Tumour aetiology with animal models
Common cancer risk factors
Genetics – SNPs and cancer genes (e.g. BRCA1, BRCA2)
Age – Most cancers need 4-6 mutations to progress to malignant invasive cell and this takes time.
Infections
Immune surveillance
Lifestyle – alcohol, diet
Environmental exposure – carcinogens
Hormonal exposure – an early full-term pregnancy is protective against breast cancer. Early period + late
menopause increases lifetime exposure to oestrogen and therefore increases risk of breast cancer.
Tissue microenvironment – e.g. udder of a cow and mammary gland of a dog constitute different tissues and
structures.
Genetic variation vs mutation
Genetic variation alters the risk of cancer e.g. chilli con carne metaphor:
Variation = more chilli than normal but is still chilli con carne
- E.g. SNPs which are usually variations in the non-protein coding region of DNA. Individually they have little
effect but cumulatively can greatly increase risk.
Mutations = used pickle onions instead of normal onions; has a massive effect and becomes useless and inedible
- Inherited mutations are rare but have a large effect.
Polygenic model: The lifetime risk of
developing cancer in a population is a normal distribution
most people have a 1/8 chance of developing breast cancer.
BRCA carriers have an 80% lifetime risk of developing breast cancer.
Some individuals only have a risk of 1/20.
Identifying SNPs: The polygenic model is
made from a combination of small risk effects (hazard
ratios) from many low penetrance genes.
Penetrance = the proportion of people who have a
particular genetic mutation who exhibit sign/symptoms of a
genetic disorder; if some people with this mutation do not
have a disorder = penetrance is reduced/ incomplete.
GWAS (Genome wide association studies) help to identify SNPs by studying lots of cases in order to discover all the
small hazard ratios.
Identifying high-risk inherited mutations (Easier than finding SNPs)
Family history studies are done using linkage disequilibrium with
genetic markers.
LD = a non-random association of alleles at two or more loci in a
general population. When alleles are in linkage disequilibrium,
haplotypes do not occur at the expected frequencies.
When two regions of the genetic material are in linkage
disequilibrium there is a reduced rate of shuffling between
those regions during meiosis. E.g. if two points are closely
together on the DNA they are less likely to be separated
You can use LD to identify high-risk mutations because the
statistical power to detect the effect depends on the strength of the This shows linkage disequilibrium because rs265… is unlikely to be
separated from the red gene because they are close together.
effect, of which high-risk mutations have a very strong effect.
Because these single variants/alleles have a very high risk you get large clusters of early onset disease and
therefore need fewer samples to confirm.
LOD score (log of the odds) = statistical score of LD
If the LOD score is >3 that is considered to indicate significant linkage
, Some examples of high-
risk cancer genes:
Founder
effect: Rare, high-penetrance alleles can become more common in
certain communities/populations due to the founder effect (bottleneck
population); e.g. BRCA1/2 mutation is 10% in Ashkenazi Jewish populations.
Finding the causative mutation
High risk:
1. Identify candidate genes closest to linked marker.
2. Typically looking for coding sequence mutation likely to be deleterious (e.g. premature stop codon).
3. Sanger sequencing or shotgun cloning. More recently would be NGS (next generation sequencing).
4. Once potential mutation is identified in index case, you must confirm its presence in all the affected
individuals.
Low risk/SNPs:
1. Unlikely to be a coding sequence variation.
2. Could be in regulatory sequence of a gene – suggesting you have some variation in levels of expression.
3. If you are unlucky you have to use guesswork to identify genes that are nearby. This method only reveals
correlation – correlation does not equal causation!
4. Once you have the candidate gene found by LD studies, SNP or GWAS you must prove the gene is causal in
tumour development.
Pathway to treatment
Correlation = is the variant/mutation associated with the phenotype
Causation = does the variant cause the phenotype
Mechanism = how does it cause the phenotype
Translation = if we interfere with the mechanism, is there a prevention or treatment
Resistance = how might resistance to therapy develop
Animal models can be used to answer these questions.
Role of model organisms
Levels of evidence: in silico in vitro in vivo
Model organisms are vital for: 1. confirming causation and deriving a mechanism, 2. Understanding the biology in
the context of a whole organism, 3. Testing novel approaches (toxicity studies), 4. Ethical and regulatory reasons
(less regulations when testing on veterinary samples).
Mice models
Pros Cons
Mammal but not primate = “more ethical” Not a primate, thus not a brilliant model of
whilst still being genetically similar to humans humans
Genetically tractable (easy to genetically Transplant of human cells requires immune-
manipulate) suppressed animals – not accurate.
Common cancer risk factors
Genetics – SNPs and cancer genes (e.g. BRCA1, BRCA2)
Age – Most cancers need 4-6 mutations to progress to malignant invasive cell and this takes time.
Infections
Immune surveillance
Lifestyle – alcohol, diet
Environmental exposure – carcinogens
Hormonal exposure – an early full-term pregnancy is protective against breast cancer. Early period + late
menopause increases lifetime exposure to oestrogen and therefore increases risk of breast cancer.
Tissue microenvironment – e.g. udder of a cow and mammary gland of a dog constitute different tissues and
structures.
Genetic variation vs mutation
Genetic variation alters the risk of cancer e.g. chilli con carne metaphor:
Variation = more chilli than normal but is still chilli con carne
- E.g. SNPs which are usually variations in the non-protein coding region of DNA. Individually they have little
effect but cumulatively can greatly increase risk.
Mutations = used pickle onions instead of normal onions; has a massive effect and becomes useless and inedible
- Inherited mutations are rare but have a large effect.
Polygenic model: The lifetime risk of
developing cancer in a population is a normal distribution
most people have a 1/8 chance of developing breast cancer.
BRCA carriers have an 80% lifetime risk of developing breast cancer.
Some individuals only have a risk of 1/20.
Identifying SNPs: The polygenic model is
made from a combination of small risk effects (hazard
ratios) from many low penetrance genes.
Penetrance = the proportion of people who have a
particular genetic mutation who exhibit sign/symptoms of a
genetic disorder; if some people with this mutation do not
have a disorder = penetrance is reduced/ incomplete.
GWAS (Genome wide association studies) help to identify SNPs by studying lots of cases in order to discover all the
small hazard ratios.
Identifying high-risk inherited mutations (Easier than finding SNPs)
Family history studies are done using linkage disequilibrium with
genetic markers.
LD = a non-random association of alleles at two or more loci in a
general population. When alleles are in linkage disequilibrium,
haplotypes do not occur at the expected frequencies.
When two regions of the genetic material are in linkage
disequilibrium there is a reduced rate of shuffling between
those regions during meiosis. E.g. if two points are closely
together on the DNA they are less likely to be separated
You can use LD to identify high-risk mutations because the
statistical power to detect the effect depends on the strength of the This shows linkage disequilibrium because rs265… is unlikely to be
separated from the red gene because they are close together.
effect, of which high-risk mutations have a very strong effect.
Because these single variants/alleles have a very high risk you get large clusters of early onset disease and
therefore need fewer samples to confirm.
LOD score (log of the odds) = statistical score of LD
If the LOD score is >3 that is considered to indicate significant linkage
, Some examples of high-
risk cancer genes:
Founder
effect: Rare, high-penetrance alleles can become more common in
certain communities/populations due to the founder effect (bottleneck
population); e.g. BRCA1/2 mutation is 10% in Ashkenazi Jewish populations.
Finding the causative mutation
High risk:
1. Identify candidate genes closest to linked marker.
2. Typically looking for coding sequence mutation likely to be deleterious (e.g. premature stop codon).
3. Sanger sequencing or shotgun cloning. More recently would be NGS (next generation sequencing).
4. Once potential mutation is identified in index case, you must confirm its presence in all the affected
individuals.
Low risk/SNPs:
1. Unlikely to be a coding sequence variation.
2. Could be in regulatory sequence of a gene – suggesting you have some variation in levels of expression.
3. If you are unlucky you have to use guesswork to identify genes that are nearby. This method only reveals
correlation – correlation does not equal causation!
4. Once you have the candidate gene found by LD studies, SNP or GWAS you must prove the gene is causal in
tumour development.
Pathway to treatment
Correlation = is the variant/mutation associated with the phenotype
Causation = does the variant cause the phenotype
Mechanism = how does it cause the phenotype
Translation = if we interfere with the mechanism, is there a prevention or treatment
Resistance = how might resistance to therapy develop
Animal models can be used to answer these questions.
Role of model organisms
Levels of evidence: in silico in vitro in vivo
Model organisms are vital for: 1. confirming causation and deriving a mechanism, 2. Understanding the biology in
the context of a whole organism, 3. Testing novel approaches (toxicity studies), 4. Ethical and regulatory reasons
(less regulations when testing on veterinary samples).
Mice models
Pros Cons
Mammal but not primate = “more ethical” Not a primate, thus not a brilliant model of
whilst still being genetically similar to humans humans
Genetically tractable (easy to genetically Transplant of human cells requires immune-
manipulate) suppressed animals – not accurate.
