Swans, black and white

Here, in UCD, our lake houses 2 swans (He, a Cob and She, a Pen). They are white and often, as I pass them, I think of the philosophy of science, and I will explain the link. How do we know something is true?

That the sun rises in the east is a given. It happens every day since time immemorial and thus we can expect it to do so forever. It is a truth. That the truth is defined as something that can be seen time after time, was challenged by Karl Popper an Austrian philosopher. He argued that the truth was best observed from an opposite viewpoint. We do not attempt to show something to be true by means of endless philosophical machinations but rather we can show something to be false. No matter how often the sun rises in the east, we cannot be certain that, one day it will rise in the southeast or east-southeast and so on. However, if ever the sun rose in the east-southeast we could say for absolute certainty that the sun does not always rise in the east. The usual metaphor is the theory that all swans are white. They always have been and they always will be. And then, some explorers on Captain Cook’s antipodean expedition discovered black swans. Thus, according to Popper, the white swan theory was abolished instantly. He saw science progressing in a cycle of conjecture (‘all swans are white”) and refutation (‘we just found some black swans’).  That all sounds fine until one reads the works of an American philosopher, Thomas Kuhn. He argues that science existed at two levels. There was ‘normal science’ and ‘revolutionary’ science. Kuhn regarded the latter as a paradigm shift, a phrase often repeated since.

We used to believe that all swans are white but now we know better. Popper and Kuhn seem to be together at this point. However, whereas Popper believed that this was normal in science, Kuhn argues that this was not in fact normal. It was revolutionary. Kuhn believed that most science fell into this classification of ‘normal science, which was the opposite of revolutionary. Normal science simply defended conventional wisdom. In fact it vigorously defended it. Let us go back to the issue of ‘all swans are white”. The explorers, excited by their discovery, decide to write a scientific paper for the International Journal of Swan Science. The editor sends it out for review and both reviewers reject it, absolutely. One argues that these are not in fact swans at all although they might look like swans. The researchers are asked to conduct extensive genetic sequencing and to re-submit the paper if that is still justifiable. The second referee goes further and argues that indeed they may be swans but they are mutants arising from some pre-mobile phone mast radiation and that they will die off. Basically, the two naïve explorers thought that science was interested in new ideas. It is, provided that the sacred central theory is not challenged. And in this instance, the sacred theory of ‘all swans are white” was not about to be dumped on the basis of some unknown swanologists out in the antipodes. Careers and egos are built on the central dogma and upstart wannabe swanologists should respect that.

Kuhn is correct. Most scientific research is simply filling in gaps of our knowledge in relation to some large paradigm. Up until 10 years ago, we believed in the Central Dogma of Biology, that there was one gene for every protein. With the sequencing of the human genome, we now know that this is not true. That was revolutionary science. Now we populate our scientific papers with explanations of how it all works. That is normal science.

Challenging conventional wisdom is laden with risk to those who would dare to do so. It is not simply that your scientific papers might be rejected. You, yourself, may be pilloried. A classic case is that of Bjørn Lomborg, a Danish statistician who wrote a magnificent work: “The Skeptical Environmentalist”. For his sins he was castigated by the high priests of global warming and suffered disgrace by the relevant Danish scientific integrity watchdog, only to be re-habilitated by his own resolute pursuit of the truth. He now heads a highly prestigious global think tank The Copenhagen Consensus Center. Lomborg simply challenged the way that environmentalists express their concerns. For example, one leading pro-environmental NGO pointed out the spiraling costs of storm damage in the coastal regions of southeast USA. Year on year the economic impact grew and thus, the conclusion was that year on year, the weather was getting worse and storms were becoming more frequent and more severe. All Lomborg did was to show that the value of real estate in this region was growing year by year and moreover, the number of properties per hectare was also growing annually. When he adjusted for the value of properties per hectare, the effect was to show no increase. This is just one of many examples of his challenge to conventional wisdom. In human nutrition, there are many sacred tenets: “Breast is best”, “Obesity is the fault of an irresponsible food industry”, “ The large bowel colonic microflora are central to human health’ etc. To challenge any of these is not easy.  To do so labels one as an agenda laden crank who sees everything in a negative fashion. However, dissent is the oxygen of science and whereas as dogmatism may be suitable for religious and scientific movements, it has no place in science. The outlier is often more revealing that the mean.

The concept of nutritional phenotyping

A little bit of Greek never hurt anybody so lets first look at the term ‘omics’. It is widely used in scientific literature to refer to the new sciences of genomics, which is derived from the term ‘genome’. According to Wikipedia, the term (‘Genom’), coined by a German scientist Professor Hans Winkler of the University of Hamburg in 1920, is from the Greek word ‘I become’. The term ‘ome’ is also of Greek origin and means ‘totality’. From the term genome, came the term genomics, the study of the genome. We then entered the freewheel of ‘omics’ where the study, not just of one protein, but of all proteins measurable within a biological sample, became proteomics. Not to be outdone, those interested in metabolites coined the term ‘metabolomics’ to refer to the science of studying not one but hundreds of metabolites at one time, using pattern recognition technology to seek patterns within the vast amount of data generated. There the matter rested although a proposal for the entry of a new term ‘astrobolomics’ was mooted at a significant scientific meeting on Copenhagen just a few years back. The thinking behind this was that all of these omics could no more predict the future that astrology and that it was all a load of b******.  Hence the proposal for astrobolomics.

All of these omics were eradicating traditional whole body physiology in what can only be described as a technology driven reductionist biomedical Klondike. They had wonderful terms such as ‘knock out’ and ‘knock in’, ‘upstream’ and ‘downstream’, ‘introns’ and exons’ and others, which I always thought would be excellent as the lyrics of some, rap song. Not only were all the omics the only show in town, but when they were blended together they ascended into a new level of mind blowing stratospheric science called ‘systems biology’, in which their blend, through great number-crunching  brontobyte computers (a 1 followed by 27 zeros....forget a google) would lead us to biological Nirvana. This was the future, the autobahn of the new biology.

But as the song goes, with minor adaptations: “And then they went and spoiled it all by saying something stupid like phenomics”. This very new term refers to the very, very old term ‘phenotype’ which comes from the Greek words phainen (‘to show’) and typos (“type”). It really means the study of what you are like: height, weight, eye colour, IQ, fitness, blood this and blood that and just about everything measurable in the human body. To study phenotype is really to study form and not function. The emergence of phenomics is not actually a surrender note from the reductionist hoards on the omics autobahn, just a realization that when you add up all your biology, you ultimately end up with a phenotype. Some of us, blessed with a traditional reverence of the totality of human biology, nutritionists in particular, saw the need for this a long time ago. One could ay that the concept of the nutritional phenotype was born about ten years ago and several people need to be mentioned here for promoting this concept: Ben van Ommen of TNO, head of the European Nutrigenomics Organisation

, Jim Kaput who headed up the FDA’s Division of Personalised Nutrition (now with the Nestle Health Institute) and if I do say so myself, mé féin (copy, paste and Google translate).

In Ireland, a consortium of four universities (Joint Irish Nutrigenomics Organisation: JINGO) received state funding to create a National Nutritional Phenotype Database

. It contains data on several cohorts which have had their phenotype characterised to a remarkable extent (food intake, physical activity, bone density, body fatness, energy expenditure at rest and at exercise, blood this and blood that, muscle function, post prandial function and so on.  In addition we payed homage to the gods of omics and collected complementary data on genomics, proteomics and metabolomics. The difference between this approach and that of systems biology is that we begin with phenotype in either the healthy state or the diseased state and we work back from there. Generally speaking, systems biology builds upwards from genomics, proteomics and metabolomics data to try to understand the mechanisms that lead to disease. Of course it isn’t a competition between systems biology and the construction of major phenotyping databases but the subtle difference is that the latter is driven by phenotype, the former less so and maybe more so now that they have discovered ‘phenomics’.

Such large phenotypic databases need to have several cohorts, central to which should be a large, healthy, nationally representative cohort. This database will always act as the reference database. If you want to know anything about the nutritional phenotype of the Irish, then we have 1,500 such subjects deeply characterised for their phenotype and of course their ‘omics’. Then you need at least one database, which involves a challenge to metabolism. We have two such. One involves a small number (210) who received test meals on two separate occasions and had their metabolic phenotype characterised after either a carbohydrate or fat meal. Almost everybody knows that when you have to attend for a blood test, you usually have to fast from the night before. If everyone were to arrive in at different times, having eaten very different breakfasts, then the interpretation of the blood tests would be confounded by this variation in food intake. So, virtually all the scientific data we have relating diet to health has blood samples measured in the fasting state. This is convenient for the clinician and researcher but it avoids the truth, which is that we eat about 5-7 times a day, and thus we spend most of our day in the postprandial or post-fed state. So, a knowledge of how dietary patterns relate to blood values at fasting is simply a measure of convenience, a means of reducing what is truly complex to a simple and manageable form. Metabolism is asleep in the fasting state and only comes alive in the fed state. Two individuals with identical levels of say fasting blood glucose may behave very differently when given a carbohydrate rich test meal. These test meals really sort out the chaff from the straw in metabolic terms. The second stressed cohort that we have is a very large cohort of older persons: 2,000 with bone disease, 2,000 with impaired cognitive function and 2,000 with high blood pressure.

We have spent the last 5 years building this database, which I equate to the building of a telescope. Now that it is almost finished, we will be equipped to peer deeper into the interaction of diet, genes and metabolism than many others can. 

Shading the sunshine vitamin

Just a few weeks ago, the nutrition community in Ireland gathered in the small town of Limavady in Northern Ireland to lay to rest one of ours, the late Dr Julie Wallace, an outstanding academic at the University of Ulster. Julie was young and in the prime of her scientific career which centered around vitamin D, the topic of this week’s blog and I will draw on a very recent paper of Julie’s in outlining what vitamin D does and doesn’t do. The main function of vitamin D is to facilitate the absorption of calcium from the gut, and then to facilitate its transport from blood into bone cells where it is used in bone growth. The earliest indication of vitamin D deficiency is rickets in children where their normal bone growth is impaired due to inadequate calcium levels, directly arising from poor vitamin D status. When expert committees sit down every so often to pour over the scientific literature to come up with dietary recommendations that ultimately find their way to your packet of Cornflakes, they usually leave vitamin D to the last. That’s because vitamin D is the “sunshine” vitamin. Every cell in the human body makes cholesterol for its own needs and one of the building blocks of cholesterol is acted on by UV light in skin cells leading ultimately to blood vitamin D. In effect, the impact of sunshine on the body’s vitamin D status was generally held by these committees to far outweigh dietary sources, thus downgrading its nutritional importance.

Then, out of nowhere, vitamin D began to climb up the popularity ratings in human nutrition. Firstly, the bizarre finding of inadequate blood levels of vitamin D in sun drenched Australia, began to raise concerns among nutritionists that over-zealous protection from sunrays to reduce the risk of skin cancer might in fact have an unexpected adverse effect. The Aussies developed a major communications programme around “Slip, Slop, Slap” (slip on long sleeved clothing, slop on sunscreen and slap on a hat). Australia has the highest level of skin cancer in the world and their cancer authorities recommend that fair skinned people can get enough vitamin D in summer from a few minutes of sunlight on their face, arms and hands before 10 am or after 3 pm on most days of the week.

This downgrading of vitamin D by the cancer specialists, began to conflict with new findings of a potential role of vitamin D in heart disease.  A significant body of data had begun to emerge ten or more years ago, linking low vitamin D status to the adverse effects of obesity, specifically the “metabolic syndrome” (insulin resistance, impaired glucose function, high blood lipid level and high blood pressure). The vast majority of these studies were “observational” meaning that in an available cohort, people with the metabolic syndrome had generally speaking lower levels of blood vitamin D. Of course, this cannot prove cause and effect. For example, people with the metabolic syndrome are overweight or obese and it could be that obesity was causing the low vitamin D levels and not the other way round. Indeed, that is exactly what a very recent paper from the late Julie Wallace shows and of course it makes intuitive sense

. Vitamin D is a fat-soluble so it prefers a fatty environment than a watery one. Thus the more fat we lay down, the more vitamin D wants to move from blood, which is a watery tissue to fatty tissue. In effect, fat people dilute their blood vitamin D levels raising the question as to whether the overweight and obese need higher than average vitamin D recommendations.

Thus the putative link between vitamin D and the metabolic syndrome and obesity is not looking so good. The only real test comes from an intervention study and myself and my colleagues here in UCD led such a study along with our collaborators in UCC. We gave 160 subjects either a reasonable dose of vitamin D (shown in previous studies to raise blood vitamin D levels within 4 weeks) or a placebo over a 4 week period and we completed this study in two phases: in the sprong just after te darkest part of the year and in Autumn, just after the sunniest part. Vitamin D levels rose in the group given vitamin D but there was no effect on any one of the 14 blood markers of the metabolic syndrome that we measured.  We then decided to rank the subjects into those with very low and very high initial levels of blood vitamin D to see if we could find an effect at the extremes. We didn’t. We then moved into new territory for this type of research using a statistical technique known a cluster analysis. Effectively, you say to the computer (You know that I know that you don’t actually speak to computers so get metaphoric please) to sort the subjects out into groups according to our 14 blood markers of the metabolic syndrome such that subjects within a cluster share a common profile and the different clusters are quite different from one another. We then asked if any one of these clusters responded to vitamin D therapy and one did. It was characterised by low levels of vitamin D in blood AND high levels of two special hormones released into blood from adipose tissue (resistin and adiponectin).

So our data would suggest that there is a subset of the population who do show quite a dramatic reduction in the adverse effects of obesity in response to vitamin D.

This is an issue that will remain on the agenda for some time to come given the intense and opposing views on this topic. What the data does show is (a) that one well designed intervention study is worth a thousand observational studies and (b) future intervention studies will have to classify people into groups which balance not just age, weight, sex and so forth but their genetic and metabolic sensitivity to the intervention. The future just got personal.