Breastfeeding women who ate eggs, peanuts and shellfish during pregnancy will protect their babies against allergy, a major study shows.
Experiments found mice that consumed allergy-triggering foods like eggs and peanuts passed on protective antibodies to their offspring through nursing.
While many studies have suggested this could be the case, this research by Boston Children's Hospital offers the clearest evidence that breast milk can prevent allergies.
The study also shows how a 'biological mechanism' could strengthen tolerance to certain foods, paving the way to new allergy treatments for at-risk infants made from purified antibodies.
Study leader Dr Michiko Oyoshi, a pediatrician at hospital's Allergy and Immunology department, said: 'Whether mothers should eat allergenic foods during pregnancy or avoid them has been controversial.
'Different studies have found different results in part because it's hard in human studies to know when mothers and babies first encountered a specific food.
'But in a mouse model we can control exposure to food.'
The researchers said the findings back recent allergy-prevention guidelines for mums-to-be not to avoid eggs and peanuts.
They showed how the antibodies caused the baby mice to produce immune cells called 'T' cells specific to particular allergens, enabling them to tolerate the foods.
Women have previously been advised to avoid highly allergenic foods during pregnancy or while breastfeeding.
The study published in the Journal of Experimental Medicine found the breast milk of mothers who consumed the foods protected against allergy.
It prevented anaphylactic shock, an extreme and potentially life-threatening reaction and the production of immunoglobulin E and mast cells, both hallmarks of an allergic response.
Breast milk was even protective when fed to unrelated offspring not exposed to allergy-causing foods in the womb.
In other experiments females who had never consumed the foods were given the specific antibodies from other mothers and this also protected their breastfed offspring.
Human breast milk fed to mice with humanized immune systems tailored to respond to human antibodies was protective too, suggesting the findings will apply to human infants.
Dr Oyoshi, who is also an assistant professor of pediatrics at Harvard Medical School, finally had mice born to the eggs and peanuts-exposed mothers nurse from those that had never consumed the foods.
'We still saw protection from the womb exposure but the protection was better when the mice were also exposed through breastfeeding,' she explained.
'If you combine both womb and breastfeeding exposure you have optimal induction of food tolerance.'
The study also revealed each step of how food tolerance is passed onto babies as they are nursed by their mothers.
Firstly antibodies in the breast milk are transferred into offspring with the help of a receptor called FcRn on intestinal cells.
These 'dendritic' cells process a complex made up of antibody and allergen - called IgG-IC - and present it on their surface with the help of FcRn.
This kick-starts the production of the allergen-specific T regulatory (Treg) cells - which interact with other immune cells to suppress food allergy.
Offspring remained food-tolerant even after the mother's antibody disappeared from the babies' circulation, suggesting a long-lasting effect.
Dr Oyoshi's team is now enrolling mothers into the first study of its kind to see if the same protective mechanisms apply to humans.
They are collecting milk from actively breastfeeding women to find out - and to better understand what factors in the milk are protective.
They will compare milk from mothers whose infants are at high or low risk of food allergy.
This is based on whether an older sibling is allergic or the baby has early risk factors such as eczema.
Dr Oyoshi said: 'We are asking: Why is this protective mechanism not functioning in the allergic population? Is it just lack of antibodies?
'Maybe the mother cannot make the antibodies or maybe something is blocking the protective pathway.
'We don't really know how tolerance is induced in normal circumstances, and what causes breakdown in normal tolerance.'
If the infant is at low risk for food allergy the researchers will compare breast milk samples before and after the mothers eat egg or peanut.
Mothers with or without a personal or family history of food allergies are eligible to take part.
Pending the results Dr Oyoshi sees the possibility of treating infants at risk for food allergy with purified antibodies to those foods.
In the meantime her team's findings indicate 'you should be eating every food to create antibodies to everything.'
She said: 'Our study does not suggest mothers' peanut eating will guarantee a healthy baby.
'Given the complicated interactions between genetic and environmental factors there is not going to be just one diet or one set of behaviours that will make children allergic or healthy.'
In the US, four percent of adults (around nine million) have food allergies, compared to eight percent of children (around six million).
Those figures appear to be on the rise. There was a 50 percent increase in food allergies between 1997 and 2013, according to the CDC.
In the UK, around two million people living with a food allergy (around one in 50 adults and one in 12 children) - one of the highest rates in the world.
That does not include those with food intolerance, like lactose intolerance or celiac.