The Problem with Nutrigenomics

"Personalized nutritional counseling is a burgeoning field. Several companies, including Vitagene, Nutrigenomix and DNAFit, are already offering individualized dietary counseling.  Their efforts are based mostly on genetic testing, but scientists have only just begun to explore the links between DNA and good nutrition. “I think companies offering personalized dietary advice are probably running ahead of the evidence,” said John Mathers, director of the Human Nutrition Research Center at Newcastle University in Britain." [NYTimes Blog]

Introduction

Science skeptics have recently reviewed these services, and found plenty of quackery. [Science Based Medicine] [Skeptical Raptor]

However, there is some research to back up the idea that genetic testing can provide insights into metabolic disorders.  There may be as many as 200 SNPs for which there are proven metabolic effects, and only a subset of these alter nutrient requirements in a significant portion of the population [e.g., the rs1801133 MTHFR SNP and folate requirement in 15–30% of the population (Solis et al., 2008) and the rs12325817 PEMT SNP and choline requirement in 20–45% of the population (da Costa et al., 2006)].

What do genetic testing services measure?

The human genome consists of about 3 billion nuceleotide bases, each of which is either A,C, T, or G.  A reference genome based on the similarities of all genotyped humans has been assembled, along with a corresponding reference database of all of the point "mutations" where individuals differ from that baseline.  So far, scientists have documented about 150 million of these single nucleotide polymorphisms, called SNPs.  The average person doesn't have all of these differences, however; most people have about 3 millions SNPs that differentiate them from the reference human genome.   Since 3 million is about 0.1% of the 3 billion base pairs, most humans differ from each other by about 0.1% of our DNA. Ancestry genomic testing services like 23andme test for a few hundred thousand SNPs for $100-$200.  For $1000-10,000 a complete genomic sequence can be obtained.  (there are about 3 billion bases total.  

Examples: MTHFR – metabolic pathways and nutrigenomics

In humans, SNPs in the gene MTHFD1 increase the demand for betaine as a methyl-donor, thereby increasing the dietary requirement for choline. Another SNP in the gene PEMT prevents the activation of this gene by estrogen, thereby decreasing endogenous production of phosphatidylcholine (a source of choline) in the liver and increasing the dietary requirement for choline. [Choline: Critical Role During Fetal Development and Dietary Requirements in Adults. Ziesel]

But note that it is not so simple.  There are several forms of the MTHFD1 gene, for example MTHFD1L and MTHFD2. If MTHFD1 is commonly mutated, it may be a pseudogene.

Complexities interpreting SNPs

No simple test can unravel the intricacies of the human genome, and consumers should be suspicious of anyone claiming to be able to interpret measurements of tens of thousands of genes, with millions of genetic variations, some of which have effects on hundreds or thousands of the small molecules of metabolism (and perhaps on thousands of peptides or proteins involved in metabolism).

[A grand challenge for nutrigenomics.  Steven Zeisel. 2010.]

Mistakes in genomics data

Note that 23andme data, like any large genome scan, can have mistakes in it.  For example, the Enlis genomics blog found more than 500 likely mistakes in a sample of 23andme raw data!  (Enlis)

Furthermore, many important nutritional SNPs are not testing by 23andme.

Solution: Metabolic Testing
There is a genetic test for MTHFR variations. But there’s also a cheaper and more accurate way to test for whetherMTHFR variations are causing disease. We simply check the levels of homocysteine in the blood...In other words, the homocysteine levels determine our actions, not the MTHFR test results.[Cleveland Clinic]

Food4me study tested Nutrigenomics...and found no benefits

Food4me is a large online study designed to test whether personalized nutrition advice based on analysis of phenotypes (waist cicrumference, blood markers: glucose, cholesterol, carotenes, n-3 index ) or genotypes (SNPs in genes such as MTHFR, FTO, TCF7L2, APOE E4, FADS1 ) could perform better than standard nutritional advice.  The study recruited more than 1,600 volunteers from across Europe to take part.  Participants performed quantified health self-analyses such as biometric measurements and movement counts. They also used a do-it-yourself blood sampling technique that involves drying blood from a finger prick on absorbent paper, which can then be analyzed for more than 92 metabolic biomarkers in a lab.  They also submitted saliva samples that were checked for more than 36 genetic variants that have been linked to nutritional needs and health outcomes.

"A scientific knowledge base was developed, capturing the current knowledge in the field of nutrition
with a particular focus on the interaction of food consumption, nutrient intakes, biomarkers,
genetic variation to health. SNP information comprises risk allele frequencies as well as gene
symbols and functions. The collected scientific knowledge represented in the data base covers
currently 35 food items, 92 biomarkers, 36 genetic variations, 16 different health outcomes, and
180 established interactions based on scientific publications and an expert assessment."

After one year, the results are in.  Although their internet-based nutritional intervention was associated with positive outcomes, a recent whitepaper concluded that, after testing various diets, there were no improved health outcomes from phenotypic or genetic information.

The researchers state that, "despite enormous efforts over the last decade to identify gene variants that define the susceptibility of an individual to a life-style dependent disease, the outcomes of the large-scale profiling studies are rather disappointing. Although a large number of genes and variants have
been found (there are for example around 60 genes that carry a susceptibility risk to develop
type 2 diabetes mellitus (T2DM)), the effect sizes of each individual gene variant are generally
very low. In almost all cases, the risk-variant increases disease risks by only a very few percent..."

Complex Science: The Role of Vitamin D Receptor (VDR) Genotype Polymorphisms (SNPs)

Vitamin D regulates the expression of hundreds of genes, with widespread hormonal and immune effects.  But whether vitamin D is good for you may depend on your genes.  According to one hypothesis, supplemental vitamin D causes allergies and asthma.  [1]

But the biochemistry is complex.  The main circulating metabolite is 25-hydroxyvitamin D or 25(OH)D, a biomarker of vitamin D status.  The active vitamin D metabolite 1,25(OH)2D3 binds to nuclear vitamin D receptor (VDR), which exists from under 500 to over 25,000 copies per cell in many human tissues including thymus, bone marrow, B and T cells and lung alveolar cells.  Gene expression can be varied over a 100-fold range by subtle modifications of introns and promoter regions outside of the gene[1]

The SNPs that seem to affect VDR are not in the exon; they may affect RNA production in the promoter region. It is unlikely that increased or decreased vitamin D sensitivity is simply mediated by a genetic variation in the VDR. Vitamin D requires several enzymatic steps to be activated, transported and degraded; receptor signalling requires several co-factors and all of these may contribute additive or multiplicative effects on vitamin D sensitivity.

Possibly because of this complexity, progress in this field has been slow.  I reviewed several papers, most of which found very small or no effects from common SNPs.  For example, although most papers found insufficient or deficient levels of vitamin D throughout the population, a case-control study only found a small effect on circulating 25(OH)D from one of the SNPs tested. [2].  A randomized controlled trial found effects from more SNPs, but each contributed very small effect sizes. All SNPs tested had, at most, +/-5% effect on circulating 25(OH)D. : "Three SNPs had statistically significant interactions: rs10766197 near CYP2R1, rs6013897 near CYP24A1, and rs7968585 near VDR, with per allele effect sizes ranging from −4% to +3% differences in [25(OH)]."[3]

These complex and unimpressive results are representative of the difficulties inherent in assaying SNPs for clinically-relevant phenotypes.  Most SNPs slightly modify expression or binding of a protein, such that it takes the combination of dozens or hundreds of different SNPs to create any significant phenotype.  Biochemistry is complex, and it is always possible that other gene or protein interactions can ameliorate or exacerbate any small perturbation from any given SNP.

Citations:
[1] Variants in the vitamin D receptor gene and asthma.  2005.  

[2] Vitamin D levels and vitamin D receptor gene polymorphisms in asthmatic children: a case–control study.

[3] Genetic Variants in CYP2R1, CYP24A1, and VDR Modify the Efficacy of Vitamin D3 Supplementation for Increasing Serum 25-Hydroxyvitamin D Levels in a Randomized Controlled Trial.

Can Patients Understand their Own Genome?

I just ran my 23andme SNP data (Single Nucleotide Polymorphism: basically, the distinct mutations that make my DNA unique) through geneticgenie.com, a website that puts the number and type of mutation in a handy table.  The website also provides nutritional recommendations based on the presumbed metabolic impact of my particular mutations.

However, after feverishly researching biochemistry I have some concerns with Dr. Yasko's conclusions cited on that site and others. These websites appear to make a number of biochemistry mistakes, and I'm not seeing a lot of citations to original research, just a lot of unpublished "physician observations".



A selection of results from G enetic Genie. There are two copies of most genes in our genomes (one from our father, one from our mother) and one or both may be mutated. The color-coded results show that I have two mutated copies of several important genes (colored red) involved in neurotransmitter metabolism and other core biochemical processes. I also have two genes with one bad copy (yellow),


Some of the statements about, for example, BH4, appear to be incorrect. Genetic Genie states that impaired BH4 production or increased BH4 utilization can impact ammonia detoxification in the urea cycle, but BH4 is not directly involved as a cofactor in ammonia to urea conversion. Instead, BH4 is involved in one of at least two pathways for generating citrulline. (Citrulline is regenerated in the urea cycle to turn ammonia into urea.)


Not to say they're not doing good work, but you have to interpret biochemistry in context. For example, I am homozygous for a mutation in CBS, which they say would upregulate CBS activity and lead to increased cystathione, cysteine, and eventually to increased taurine and sulfite. But I also have a heterozygous mutation in CTH, which would limit the amount of cystathione converted into cysteine, effectively stopping that cascade at the starting line.

I hope we're just a short ways off from a website or interface that can actually map all of our unique (SNP-dependent) metabolic pathways, but I think we're still in the dark ages when it comes to interpretting SNP genome results. Promethease is the online tool that has replaced 23andme's health-specific genetic information, but the website only summarizes Pubmed results:



The Promethease website is great, but is based on observational studies with tiny effect sizes. Trying to infer causation from those correlational studies is a textbook example of how not to interpret statistics.

Faced with the complexity of ~20,000 SNPs and less-than-user-friendly professional tools like ENSEMBL, I don't think it is possible for individuals to understand how SNPs influence protein function to the extent necessary to make informed decisions about our biochemsitry.

Core biochemical methylation pathways.

Two common mutations can cause decreased levels of BH4. The first mutation increases the activity of CBS, which converts Homocysteine to Cystathione and eventually to cysteine and then taurine.  The second mutation is directly involved in the regeneration of BH4 in the methylation pathway.

But before I get ahead of myself, why is BH4 so important?

Tetrahydrobiopterin (BH4) has five major responsibilities as a cofactor.  It is needed to work with:

Tryptophan hydroxylase (TPH) for the conversion of L-tryptophan (TRP) to 5-hydroxytryptophan (5-HTP)
Phenylalanine hydroxylase (PAH) for conversion of L-phenylalanine (PHE) to L-tyrosine (TYR)
Tyrosine hydroxylase (TH) for the conversion of L-tyrosine to L-DOPA (DOPA)
Nitric oxide synthase (NOS) for conversion of a guanidino nitrogen of L-arginine (L-Arg) to nitric oxide (NO) in the Urea Cycle
and
Alkylglycerol monooxygenase (AGMO) for the conversion of 1-alkyl-sn-glycerol to 1-hydroxyalkyl-sn-glycerol

The first three reactions are critical to producing adequate levels of serotonin and dopamine. Low levels of BH4 can impair neurotransmitter production and lead to the build up of toxic intermediates, like phenylalanine.



This chart shows that without BH4 phenylalanine (found in all foods) is not converted to tyrosine. Tyrosine one of the 22 amino acids used to build protein and is normally non-essential. It is found in most foods.

BH4 is synthesized in four ways:



BH4 is involved in the major biochemical cycles:



Methionine is a major source of sulfur groups in the diet, so limiting methionine-rich foods helps limit problems from an overactive transulfuration pathway (i.e. overactive CBS enzyme). Other sulfur rich foods include crucifer vegetables and onions and garlic.

BH4 is tangentially involved in both the urea cycle and the folate cycle.

The urea cycle in detail: Arginine from our diet or from protein metabolism is converted to ornithine and urea by the enzyme Arginase. Ornithine is then converted to citrulline by ornithine transcabamoylase (OTC). Citrulline is converted back to arginine. This cycling of Arginine through the various intermediates is what converts ammonia to urea. (More info)


Source.

Arginine is also required for the production of Nitric Oxide (NO) by the enzyme nitric oxide synthase (NOS or eNOS). This reaction is dependent on the levels of BH4 available from the BH4 cycle. Remember two molecules of BH4 are needed to generate Citrulline and NO. One molecule of BH4 will in turn generate peroxynitrite and if there is no BH4, super oxide is formed. (Source. )

The Urea Cycle and the Nitric Oxide Cycle are interconnected by arginine as follows: Citrulline, made from ammonia (and ornithine), is recycled to arginine in the Urea Cycle. That arginine can then enter the Nitric Oxide Cycle where it is converted to nitric oxide by nitric oxide synthase using tetrahydrobiopterin (BH4). (Note that citrulline is also generated during the formation of nitric oxide from arginine.)

From this we can conclude that a deficiency in BH4 does not impact ammonia detoxification in the urea cycle.