Lorne 2009 profile: Nick Hayward

Around 70,000 years ago, dark-skinned modern humans began moving out of tropical Africa and colonising temperate and sub-Arctic regions of Eurasia.

With the lower angle of the sun at higher latitudes, the dark skin that had protected them against the fierce African sun now became a liability, leaving them deficient in vitamin D, synthesised in the skin under the action of UV light.

Natural selection favoured mutations that reduced melanin levels, increasing vitamin D synthesis. Blonde Scandinavians, red-headed Celts and other pale-skinned peoples are the result of adaptation to life at high latitudes.

Today the light-skinned descendants of the first Europeans migrate seasonally from Scandinavia, Scotland and Germany to the Costa del Sol in Spain, or the Greek islands. Others bake themselves on beaches under an Antipodean sun, unwittingly gambling that they do not carry gene variants that increase their risk of melanoma.

Australians of European descent have the world’s highest rate of melanoma, the most lethal form of skin cancer. Oncogeneticist Dr Nick Hayward, of the Queensland Institute of Medical Research, says around one per cent of Australians carry inherited ‘familial’ mutations that significantly increase their risk of melanoma.

A host of as-yet unidentified deletion mutations, duplications and single nucleotide polymorphisms contribute to a greater or lesser degree to the other 99 per cent of melanomas.

Hayward, who spoke at the Lorne Cancer conference last week, is searching the human genome for variation associated with an increased risk of non-hereditary melanomas.

“We’re taking two different approaches,” he says. “The first is a candidate gene-sequencing approach, which involves making educated guesses about which genes might underlie susceptibility, and sequencing them. These genes lie within two key genetic pathways regulated by the tumour suppressors p53 and pRB (retinoblastoma).

“The other approach involves an unbiased, genome-wide linkage analysis.”

Hayward’s Brisbane team is a member of an international collaboration called the Melanoma Genetics Consortium, or GenoMEL, coordinated by the University of Leeds, that is systematically searching the human genome for melanoma susceptibility genes, and evaluating environmental risk factors that may amplify the genetic risk.

In a recently completed study, the consortium analysed blood samples from 180 families with a history of melanoma. Hayward’s team contributed data from 100 Australian families.

The researchers are looking for genetic markers that occur at higher frequency in individuals who have been diagnosed with melanoma than in unaffected individuals from the same pedigrees.

They will then check to see if any candidate susceptibility genes occur in proximity to ’linked’ markers. The genes will be sequenced to identify any variation that distinguishes them from alleles carried by family members who have not developed melanoma.

“We might do unbiased, comprehensive sequencing of all genes in the key linkage regions using something like the Solexa next-generation sequencing platform from Illumina,” Hayward says.

The Illumina rapid-sequencing system scans libraries of randomly cleaved, short DNA fragments, recording the sequences of the nucleotides at the ends of each fragment. Computers then search for matching sequences to align the fragments and reconstruct whole segments of chromosomes.

The current sequencer works on 36-nucleotide fragments, but Hayward says Illumina is about to release improvements that will generate 75-nucleotide fragments.

---PB--- Founder effect

Some familial genetic disorders in Australia, such as hereditary epilepsies and glaucoma, are the result of a “founder effect” – they can be traced to a single individual who migrated to Australia in the 18th or 19th centuries, when the nation’s population was still quite small.

But Hayward says there is no evidence for a common founder effect with melanoma. “Australia has had fairly steady migration from Europe, so we got founders who represent the entire continent of Europe, especially the UK and Ireland,” he said.

“While Anglo-Celtic immigrants are strongly represented in our population, the pattern of melanoma risk – and the risk for most other heritable disorders – is pretty much what we see in Europe.

“We already know of three genes that account for around 50 per cent of all familial melanoma. These genes have a number of mutations, and we can see where in Europe they originated – they are found among Norwegians, Italians, Scots and French, among others. Assuming there are still risk genes we have yet to identify, we expect they will show a similar distribution of founders.”

Hayward says that several hundred years ago – and even as recently as the early 20th century – melanoma was a rare form of cancer. Its historic rarity is partly due to the fact that melanoma is predominantly a disease of ageing, like most cancers – despite its rising incidence in young sunseekers.

“In the past, life expectancy was much shorter, and people didn’t live to the age where they developed melanoma – or if they did, they didn’t die from it. They didn’t go out into the sun as much, and when they did, they wore clothing that covered more of their skin. If you went swimming at the beach a century ago, you had to wear a neck-to-knee costume.

“In the near absence of exposure to ultraviolet light, melanoma rates are very low. I’m surprised that everyone didn’t develop rickets.”

Hayward says familial melanomas are due to deleterious mutations affecting the protein-coding regions of genes upstream of p53 and pRB, disrupting the mechanisms that would normally force mutant cells into apoptosis.

“These are typical Mendelian mutations, with a dominant pattern of inheritance,” he says. “Some change amino acids so proteins don’t function, others cause truncations or frameshift errors, or in some cases, the gene simply goes missing – basically, anything that can go wrong in the pathway will go wrong.

“One thing that the GenoMEL consortium has been doing is to compare the effect of identical mutations in different continents, using the continents as surrogates for UV exposure.

“It turns out that the lifetime or age-adjusted penetrance of these mutations differs from continent to continent. In Australia, there is a 90 per cent risk of a mutation being penetrant, while in North America the risk is 70 per cent, and in Europe, 50 per cent.

“If you inherit one of these pretty harmful mutations, it’s not inevitable that you will get melanoma. It depends to a large degree on your lifetime exposure to UV light.”

---PB--- SNP associations

Non-Mendelian melanomas, which account for 99 per cent of melanomas, have a complex, polygenic pattern of inheritance, according to Hayward. The consortium is using single nucleotide polymorphism (SNP) association to identifying these genes and to distinguish high-risk from low-risk variants.

The same, complementary approaches are used: search for candidate genes in the region of high-score SNPs to determine whether they contribute to the risk of melanoma, and compare them to null variants.

“We then take the high-risk and normal alleles and compare their effects in people of the same ethnic background, to see which variants of the SNPs are over- or underrepresented,” Hayward says.

“We also use the whole-genome, unbiased approach where we survey SNP markers across the genome, to see which ones consistently associate with melanoma. It turns out that the most robust signals we have identified to date come from one general class of genes that are obvious melanoma candidates – genes that control pigmentation.

“In fact, pigmentation genes are the only ones we and our overseas colleagues have linked to an increased risk of melanoma. There may be others, but the effect of non-pigmentation genes is going to be far less significant.

“The strength of the signals from the pigmentation genes reduces our power to detect other risk genes. We’ve picked the low-hanging fruit, and we’re going to need larger case-control samples to convincingly implicate non-pigmentation genes.”

---PB--- Multigenic risks

In the multigenic area of melanoma risk, the main contributors are pigmentation-related, he says. Biologically, this confirms that the classic risk phenotypes in people are light hair and eye colours, moles on the skin or naevi, pale skin, freckles and the inability to tan.

“We’ve known this for some time, from epidemiological studies. Now we’re finding the genes involved. Putting pigmentation aside – which is clearly the dominant phenotypic risk factor – we can now find other molecular or pathway targets that contribute to melanoma risk.

“If, for argument’s sake, we identify genes in the cell-cycle pathway that contribute to melanoma risk, they might operate as second-tier genes, after the primary melanoma lesion arises.

“It could be that the gene marked by a particular SNP manifests itself only after DNA damage has occurred – we need to know the tipping point at which the risk exceeds a certain threshold, and the individual develops melanoma.

“These are more likely to be candidates for loci that affect the progression or outcome of melanoma, rather than predisposing the individual to melanoma. That’s something we hope to look at down the track, in relation to the outcomes of clinical therapy.”

Hayward says that, given the role of the immune system in detecting and eliminating cancerous cells, it is “highly probable” that the study will detect high-risk HLA haplotypes.

The public health benefits of the research falls into two categories. “First, we’re looking at the basic biology of the pathways that underlie risk, which eventually should give us some new therapeutic options to intervene at different points in these pathways.

“That’s particularly likely to be the case with familial melanomas, because we already know the types of mutations that occur in 50 per cent of families, and there are a number of very promising therapeutic strategies to intervene.

“The familial mutations give us clues to the nature of sporadic mutations in individuals with no history of familial melanoma.

“Once we get away from the pigmentation genes, and begin to identify the lower-penetrance polygenic risk genes, that will uncover new pathways that underlie risk, and should offer possibilities for new treatments.”

A second area of clinical benefit is in risk prediction, he says. If a person has a 50 per cent chance of inheriting a high-risk allele, and an 80-90 per cent lifetime risk of developing melanoma, they can be counselled about the hazards of excessive exposure to UV.

“More importantly, they can be screened regularly, and surgical intervention can occur in the early stage of melanoma, which is the best way to improve their survival and prognosis,” he says.

And with access to gene-testing tools for polygenic loci that contribute to a person’s individual risk of melanoma, clinicians should be able to test individuals for their risk of melanoma.

“If I have a set of phenotypic characters, like freckles and red hair, then a blood test will give me a genotypic set of risk factors, and the two sets of data will predict my absolute personal risk of developing melanoma.

“From there, we can stretch out the gradient of risk across the entire population, and counsel people at the extremely susceptible end of the spectrum about the need for behavioural changes to avoid melanoma.

“These are the people who would eventually benefit from extremely vigorous, targeted surveillance. They would come in for a six-monthly checkup.

“If we can narrow the high risk cohort to a small proportion of the population, for which we have very high accuracy in risk prediction, testing could be deemed to be a medical procedure and subject to a Medicare rebate.”