A new book claims key genes can make or break our efforts to eat or exercise our way to better health. As direct-to-consumer marketing of genetic tests gathers pace globally, Donna Chisholm explains what the results can and can’t tell us.
Good timing, then, for US genetic-testing company 23andMe, which in April, after a four-year Food and Drug Administration-imposed marketing ban, gained approval to offer direct-to-consumer tests for genes associated with an increased risk of 10 disorders, including Parkinson’s, Alzheimer’s and coeliac disease.
Fortunate, too, for entrepreneurial Auckland skin-cancer doctor and author Sharad Paul, whose new book The Genetics of Health – Understand your Genes for Better Health was released in the US the same month.
Paul’s book doesn’t traverse the genetic links to cancers or dementias, but focuses on how we might tweak our diet and lifestyle according to the genetic hand we’ve been dealt. It explains the evolutionary and genetic associations between our genes and anxiety, exercise preferences, diet, gut, skin pigment, metabolism and responses to food and medication.
“If we share two-thirds of our genes with a lowly worm, nature has given us some genetic wiggle room,” he writes. “With our actions and diets, we can make real changes in our gene expression. Genes are like miniature master chefs, valuing simplicity but giving diners plenty of options.”
The book and its associated website promote a new genetic test he’s developed, which he says will help readers understand how their genes can influence their stress levels, metabolism and sports preferences. “Most importantly, you can learn to eat for your gene type.”
He’s had the test himself, he tells North & South from the United States, where he’s on a speaking tour to promote his book. As a result, he now knows he tends to retain salt, putting him at greater risk of high blood pressure if he has a diet high in salt, but that he has the “good coffee gene”, meaning he metabolises caffeine more quickly, which may reduce his risk of hypertension. And he’s sensitive to saturated fat. “If I eat fat, I’ll get fat.”
Paul says he’s uncomfortable with the book’s title – “I hate it,” he told the publishers – and the way the cover promotes its contents. He would have preferred the title to be The Genetics of Being Human, a subtle but fundamental difference that he says better reflects its tone. Promoting the $US299 gene test was, he says, an “afterthought” in response to early readers of the manuscript who said it would add value. He says it’s not a profitable model, given shipping costs of samples to the US, lab charges and currency conversion.
But the test offering is yet more evidence of the increasing popularity of DNA profiling kits among consumers wanting to know more about their health and risk of disease. So, are they worth the money? And is the information they provide reliable?
With most direct-to-consumer kits testing only 21 genes and around 45 markers, genetic experts told North & South that’s simply too few to provide truly valuable data. Paul’s test covers 21 genes and 35 markers.
“I wanted to include only the ones we have peer-reviewed evidence for. It’s not treatment; it just tells you whether you have a propensity,” he says.
Cambridge University geneticist and obesity researcher Giles Yeo, who’s also a BBC TV presenter, took a similar test as a gimmick, paid for by the UK’s Daily Mail. “It’s bollocks,” he says cheerfully of the 40-page report he received.
He told North & South it would take another 15 years for personalised predictions on health and diet to be meaningful. “They predicted all kinds of things about my diet and exercise, which was all bollocks, but I did get some useful information.” He discovered he’s lactose intolerant and has stopped drinking milk. It’s something he could well have worked out from his own family history and symptoms. His father was severely lactose intolerant but he hadn’t attributed his own, more mild reactions to the same thing.
The problem is, says Yeo, that hundreds of genes and thousands of interactions are involved in complex traits – so tests of most single genes can’t tell us very much (see Of Onions and Six-Toed Cats, page 44). “A lot of it is bumpf, but some of it is sound. Do I have a predisposition to coeliac disease? That’s good, that’s sound. But I already know the answer to that, because I can eat gluten and I’m fine. The lactose intolerance is a single gene thing, so is alcohol sensitivity.”
The test he had done also looked at what types of exercise – power or endurance – suited him better, and purported to measure his aerobic capacity.
“Now, immediately you know that’s complete tosh. From a gene test, they’re going to calculate my aerobic capacity? Are you kidding me? They didn’t even ask my weight. Then there’s antioxidant need. How do you measure that? Omega 3 need. Vitamin B need, cruciferous vegetable need. Are you serious? And this is true bollocks: detoxification ability.
“They give all the right caveats, that this is not a diagnosis, but then proceed to give you a diagnosis.”
He says the tests take population level risk and try to apply it to individual predictions.
“The example I give to medical students is pregnancy. The younger you are, the more likely you are to be able to get pregnant, and the likelihood decreases as you age. That’s very robust on a population level but I cannot take a random 34-year-old woman off the street and, without biological tests, predict if she’s going to become pregnant. But it doesn’t change the fact that for a woman of that age, there’s a 52 per cent lifetime chance of her getting pregnant.”
Without the ability to properly interpret results, users can respond to recommendations in the tests in different ways. One large study published in April, the Food4Me European randomised controlled trial, found people who knew they had the obesity-linked FTO gene reduced their weight and waist measurements more than people who were told they didn’t have it. The findings were, however, at odds with a 2016 meta-analysis from eight other trials that showed no detectable effect on weight loss according to genotype.
At Cambridge, Yeo says tests to determine the potential risk of obesity would cover 100 genes “and I don’t know any direct-to-consumer tests that do 100 genes”. He could have his entire genome sequenced tomorrow at Cambridge.
“I run the facility that does it. I could do it myself. But would I? The answer is no. I don’t know what I’d do with the information. Do I want to know it? My big fear is what if I find some hidden mutation, say a predisposition for cancer, or that I’m going to get Huntington’s disease: will I be able to do anything about it? If I can’t, I’d rather just lead my life and not know.”
The Prime Minister’s chief scientific adviser, Sir Peter Gluckman, is another who could have such a test but won’t, saying the technology is still a decade away from being useful.
He has had a test for a specific disorder, however. When he wed several decades ago, he was tested to discover if he was a carrier for Tay-Sachs disease, a rare and fatal genetic disorder in children. The condition is more prevalent among Ashkenazi Jews. His wife Judy is also Jewish, and both partners must carry the gene before it afflicts their offspring. “In New York among very religious Jews, the rabbis arrange marriages according to who has it or doesn’t have it.”
But he says without a family history of a condition, or having a newborn with an unusual and undiagnosed condition, “I can’t honestly think of a compelling case to get a gene test. We are on the dawn of precision medicine, but this is still very early days. We have known about single gene mutations for a long time and they’re already helping us make better decisions in treatment, for example in cancer. But what precision medicine is going to be about in the future is the small effects of lots of genes.”
He uses the analogy of a mixing console at a rock concert or recording studio, with hundreds of sliding levers subtly changing the sounds. Those sliders are akin to the sorts of mutations that occur outside the genes, in the non-protein-coding area of the genome formerly described as “junk DNA”. A mutation inside a gene, which switches it off, for example, can have a dramatic effect. But an accumulation of many different mutations outside the genes causes much smaller effects.
“There are hundreds of switches affecting your propensity to get obese or get depressed or how fast you age biologically, and most of that is probably not solely genetically determined, but has been affected by environmental changes in your earlier life. The complexity in this area is immense.”
He’s seen one calculation estimating that the genetic information we know currently can explain about three per cent of the risk of obesity. “In theory, if you did the whole genome of everyone in the world, you’d get to about 11 per cent.”
Auckland doctor and scientist Andrew Winnington, who gave up medicine five years ago to set up XY Leap, a gene-testing software company that sells its technology to labs internationally for medical and clinical use, says most direct-to-consumer test kits are grossly underpowered. “It’s doing the industry a disservice. The most powerful part of this technology is drug safety and matching medications to your DNA.”
With his technology, more than 60,000 genetic markers can be tested. Around 6000-8000 markers are used in a test suitable for sportspeople, for example, covering injury prevention, response to hormones and maximising performance. Many businesses, he says, are overselling underpowered tests. “The science behind tailoring your healthcare to your DNA is decades old and super robust. There’s truckloads of evidence in studies where they have looked at people’s genetics, altered their diets, measured parameters and they’ve responded in the way you’d predict, but you need to test far more markers.”
The experts we spoke to say that in terms of your own health, you’d be better doing a little sleuthing through your own family tree rather than spending hundreds of dollars on the tests currently marketed to find out what you may already know.
As Otago University geneticist Professor Stephen Robertson puts it: “If you want a really good idea about what your genes are leading you towards, invest in a full-length bathroom mirror and some high-quality time with your grandmothers on a Sunday afternoon with a full teapot.”
The pros and cons of gene testing for disease risk.
University of Auckland geneticist and genetics researcher Dr Austen Ganley had his DNA taken by US company 23andMe in 2010, before the FDA stepped in to stop the company marketing the health tests directly to consumers. He says his parents and siblings were also tested, because they were interested in their ancestry.
“For me, it was more of a gimmick, a bit of fun. It wasn’t just medical, but every kind of imaginable trait – do you like coriander, do you sneeze when you look at the sun…”
He can’t recall most of his results, and the data was taken offline when the testing was stopped. But he has kept five: the test suggested he was at increased risk of coronary heart disease and psoriasis but reduced risk of Type 2 diabetes, gout and prostate cancer. And has it changed his life? Not one iota.
The test put his risk of heart disease at 69.8 per cent, compared with the average risk of 46.8 per cent: an estimate for which there’s no compelling evidence.
“The risk is calculated from adding up the contributions of many gene variants to risk, and then just giving the sum. For many of the variants, I don’t think there’s strong, validated evidence for the risk (positive or negative) they associate with it.
“The idea that I should be happy with an average risk and do nothing, but should be super-worried and take action if my risk is 70 per cent, seems irrational to me. I suppose it stems from the fallacy that being ‘average’ for anything is okay.”
Ganley says part of the problem with the tests is the language around genetics. “They say you have the gene for this or the gene for that. Everyone on the planet has the gene, so it’s not about whether you have it, but whether you have one with slightly altered function that can potentially lead to other things.”
He can see the benefit for someone in US actress Angelina Jolie’s position in taking a test. Jolie had a family history of breast cancer and had a double mastectomy and her ovaries removed after finding she had the faulty version of the BRCA1 gene, which put her at very high risk.
But should you take a gene test to see what diet and exercise regime is best for you? “I don’t think we have anywhere near enough knowledge to predict that currently. Exercise more and eat better. That’ll help no matter what your genes are.”
Health Revolution! Not So Fast
North & South spoke to scientists at the forefront of using DNA technology in their research. They say revolutionary change is coming but much of their work is raising more questions than answers.
Professor Parry Guilford, cancer researcher, University of Otago.
If you’re the ‘worried well’, you could think about having a look for the 25-30 genes associated with a very high cancer risk. You could have inherited a mutation, but the mutation may have occurred in the sex cells of your parents, so if you’re unlucky, you would be the first in the lineage to have an inherited cancer.
I haven’t had a gene test and I could do it myself, tomorrow. I don’t think I have a sufficient family history to think I have an inherited risk. The information you get from your family history is way more useful than what you’d get from having your genome sequenced.
As you age, you start to accumulate mutations in genes, but until you develop a cancer, they are occurring at a single cell level. Whereas when you take a gene test, you’re testing the DNA you were born with, not the changes which occur since.
Cancer is a multi-gene process, so you might have a mutation in a gene which causes cancer, but alone it won’t cause the cell to divide and cause a tumour until by chance there are two, three or four other mutations that occur in that same cell.
We are trying to develop therapies for patients with the CDH1 gene mutation, which leads to a high risk of stomach cancer – in much the same way as the BRCA1 and BRCA2 genes strongly predispose to breast cancer. We’re trying a chemotherapy prevention approach to try to get a drug which will kill early-stage cancers before we can see them by endoscopy. We have a short list of known drugs, already approved for other applications, and repurposing them. They should be relatively cheap because they’ve already been through all the clinical trials for other uses.
In terms of the future of targeted cancer therapy, the progress is all around drugs like Keytruda, which has shown excellent results in patients with advanced melanoma, and some other cancers. About 25-30 per cent of patients will respond to it, and 15 per cent will have what are effectively cures. It’s really, really dramatic, but what we don’t know is how to choose which patients will respond and which won’t, and that’s particularly important given the cost. There’s also huge interest in trying to work out how to extend the range of drugs like Keytruda to work on different cancers.
Beyond the 30 high-risk genes, we have a very, very poor grasp on the meaning of mutations in other genes. Although they might attribute a certain risk, the errors around those risks are huge and they’re unlikely to be accurate or useful.”
Dr Patrick Gladding, Auckland cardiologist and founder of the Theranostics Lab, which offers DNA testing for coronary artery and atrial fibrillation risk.
Using genetic risk scores to predict coronary disease is still in its infancy and only in the last two years has it evolved to the point where it’s useful clinically. There’s emerging evidence to support that, particularly with regard to how statins work. If you have a high-risk score, you’ve got a much greater likelihood of a statin being effective for you than if you’re at low risk. You can’t really gauge how useful they are until you’ve put them into practice. It’s changed decisions I’ve made in treating patients, but I always give them the caveat that this is pretty early-stage technology.
I have, for instance, decided not to give statin treatment to some people whose overall risk is low, even when their cholesterol is high; the counter also occurs – where you have people who might have borderline elevated cholesterol but a high risk overall, you’d perhaps treat them more aggressively. But this requires a person who understands the score and what it means, and can have a frank discussion with the patient about that.
We’re not yet at the point of using genetic risk as a screening tool, but there’s actually literature that suggests, depending on the number of genes used, you can get to a very high level of accuracy.
In private practice, since 2013, I’ve also been testing patients for a rare variant of the CYP2C19 gene, which makes a commonly used cardiac medication (clopidogrel) ineffective. The variant is more common in some ethnic groups, including Maori. One patient came into my care after suffering a cardiac arrest following multi-vessel coronary stenting, and being resuscitated. I tested him retrospectively and found he had the variant. He was put on the right medication for him.
Apart from this, at the moment there’s no evidence that one gene is going to be able to have a major influence on how you do things. The nature of the predictions, particularly those in test kits marketed by places such as 23andMe, is very weak. They tend to focus on one variant and presume that’s the predictive gene for that condition, when there is a lot we don’t know, because it’s testing only 21 genes.”
OBESITY, DIABETES & SMOKING
Professor Terrie Moffitt, psychologist and co-director of the Dunedin multidisciplinary health and development study
We’ve reported that people who carried more gene variants for obesity had their first big weight gain between babyhood and age three, and if they did become a pudgy baby, their chances increased of them being overweight in their 30s. So, if you knew your (or your child’s) polygenic obesity score, you might take care with sweets and sugary drinks in toddlerhood, and work hard to teach your kids healthy eating habits. In the study of smoking, we reported that people with a high polygenic score for nicotine addiction went extremely rapidly as teenagers from their first experimental puff on a cigarette to being a daily smoker. In contrast, those who can smoke at parties without ever getting hooked had the lowest polygenic scores, even lower than normal. It could be useful to know your polygenic score for nicotine addiction, especially if you could know it before the age when most people try cigarettes, which is in early adolescence. But teens are notoriously immune to health-risk information.
However, there are big caveats about this polygenic-score work. First, having your whole genome tested and the polygenic scores made up for you is not cheap or readily available to the public. Second, the scores have only been worked out for white European ancestry so far. Third, the statistical-effect sizes of the polygenic-score associations with outcomes are uniformly very small indeed, so genetic information is still a very minor and even trivial piece of information, as compared to other things you might know about yourself. Fourth, in Dunedin, we compared the polygenic score to simply knowing one’s family history, which is free. Both polygenic score and family medical history predict your own health outcomes about the same, which is modestly, but better than zero. If you want to know about your risk of addiction, obesity, heart disease, depression or many other diseases, just sit down with your mum and ask about your family history.”
Dr Justin O’Sullivan, molecular biologist and obesity researcher, Liggins Institute and University of Auckland.
In diabetes and obesity, certain parts of the body are more affected by genetic variants that alter the dimmer switches – the thyroid gland, for example, as well as the oesophagus, the lung, the subcutaneous fat, the skeletal muscle and the tibial artery and nerves.
In terms of the subcutaneous fat, we think the variants reduce the ability of the fat to moderate and store the lipids circulating in the body, so the excess spills over into visceral fat, which is bad. The thyroid is a major regulator of metabolism. We don’t understand the roles of the genes in all the different pathways, but we do know some affect insulin signalling and the relationship between glucose and insulin signalling. We also know some affect leptin (the so-called satiety hormone).
If we start to integrate these things into understanding how different tissues are affected, it could lead us to novel approaches for diagnostic prediction and treatment, or even prevention.
The FTO gene variant, which has been linked to obesity, is very interesting, because the changes don’t actually control the FTO gene itself; they control another gene that is a long way away.”
Professor Stephen Robertson, CureKids professor of paediatric genetics, University of Otago.
For complex traits, about 50-60 per cent of the predisposition is attributable to genetic influence. Cardiovascular disease is in that ballpark, but conditions such as leukaemia are down at one per cent. However, with some disorders, such as autism and schizophrenia, it’s around 70-80 per cent.
By looking at genes, you’re only getting a partial view – it’s predisposition, not determination. We are given this loading and what’s important is how it plays out in the environment in which we live. There’s a lot of difficulty in people interpreting that.
I haven’t had my genes tested, because I’m one of those boring people in the middle of the curve who has an unremarkable family history of late-onset diseases of indolence and excess, which is going to kill me, as opposed to something early-onset where there looks as if there might be some sort of genetic loading. If I had a strong family history, I think we’re moving to the point people might get useful information out of it, for things they can act on and modify. But If I learnt tomorrow I had a predisposition to early-onset dementia, there’s not too much I can do about it, so the conversation is different.”
Geneticist professor Russell Snell, Centre for Brain Research, University of Auckland.
The most common gene associated with late-onset Alzheimer’s is the APOE gene. If you inherit one copy of the predisposing allele or variation, the E4 variant, your risk is increased approximately fourfold above the population average. If you inherit two copies, it’s eightfold.
In the special case of early-onset disease, we know that mutations in the amyloid precursor protein, the presenilin 1 and presenilin 2 genes, absolutely cause Alzheimer’s because we have seen it transmitted in families and found the mutations.
About seven or eight other risk-conferring genes have been found and more are popping up as studies get larger, but they will have less and less effect than those we’ve already found.
There’s a lot we don’t know about how genes and the proteins they code for interact. We can test for this variation or that variation but we know in the majority of people the body and the brain are influenced by these interactions, not just a single cause. What we’re really useless at is saying, is the condition of this gene and that gene more predisposing than this one alone? We’re not good at it because studies aren’t large enough and mostly people don’t ask those questions.
Patients being recruited into our dementia-prevention research clinics are being genetically profiled, so this will be the first time we get this information on a large scale.
The decision to be tested is incredibly personal. There are 100 reasons to be tested and 100 reasons not to be. I’ve seen people who want to be tested because they want to know how life is going to turn out and it’s the best thing they ever did. Others say they tend to live every day as if it’s their last anyway so it doesn’t matter if they know or not.
We need to safeguard genetic information so it belongs only to the individual. This protection is becoming an imperative with the cost of DNA sequencing rapidly decreasing and our knowledge of genetic effects rapidly increasing. In my opinion, the state or companies have no right to ask for or have access to this information and this should be protected in law. Without this protection, we could easily end up with a genetic underclass.”
GENETIC TESTING: FACTORS THAT MATTER
Who should consider genetic testing for cancer risk?
The US National Cancer Institute says it should be strongly considered when all three of these criteria are met:
1) The person being tested has a personal or family history that suggests an inherited cancer risk.
2) The test results can be adequately interpreted, i.e. they can clearly tell whether a specific genetic change is present or absent.
3) The results provide information that will help guide future medical care.
The institute says features of a person’s history that, particularly in combination, may suggest a hereditary cancer syndrome include:
- Cancer at an unusually young age.
- Several different types of cancer occurring independently in the same person.
- Cancer that has developed in a set of paired organs, such as both kidneys or breasts.
- Several close blood relatives that have the same type of cancer, for example, a mother, daughter and sisters with breast cancer.
- Unusual cases of a specific cancer type, for example, breast cancer in a man.
- The presence of birth defects, such as non-cancerous skin growths or skeletal abnormalities known to be associated with inherited cancers.
- Being a member of a racial/ethnic group known to have an increased chance of having a certain hereditary cancer syndrome and having one or more of the above features.
This was published in the June 2017 issue of North & South.