Cancer remains a leading cause of morbidity and mortality world-wide. It is estimated that at least 14.1 million people are diagnosed with cancer every year and it currently accounts for at least 15% of all human deaths. Recent advances in modern medicine have significantly improved the cure rates over the last few decades, but an increase in adverse drug reactions (ADRs) has been seen with the increased intensity of treatment. Anthracycline-based cancer treatments are highly effective for many types of cancers, especially childhood malignancies and breast cancer and have significantly improved the 5 – year survival rates of many cancers from 30% to greater than 80%. However, the use of Anthracycline-based cancer treatments is being questioned due to a highly individualized ADR, cardiac toxicity. In order for Anthracycline-based cancer treatments to remain on the market as highly effective therapies for many severely ill patients, it is imperative to identify the inherited genetic predispositions causing these ADRs, build an ADR prediction algorithm and implement this in a cost effective pharmacogenetic ADR prevention program. The ultimate goal is to tailor the right treatment to the right patients and improve the safety of these medications currently on the market. The Canadian Pharmacogenomics Network for Drug Safety (CPNDS), recently discovered a novel gene (RARG) responsible for cardiomyopathy and congestive heart failure in cancer survivors and has developed clinical practice recommendations for genetic testing to reduce the incidence of anthracycline-induced cardiotoxicity in children after cancer treatment. Join us as Dr. Folefac Aminkeng, presents the CPNDS’ important research efforts focused on understanding the role of genes in ADRs and developing drug safety solutions for cancer patients, initiatives which are critical to improving long-term survival outcomes.

1. Pharmacogenomic Prediction of Anthracycline-
induced Cardiotoxicity in Childhood Cancer
Folefac Aminkeng
The University of British Columbia
1

2. Outline
Background and Rationale
Aminkeng F et al, Nature Genetics 2015
Sep;47(9):1079-84
Importance of SNP & Variation Suite (SVS)
2

3. Background and Rationale
3

4. Effectively treats or
prevents disease
The Ideal Medication
Has no adverse effects
4

5. Paradox of Modern Drug Development
1. Clinical trials provide evidence of efficacy
and safety at usual doses in populations
2. Physicians treat individual patients who can
vary widely in their response to drug therapy
+ =
+ =
Safe & Effective
Effective
No Response
Adverse Drug Reaction5

6. Anthracyclines
Doxorubicin, Daunorubicin, Idarubicin, Epirubicin,
Valrubicin, Mitoxantrone
Administered to 70% of all childhood cancer patients
Adjuvant chemotherapy for 50-90% of breast cancer
• 22,000 patients/year in Canada
At least 970,000 patients receive it each year (N. America)
Highly effective
Improved childhood cancer survival: from 30% in 1960s to
>80% today
• 1 in 750 young adults is childhood cancer survivor;
6
Kremer LC et al (2004), N Engl J Med 351: 120-121.
Lipshultz SE (2008) Heart 94:525-533.
Altekruse SF et al (eds): SEER Cancer Statistics Review, 1975-2007.

7. Anthracyclines-induced cardiotoxicity (ACT)
1 in 2 patients develop detectable
cardiac abnormalities (57%)
1 in 5 patients suffer Congestive heart
failure
May require intra-ventricular assist
device or heart transplant
Increased risk in children, especially
under 4 years old
ECMO Unit
Heart Transplant
Lefrak EA et al, (1973). Cancer 32: 302-314.
Lipshultz SE et al, (2008). Heart 94: 525-533.
Felker GM et al, (2000). N Engl J Med 342: 1077-1084.
7

8. Aminkeng F et al, Nature
Genetics 2015 Sep;47(9):1079-84
8

9. Research Question
What are the genetic factors that modify the risk of ACT
and can potentially inform treatment decisions?
9

10. Analysis Plan Implemented in SVS from Golden Helix
Step 1: Quality Control (Samples & SNPs) - Exclude poor quality data
Step 3: Replications – Identification of associated gene region
10
Step 2: Discovery - Screening & Prioritisation of Candidates for follow-up
Step 4: Fine mapping - To identify the causal variants within the region
Step 5: Biological Plausibility – Functional relevance of gene & variant

11. EXAMPLE FROM ONE OF OUR GWAS
Stage 1. Quality Control (Sample and SNP) using SVS
11
Before QC: Sample size = 434 patients and 740 000 variants
Detail QC Step takes place right after calling the genotypes
After QC: Sample size = 280 European ancestry patients
and 657,694 variants
Samples
Sample Call Rates
Gender Misspecification
Check
Cryptic Relatedness
Verification
Population Stratification
SNP
SNP call rates
Minor allele frequency
Hardy-weinberg Equilibrium

12. GWAS Work Flow Using SVS
Stage 1. Quality Control (Sample and SNP)
12
Stage 2. Screening and Prioritisation of Candidates for follow-up in
an initial patient cohort
Screened and Prioritized candidates variants in specific genomic
regions for follow-up in an independent patient cohort
Genetic Association Analyses
Genotype distribution
Regression Analysis
Multiple testing correction or select top candidates by p-value
SVS provides a variety of statistical tests to
perform all these analysis which largely
depends your data set and research question

13. GWAS Work Flow Using SVS
Stage 1. Quality Control (Sample and SNP)
13
Stage 2. Screening and Prioritisation of Candidates for follow-up
3. Identifying associated gene regions using another patient cohort
We Identified a new genomic region
Genetic Association Analyses
Genotype distribution
Regression Analysis
May or may not implement multiple testing correction
SVS provides a variety of statistical tests to
perform all these analysis which largely
depends your data set and research question

14. GWAS Work Flow Using SVS
Stage 1: Quality Control (Samples and SNPs) - Exclude poor quality data
Stage 3: Replication - Identifying associated gene regions
14
Stage 2: Discovery - Screening and Prioritisation of Candidates for follow-up
Stage 4: Fine mapping to identify the causal variants within the region
Imputation
Available scripts for import and export of data
for any of the imputation software programs
Data analysis after imputation
Sequencing
SVS can also analyzed sequence
data
The causal or most important variant is studies for the mechanistic basis of the ADR phenotype

15. Stage 1
Discovery
Stage 2
Replication
Europeans from Canada
N = 280 patients: 32 cases & 248 controls
Europeans from The Netherlands
N = 96 patients: 22 cases & 74 controls
Stage 3
Replication
Africans
11 patients
Latinos
23 patients
A. Canadians
15 patients
East Asians
31 patients
15
Patients (Children treated with Anthracyclines )

16. Genotyping: Genome-wide Association Study
 GWAS examines genetic variation across the entire genome
Unbiased and hypothesis-free
Specifically target common variations
Potential to discover novel genes, variants, pathways & inform
drug development
Illumina Infinium HumanOmniExpress assay (738,432 SNPs)
First genome-wide study of anthracycline-induced cardiotoxicity
16a

17. Genome-wide Association Study
We screened with statistical tests implemented in SVS and prioritized with P < 1.0E-05
17

18. GWAS Uncovered RARG as Novel Gene for Cardiotoxicity
Canada
Stage 3 – Replication, Worldwide:
Africans Hispanics
Aboriginal
Canadians
East Asians
Variant P-value
rs2229774
O.R.P-value
9.9 0.012
O.R.P-value
12.3 0.052
O.R.
5.9 0.0859.5 0.026
P-valueO.R.
Variant
rs2229774 > 6 0.00012
P-valueO.R.
Stage 1& 2 – Discovery & Replication, European Patients
The Netherlands
rs2229774 7.0 4.1E-08
Variant P-valueO.R.Gene
RARG 4.1
P-valueO.R.
0.0042
Combined
5.4
P-valueO.R.
1.2E-09
18
Aminkeng F et al, 2015. Nature Genetics Sep;47(9):1079-84

19. GWAS and
imputation identified
putative haplotype
(5 SNP) associated
with ACT
rs2229774 (S427L)
only coding variant
in haplotype
Fine Mapping and Haplotype Analysis
19

20. RARG regulates TOP2B, ACT causative gene
n = 12
*** = P < 0.0001
Normal
RARG
Variant
RARG
Top2bRepression
(Normalized)
-4
-3
-2
-1
0
***
We showed that RARG
transcriptionally regulates Top2b
Top2b critical to development of
ACT (Zhang et al. 2012 Nat Med 18:1639)
RARG variant impaired in Top2b
regulation
20
Aminkeng F et al, 2015. Nature Genetics Sep;47(9):1079-84

21. Genetic Association: Novel gene (RARG), Novel variant
(rs2229774) & Novel haplotype (5 SNPs) for ACT
Functional Validation: RARG & rs2229774 regulates Top2b
expression; Top2b - known ACT-susceptibility gene
Conclusion: RARG rs2229774 is a novel pharmacogenetic
biomarker & provides novel insight into the pathophysiology of
ACT
Summary of Main Findings
21

22. Genetic Association (Drs. Carleton BC, Ross CJD and
Aminkeng F): We are current studying the genetic association
of RARG rs2229774 in adult breast cancer patients
In vitro & in vivo functional studies (Drs. Ross CJD and Bhavsar
AP): Mechanistic studies of RARG & rs2229774 are ongoing &
will inform future drug development in the following ways:
Development of less heart failure prone cancer treatments
Development of more advance cardio protectants
Patients Studies (Dr. Bernstein D): A collaboration with a
Stanford based NIH project is studying the role of RARG &
rs2229774 in cardiotoxicity using real world patient populations
Next Steps: Ongoing Projects

23. Personalized Medicine Project (Drs. Carleton BC, Rassekh RS,
Ross CJD): Pilot project on implementation of PGX Testing
Next Steps: Ongoing Projects

24. What is currently done?
Children’s Oncology Group Long-Term Follow-up Guidelines v3.0
Treatment decisions based on clinical risk factors
24

25. Prediction Based on Clinical & Genetic Risk Factors
Genetic factors improve the prediction of ACT beyond clinical
factors and can potentially inform treatment decisions25

26. Genetic Risk Stratification for Anthracycline-
induced Cardiotoxicity in Children
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0% 20% 40% 60% 80% 100%
Patient Anthracycline-Induced Cardiotoxicity PGx Risk (Percentile)
RiskofCardiotoxicity(%)
14% Risk
21% Cardiotoxicity Risk
39% Cardiotoxicity Risk
45% Cardiotoxicity Risk
89% Cardiotoxicity Risk
(~23% of population.
Risk estimate based upon 139
patients. Includes carriersof
protective SLC28A3 variant.)
(~60% of population.
Risk estimate based upon
356 patients. Includes non-
carriers,and carriers of 1 risk
+ 1 protective variant).
(~13% of population.
Risk estimate based upon 80
patients. Includes carriersof
1 risk variant,or 2 risk +1
protective variant).
(~2% of population.
Risk estimate based
upon 11 patients.
Includes carriers of 2
RARG risk variants).
(~2% of population.
Risk estimate based
upon 9 patients.
Includes carriers of
1+ RARG and 1+
UGT risk variants).

27. Clinical Pharmacogenomic Testing
Avoid adverse drug reactions
Maximize drug efficacy for individual patients
All Patients Treated
with Same Drug
Moderate risk of ACT
Low risk of ACT
Pharmacogenetic Risk Profile:
High risk of ACT
27

28. Genetic Risk Stratification for Anthracycline-
induced Cardiotoxicity in Children
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0% 20% 40% 60% 80% 100%
Patient Anthracycline-Induced Cardiotoxicity PGx Risk (Percentile)
RiskofCardiotoxicity(%)
14% Risk
21% Cardiotoxicity Risk
39% Cardiotoxicity Risk
45% Cardiotoxicity Risk
89% Cardiotoxicity Risk
(~23% of population.
Risk estimate based upon 139
patients. Includes carriersof
protective SLC28A3 variant.)
(~60% of population.
Risk estimate based upon
356 patients. Includes non-
carriers,and carriers of 1 risk
+ 1 protective variant).
(~13% of population.
Risk estimate based upon 80
patients. Includes carriersof
1 risk variant,or 2 risk +1
protective variant).
(~2% of population.
Risk estimate based
upon 11 patients.
Includes carriers of 2
RARG risk variants).
(~2% of population.
Risk estimate based
upon 9 patients.
Includes carriers of
1+ RARG and 1+
UGT risk variants).
PATIENT

29. Importance of SNP & Variation
Suite
29

30. SNP & Variation Suite
Well suited for big & complex data analysis,
visualization & interpretation
Important application for pharmacogenomic
studies to uncover the genetic & mechanistic
basis of drug response
30

31. 10 reasons SVS remains our software of choice
1. Very user friendly and great for beginners
2. Detailed user manuals available which is also easy to use
3. Wide range of statistical tests available
4. Great for data visualization
5. Very fast and have lots of computation power
6. Great variety of data manipulating tools
31

32. 10 reasons SVS remains our Software of Choice
7. Technical support is great & rapid
8. Lots of tutorials and training available on the website
9. Constantly investing heavily in educations & training of its
customers on a wide variety of topics in medical and
biomedical sciences
10. With the amount of genetic data growing faster than
computation capacity of most standard statistical software
packages, SVS is the software of choice, especially for
scientist & clinicians who are neither statisticians nor bio-
informaticians
32

33. ACKNOWLEDGEMENTS
33

34. ST. JOHN’S
Janeway
Children’s
Hospital
HALIFAX
IWK Grace
Health Centre
MONTREAL
Montreal
Children’s
Hospital
MONTREAL
Sainte-
Justine
Hospital
OTTAWA
Children’s Hospital of Eastern Ont.
TORONTO
Hospital for Sick Children
HAMILTON
Hamilton Children’s Hospital
LONDON
Children’s
Hospital of
Western
Ontario
KINGSTON
Kingston
General
Hospital
WINNIPEG
Winnipeg
Children’s
Hospital
CALGARY
Alberta
Children’s
Hospital
EDMONTON
Stollery
Children's
Hospital
Adults: 3 sites-BCCA,
VGH, SPH, KGH, PMH,
SUN
5 MS Sites-UBC, WIN,
LON, HAL, CHUM
VANCOUVER
CFRI/BC.Children’
s Hospital
CPNDS Paediatric
Surveillance Sites
13 Adult Sites
CPNDS Adult
Surveillance Sites
13 Paediatric Sites
• 8 CPNDS
• 5 C17 Sites
CPNDS Network in Canada
34

35. Stanford
(Poole,
Bernstein)
Mexico
(Casteneda,
Rivas,
Medeiros,
ADR
Surveillance)
Indiana University
(Renbarger,
Vincristine)
U Chicago
(Dolan)
S Illinois U
(Rybak)
Baylor & Texas
Children’s
(Berg)
Kansas City
(Leeder)
(Neville, Warfarin)
CPNDS Collaborations
CPNDS Planned/Potential Collaborations
Brazil
(Galvoa)
South Africa
(Cristina Stefan)
African
Populations
(S. Tishkoff)
Zimbabwe
(Collen
Masimirembwa)
Tanzania
(Julie Makani)
Ethiopia
(Beleke Endashaw)
Australia
(Oncology, R.
Conyers,
L. Orme)
New Zealand
(Oncology,
M.Sullivan,
Winstanley)
China
(Oncology, SJS)
Singapore
(Oncology,
Population PGx
Diversity)
Hungary (Oncology, D. Erdelyi)
Germany (SJS, Mackenhaupt)
The Netherlands
(Anthracyclines, Kremer, Caron)
Ireland
(SJS)
UK
(SJS, Pirmohamed)
(ADRIC study)
Uganada
(J. Mukonzo)
Switzerland
(Amstutz)
Taiwan
(Chung, SJS)
International Collaborations

37. Backup Slides
37

38. Anthracycline-induced Cardiotoxicity
many risk factors
Adapted from: Lipshultz et al. Heart. 2008;94(4):525-33
Risk factor Risk
Cumulative dose Cumulative doses >500 mg/m2 associated with significantly elevated long term risk
Length of follow-up Incidence of clinically significant cardiotoxicity increases progressively post-therapy
Radiation therapy Cumulative radiation dose (>30 Gy); prior or concomitant anthracycline treatment
Age Both young and advanced age at treatment are associated with elevated risk
Sex Females are at greater risk than males
Rate of anthracycline
administration
Prolonged administration to minimise circulating dose volume may decrease toxicity;
results are mixed
Individual anthracycline dose Higher individual anthracycline doses are associated with increased late cardiotoxicity,
even when cumulative doses are limited
Type of anthracycline Liposomal encapsulated preparations may reduce cardiotoxicity. Conflicting data exist
about anthracycline analogues and cardiotoxicity differences
Concomitant therapy Trastuzumab, cyclophosphamide, bleomycin, vincristine, amsacrine, and mitoxantrone
may increase susceptibility/toxicity.
Pre-existing cardiac risk
factors
Hypertension; ischaemic, myocardial, and valvular heart disease; prior cardiotoxic
treatment
Comorbidities Diabetes, obesity, renal dysfunction, pulmonary disease, endocrinopathies,
electrolyte and metabolic abnormalities, sepsis, infection, pregnancy
Additional factors Trisomy 21; African American ancestry
38

39. Paradox of Modern Drug Development
1. Clinical trials provide evidence of efficacy
and safety at usual doses in populations
2. Physicians treat individual patients who can
vary widely in their response to drug therapy
+ =
+ =
Safe & Effective
Effective
No Response
Adverse Drug Reaction39

40. 50% of newly approved therapeutic health
products have serious ADRs, discovered
only after the product is on the market
- Health Canada, 2007
40

41. Reports of Severe ADRs are Increasing
Moore et al, FDA, 2007
Serious ADRs
Prescriptions
41