Three things high folic acid does to animals

Dr Clare Craig

Rodents handle folic acid far more efficiently than humans but nevertheless they see three clear types of harm: altered brain development, weakened immune defence and faster growth of tumours that already exist. There is twenty five years of evidence on these topics.

Most of the evidence base regarding folic acid fortification is either epidemiological associations, which are often confounded or biochemical mechanisms, which do not always play out in real life. Both are helpful in their own way. The missing link between them can be partly filled by animal work. There is an extensive body of work using controlled feeding experiments in mice and rats that has not been referenced in government decision making. Animal studies have used randomisation with folic acid specifically added and outcomes directly measured.

Why not stick to epidemiological evidence?

There are major differences between the women who take folic acid supplements before and during pregnancy and those who do not. They are older, more educated, richer, more likely to have planned the pregnancy, to have booked early antenatal care and less likely to smoke. Each of those is independently known to relate to a better outcome for the baby. It becomes impossible to know whether it was really these factors affecting an outcome or a supplement that the group happens to take disproportionately. Causation can only be shown in a clinical trial.

Randomisation in a trial removes the problem by allocating every type of person equally to receive or not receive the supplement. Most of the human trials that established the neural tube defect benefit failed to follow the mothers and children for endpoints beyond birth. The pregnancy trials that did follow up over years are small, for example the FASSTT trial 2005-2007, described in a Nutrition Society review as the only randomised trial to date to test 400 micrograms of folic acid without other vitamins. It was an odd trial. Women were enrolled who had already taken folic acid in the first trimester and were then randomised to have more from the fourteenth week of gestation. The children were assessed at intervals up to eleven years. They were given cognitive tests, and the supplemented group came out ahead on one out of eight tests at each age, a different test each time: general cognition at three years, word reasoning at seven, then two speed of processing tasks at eleven with a verbal score in girls alone. Nothing else was significant. Only 39 children were tested at three years, 70 at seven and 68 at eleven. The longest follow up was the Aberdeen trial, run in the late 1960s, which measured only DNA methylation in saliva in 86 adult offspring at a mean age of 47. Methylation is the chemical tagging that switches genes on and off. The tags had shifted in the folic acid group and were still shifted 47 years later, across 46 regions of the genome including two immune genes. The question as to the consequence of this altered methylation remains open. In the meantime, there has been no study that has followed up the children to report on immune function, susceptibility to infection or cancer. Those are the three endpoints for which the animal work shows clear safety signals.

The dose objection

Much of the animal work has been dismissed on the basis that the doses are too large to translate to humans. The studies used five, ten or even twenty times the amount recommended for rodents. This dismissal is not reasonable. Rodents metabolise folic acid far more efficiently than humans. To replicate the issue of circulating unmetabolised folic acid a higher dose is necessary.

To be specific, human liver handles natural folate at around 3 percent of the rate a rat manages and folic acid at under 2 percent, and within human liver folic acid is processed 1,300 times less efficiently than natural folate. Unmetabolised folic acid is a serious problem as this is a synthetic drug. Random testing found it present in 95 percent of US citizens exposed through fortification of flour. The authors of one study stated they thought they might in fact be underdosing the rats in terms of the amount of unmetabolised folic acid that was circulating.

The rodent doses are given as milligrams per kilogram of feed, not per kilogram of animal so the researchers aligned them to human dosages as a proportion of the daily requirement. The daily requirement of 2 mg per kg of feed in rodents was taken as equivalent to the official human daily requirement of 400 micrograms a day. Both figures are likely overestimates with the rat figure potentially ten times too high and the human figure eight times higher than human daily folate losses. A 5 mg per kg dosage  is 2.5 times the RDA and equivalent to the 0.8 to 1.0 mg a day which would be the typical intake once fortified food and a supplement are combined. That is also the dose of folic acid used in the Czeizel trial on which the whole policy depends. A 20 mg per kg dosage raises plasma folate about 2.5 fold, the same rise measured in women taking 4 mg a day, and critically it is the dosage at which unmetabolised folic acid becomes detectable.

There is no equivalent conversion

Rodents differ from us in multiple relevant ways:

• Processing. Human liver reduces folic acid at under 2 percent of the rate measured in rat liver and human activity varies fivefold between individuals whereas rat activity varies less than twofold.
• Metabolic rate. A mouse burns around seven times more energy per kilogram than a human and a rat around four times. Nutrients per kg are therefore disproportionately high at baseline.
• Spiking. Rodents feed through the night while a person has occasional meals and the spike in uptake in humans readily saturates the enzyme in a way that is unlikely to occur in a rodent.
• Other sources of folate. Gut bacteria in rodents synthesise folate and they eat their own droppings which recycles folate. Both these lead to higher levels of folate, rather than folic acid.
• Liver renewal. The enzyme that processes folic acid, dihydrofolate reductase, is switched on as liver cells prepare to divide. Somewhere between 0.1 and 0.3 percent of rat cells are dividing at any one time whereas for humans it is more like 0.001 to 0.01 percent.

With all these variables there is no equivalent dosage for a rat or mouse that matches the human dosing. The only fair comparison is how much unmetabolised folic acid is present after dosing.

When mice were fed dosages ranging from 0 to 40 mg per kg of feed plasma unmetabolised folic acid was not detected at 5mg and rose significantly only at 20 and 40 mg per kg. There was no measure  between 5 and 20 mg. Inevitably the folate levels were high as a consequence and the fact the folate levels were higher than what has been seen in humans is used to dismiss the findings. Any model attempting to measure the effects of unmetabolised folic acid is bound to end up with excess folate in a rodent. Until a model that somehow keeps the folic acid in an unmetabolised state is produced, this is the best evidence we have of the harm from unmetabolised folic acid.

Brain development

When female mice were given either 2mg or 20mg of folic acid per kilogram of feed from weaning and therefore through mating, pregnancy and lactation gene methylation was altered. The diets were identical in every other respect. They tested the mothers, then the embryos, then the offspring. In the high dose group the mothers’ cells produced less of the important enzyme MTHFR which is needed to use folates in methylation pathways that control DNA and RNA gene switching. Levels of the MTHFR enzyme fell threefold and what remained was more likely to be in an inactive form. The authors named it pseudo-MTHFR deficiency, because a diet had produced the same shortage that some people inherit as a genetic variant.

Late in pregnancy the embryos and placentas weighed less. Embryo livers had lost MTHFR activity too. There is a backup route the body uses to supply methyl groups when the folate route is blocked and two key chemicals for this route, betaine and phosphocholine, were reduced showing this backup was being relied on.

The three-week-old male pups showed evidence of impaired short-term memory and had physical changes in the brain with a smaller hippocampus and thin dentate gyrus. The hippocampus is the brain’s memory structure and the dentate gyrus is the layer inside it where new nerve cells are made. The pups also had less Dnmt3a, the enzyme that attaches methyl tags in maturing nerve cells.

The same group had already shown that in adult male mice 20 mg of folic acid per kg of feed for six months produced the same enzyme shortage with disturbed fat metabolism and liver injury. A separate paper found that at half that dosage folic acid cut the capacity to supply methyl groups in placenta and embryonic liver. Higher doses led to embryo loss, developmental delay and defects.

There is a paradox in that mice mothers who lacked MTHFR gene had better outcomes suggesting that the methylation pathway is not the full story. The brain development issues have been replicated with cortex growth and nerve branching issues seen in both low folate and high folic acid exposure. It is known that folic acid can block folate uptake into brain cells so this is not a contradictory finding.

Immune protection

Natural killer cells are white blood cells that find and destroy virus-infected and cancerous cells in a non specific way i.e. without relying on immune memory. They are the body’s first line of attack on both viruses and cancer.

Aged female mice given folic acid had natural killer cells that killed less effectively and fewer of these cells had matured into the killing form. That experiment was carried out after observations in humans suggested a link. Postmenopausal women with detectable unmetabolised folic acid in their blood had been found to have reduced natural killer cell activity. The mouse study reproduced this finding under randomised conditions.

Mice with malaria on a folic acid supplemented diet died sooner and carried more parasites. They had fewer immune cells of various kinds and more inflammatory proteins present in the brain.

A further human study confirms the finding. Thirty healthy Brazilian adults given 5 mg of folic acid daily for 90 days showed rising unmetabolised folic acid in the blood which tracked falling natural killer cell activity. There was no control group, so it cannot exclude drift over three months, but it was a marked difference.

In a laboratory adding folic acid to human immune cells had no effect. That suggests the problem is not folic acid directly but something upstream in how those cells mature.

There is one human dataset on the effects on infection. Among 380,380 UK Biobank participants those with a GP prescription record for 5mg of folic acid were about 1.5 times more likely to be diagnosed with covid and 2.6 times more likely to die of it. Those taking it would have been sicker to begin with. The authors tried to adjust for the medical conditions they could identify and the association remained. Ultimately, the question has not be properly answered.

Tumour growth

The key study on tumour growth took female rats given a chemical carcinogen at puberty. They were fed low dose folic acid throughout until one breast tumour grew to about 8 mm across. At that point they were randomised to 2, 5, 8 or 10 mg folic acid per kilogram, with 44 animals per group, for twelve weeks. Those on higher doses all ended up with bigger and heavier tumours than the low dose group. There was also a dose response relationship in terms of the amount of the growth signal receptor HER2 expression.

The tumour growth effect was not confined to the highest dose. In fact, it was strongest at the lowest supplemented dose tested (5mg/kg) which the authors had chosen specifically to approximate what North Americans consume after fortification, roughly 0.8 to 1.0 mg of folic acid daily. It is important to note that this dosage does not produce unmetabolised folic acid. Above that level the folate stores stopped rising. The authors attributed this to the body’s capacity process it being saturated. This is yet more evidence that the effects at low dose are a cancer concern even when the effects at higher doses are different.

In mice bred to develop bowel tumours the effect changed over time. Folic acid reduced the number of early lesions at three months but had the opposite effect at six months. In rats, increasing folic acid increased the number of the earliest visible abnormalities in the bowel lining, although the number of tumours themselves did not differ. A recurrent finding in these rodent models is of two quite different effects. Folic acid is apparently protective in rodents prone to cancer in terms of initiation of a cancer but once one exists it promotes growth.

Maternal folic acid supplementation in rats reduced the incidence of colorectal adenocarcinoma in offspring by 64 percent. There are clearly certain genetic groups in rodents where a benefit can be seen as well as ones where it is clearly harmful and certain cancers that are increased even while others might not be.

Conclusion

Under randomised conditions in rodents, folic acid reduces MTHFR activity, restricts growth in the womb, shrinks the memory structures of the brain and impairs memory, weakens the immune cells that police viruses and tumours, shortens survival after infection and accelerates the growth of tumours that already exist. That is a long list of problems for a supposed vitamin that is being added to our food as a legal requirement.

Human safety evidence has not been actively sought so why is the animal evidence still being ignored?