A higher likelihood of autism spectrum disorder (ASD) and autistic traits have been linked to maternal grandmothers smoking while pregnant.
Girls whose grandmothers smoked while pregnant were 67 percent more likely to show autistic traits, and both boys and girls were 53 percent more likely to be diagnosed with ASD.
'In terms of mechanisms, there are two broad possibilities. There is DNA damage that is transmitted to the grandchildren or there is some adaptive response to the smoking that leaves the grandchild more vulnerable to ASD,' said Professor Marcus Pembrey at University College London, and one of the study authors.
'More specifically, we know smoking can damage the DNA of mitochondria - the numerous "power-packs" contained in every cell, and mitochondria are only transmitted to the next generation via the mother's egg.'
Researchers at the University of Bristol used data collected from the Avon Longitudinal Study of Parents and Children (ALSPAC), which has followed approximately 14,500 individuals who were born in the 1990s. They assessed the grandchildren of maternal and paternal grandmothers who smoked during pregnancy for ASD or four known autistic traits.
Of the participants, 7000 were found to have autistic traits. The researchers discovered that girls were 67 percent more likely to show at least one of two autistic traits, including poor social communication and repetitive behaviours, if their maternal grandmother had smoked during pregnancy.
'We have no explanation for the sex difference, although we have previously found that grand-maternal smoking is associated with different growth patterns in grandsons and granddaughters,' said Professor Pembrey.
There was also a 53 percent higher risk of the smoker's grandchild being diagnosed with ASD. As only 177 participants in the ALSPAC study were diagnosed with ASD, this group was too small to distinguish any difference between the sexes.
The discovery suggests that cigarette smoke could affect the developing eggs of females in the womb, and eventually the development of the children born from those eggs, either through epigenetic or mitochondrial DNA damage.
The researchers emphasise that many other genetic and environmental factors are believed to influence an individual's chance of developing ASD. Further research will be needed to identify what molecular changes are caused by maternal grandmother's smoking, and the researchers also plan to investigate if the effect can be seen in great-grandchildren too.
Professor Peter Hajek of Queen Mary University of London said the study suggested 'an interesting epigenetic effect' but said that more data on children actually diagnosed with ASD would be useful to rule out a 'chance finding'. 'Several large studies examined possible links between maternal smoking during pregnancy and autism and their meta-analysis shows no link, so this finding, if real, would be surprising,' he said.
The study was published in Scientific Reports.