for ER and BRCA1 status

Vant Veer et al. Nature 2002, 415530-36
15
Debate about size of prognostic group of genes

• 70 (van de Vijer et al. NEJM 2002, 3471999-2006)
• 23 (Bertucci et al. Hum Mol Gen 2002, 11863-72)
• ? More or less
• Too early to say

16
Gene expression profiles as predictors

• Response to docetaxel
• 92 gene set
• Primary breast tumours from 24 patients before
  treatment with neoadjuvent docetaxel
• Monitored tumour size

Chang et al. Lancet 2003, 362 362-69
17
Leukemia

• Although cytogenetic abnormalities characterise
  many cases of AML, still 45 show a normal
  karyotype
• Now 75-80 can be further characterised by
  molecular methods e.g FLT3, MLL, NPM1
• Prospective study in GEP of 4000 cases in 11
  european centres

18
DNA - Karyotyping
www.pathology.washington.edu/galleries/Cytogallery
/
19
CGH Microarray Brief Protocol
3 X 2 ?g Control DNA
3 X 2 ?g Test DNA
Bioprime DNA Labeling Kit (Invitrogen)
10 ?g tRNA, 40 ?g Human Cot-1 DNA
Speed Vac
Hybridize 14-18 hours at 37ºC
Pool and dissolve in 80 ?L DIG Easy Hybe
5 min spin
Denature, incubate 30 mins at 37ºC
20
mCGH
www.upci.upmc.edu/facilities/Cytogen/
21
Microarray CGH
On the left is a 19K cDNA microarray slide
containing 19200 genes and/or ESTs. Above is one
subarray of the 19K array showing the location of
an amplified gene (red spot) and a deleted gene
(green spot).
22
Array CGH

• Choice of platform
• BAC
• Oligonucleotides
• cDNA
• Factors to consider include
• Annotation
• Size of targets
• Analysis software
• Cost

23
An overlay of the two profiles generated from a
formalin-fixed paraffin-embedded breast cancer
case without amplification (black spots) and with
SCOMP-amplification (grey spots) shows identical
profiles with common regions of amplification
highlighted at 1q, 7q11-21 and 16q12
Ghazani, A A et al. J Clin Pathol 200659311-315
24
MCF7 CGH profiles of chromosomes 17 and 20
A. Fresh vs. FFPE MCF7 B. DOP-PCR MCF7 vs.
DOP-PCR FFPE MCF7 C. SCOMP MCF7 D. SCOMP FFPE MCF7
Ghazani, A A et al. J Clin Pathol 200659311-315
25
aCGH

• Prostate 6q15 deletion (Lapointe 2007)
• Cleft lip and palate identified new candidate
  genes (Osoegawa 2007)
• Copy number variations

26
Inherited diseases

• Cystic fibrosis American College of Medical
  Genetics 23 mutations - 2004
• Tay-Sachs, Gaucher Type 1, NiemannPick A and B,
  mucolipidosis Type IV, familial dysautonomia,
  Canavan, Bloom syndrome and Fanconi anemia type C
  combined carrier rate of 16 in the Ashkenazi
  Jewish population

27
Disease susceptibility - SNPs

• HapMap project (www.hapmap.org)
• High density genotyping arrays 500,000 single
  nucleotide polymorphisms (SNPs)
• Platelet genomics and risk of thrombosis 110
  genes
• Response to therapy in asthma and COPD

28
Cellular heterogeneity is a feature of invasive
breast cancer
29
Tissue Microdissection

• Allow study of histologically defined lesions
• Necessary because of heterogeneity of many
  tissues
• Techniques have evolved from excavation of
  formalin-fixed paraffin-embedded (FFPE) tissue
  blocks to laser capture microdissection of single
  cells

30
Microdissection
Area of interest identified and microdissected
using the Arcturus Laser Capture Microdissection
system (PixCell II). Larger areas are removed
with two 18G needles under a stereomicroscope.
31
Whole Genome Amplification (WGA)

• Developed to allow study of small numbers of
  cells
• Many protocols
• Two broad categories
• PCR-based
• Isothermal DNA amplification

32
Whole Genome Amplification PCR based methods

• Interspersed Repetitive Sequence PCR (IRS-PCR)
  Nelson PNAS (1989) 866686
• Degenerate-oligonucleotide primed PCR (DOP-PCR)
  Telenius Genomics (1992) 13 718
• Primer-extension Pre-amplification PCR (PEP-PCR)
  Zhang PNAS (1992) 895847

33
Whole Genome Amplification DNA fragmentation

• Random
• Sequence specific Klein PNAS (1999) 964494
• Adapter ligation
• Addition of poly(dT) tails by terminal
  deoxynucleotidyl transferase

34
Whole Genome Amplification (WGA) and Array CGH
Genomic DNA Mse I digestion
Adaptors annealing and ligation
5
3
5
3
PCR Amplification Purification
Cy3 (Test) Cy5 (Control) Labeling
Combine Cy3 Cy5 labeled samples for competitive
hybridization to the H-19k-cDNA array, UHN
Scan Analysis
Klein CA, etc Proc Natl Acad Sci U S A
199996(8)4494-4499
35
Whole Genome Amplification DNA polymerases

• Multiple displacement amplification (MDA) Phi29
• Large Fragment Bst DNA polymerase Aviel-Ronen BMC
  Genomics (2006) 7312
• Hyperbranched Strand Displacement Amplification
  (HSDA)
• Isothermal
• Ability of polymerase to cause strand
  displacement
• Random initiation points using random primers
• Products in absence of template

36
Whole Genome Amplification - Applications

• Microdissected tissue samples
• Preimplantation diagnosis
• Bacterial cells for genomic sequencing

37
Whole Genome Amplification - Limitations

• Accuracy of representation
• Preferential amplification related to chromosomal
  location or genomic sequence
• How to assess?

38
Comparison of whole genome amplified array CGH
data by ArrayNorm software
Ghazani, A A et al. J Clin Pathol 200659311-315
39
TMAs
www.worldcommunitygrid.org
40
Permits study of large numbers of cases

• Low cost of IHC
• Time consuming to constuct
• Can be used for multiple studies
• Stromal mast cells a marker of favourable
  prognosis in 4,444 breast cancers (Rajput2007)

41
Reverse Phase Protein Microarrays

• Cell lysates
• Plasma
• Serum
• Protein quantity and presence of post
  translational modifications
• Limited by availability of antibodies

42
Protein Microarrays

• Peptides or antibody libraries
• Protein expression
• Ligand specificity
• Serotyping
• Post-translational modifications

43
Are Gene Microarrays Ready for Routine Clinical
Use?

• Some problems validating studies probably
  relate to differences in gene sets, different
  arrays, etc.
• Need to standardise the technology
• Need to conduct larger studies
• MIAME (Minimal Information About a Microarray
  Experiment) guidelines www.mged.org

44
Array Issues

• Need fresh frozen tissue RNA may be degraded
  during routine specimen handling
• Technology expensive cost will probably come
  down as we move to more focused arrays

45
RT-PCR panel

• Reverse transcriptase PCR
• RNA in formalin-fixed paraffin-embedded tissue
  blocks
• Panel of 21 genes (5 control) selected based on
  literature
• Devised model using cases from NSABP-B20 and two
  other groups and validated using NSAPB-B14
• ER-positive, node-negative
• Low (lt 10), intermediate (10-30) and high risk
  of recurrence

46
Gene expression issues

• RNA
• Standard prognostic markers not taken into
  account
• May result in an immunohistochemical panel

47
Empty promises?

• Multiple microarray studies reviewed lymphoma,
  leukemia, breast cancer and lung adenomcarcinoma.
  Few genes supported by cross validation. (Mtzani
  and Ioannidis Lancet 2003)
• In five of seven studies patients were not
  classifies better than by chance (Michaelis
  Lancet 2005)

48
Developing a new clinical test Factors to
consider

• Clinical need - Significant number of patients
  affected
• Clinical need Problems with existing methods
• Reliable
• Low cost

49
Clinical trials in Breast Cancer using Gene
Microarrays

• MINDACT comparing NKI prognostic signature to
  Adjvant!-Online
• Massachusetts General Hosp. validating NKI
  prognostic signature (5000 patients)
• M.D. Anderson validating own signature for
  predicting drug (paclitaxel-FAC) response (210
  patients)
• Baylor College validating own 90-gene prognostic
  signature (100 patients)
• Oncotype DX 18-75 years LNN ER/or PR pos
  (10,046)

50
Impact

• New disease subtypes
• New prognostic/predictive panels
• New therapeutic targets
• Personalised medicine