Scientists from the Wellcome Sanger Institute and their collaborators analyzed over 20,000 families’ genetic information and identified 12 children with between two to seven times more mutations than the general population.
The team linked the majority of these to increased mutations in the sperm of the biological father.
The research, published today in Nature, shows that just under half of these fathers had been treated with certain types of chemotherapy earlier in life, which could be linked to the increased number of mutations in their sperm cells.
While these cases of hypermutation in children are rare, and in the vast majority of children will not lead to genetic disorders, hypermutation will increase the risk of a child having a rare genetic disorder. It is important to investigate this further due to the implications it has for patients who receive chemotherapy and want to have children in the future.
Genomes are copied with a very low error rate when they are passed from one generation to the next. Nevertheless, as the human genome contains three billion letters, random mutations in the sperm and the egg are inevitable and pass from the parent to the child. This means that typically every child has around 60 to 70 new mutations that their biological parents don’t have.
These mutations are responsible for genetic variation along with many genetic diseases. Around 75 percent of these random mutations come from the father.
Most genetic disorders only occur when both copies of an important gene are damaged, resulting in what is known as a recessive disease. If only one copy is damaged, for example, by a new mutation, the remaining functioning copy of the gene will be able to prevent disease.
One of the main factors influencing mutation rate is the age of the parents, with mutations increasing by 1.3 mutations per year in the fathers and 0.4 mutations per year in mothers. If there is a higher number of germline mutations, there is a higher risk of a child being born with a dominant disorder.
However, hypermutation in children does not always mean they will have a dominant disorder.
In new research, from the Wellcome Sanger Institute and collaborators, scientists used genetic data and family health histories from existing databases to identify children that had unusually high mutation rates, between two and seven times higher than average, to investigate where these might have originated from.
The team analyzed data from over 20,000 UK families with children with suspected genetic conditions participating in the Deciphering Developmental Disorders and 100,000 Genomes projects.
They found that children with hypermutation were rare among these families. As the number of children with hypermutations was only 12 out of around 20,000, these rates of increased mutations could not have been caused by common exposures, such as smoking, pollution, or common genetic variation.
The other five men had all previously been treated with chemotherapy before conceiving a child. Three of these children had a pattern of mutations characteristic of chemotherapy using platinum-based drugs and the fathers of the other two children had both received chemotherapy with mustard-derived alkylating agents.
However, by linking the genetic data to anonymized health data, it could be shown that most fathers and all mothers who had received chemotherapy prior to conceiving a child did not have children with a notable excess of mutations.
This study exemplifies the value of linking nationwide genetic data and routine clinical records in secure, anonymized and trustworthy ways to provide unique insights into unanticipated, but important, questions.
Through the efforts of Health Data Research UK and its partners, these kinds of responsible analyses of potential clinical relevance will be easier to perform in the future.
While chemotherapy is one of the most effective treatments for cancer, it is widely recognized that it can have disruptive and debilitating side effects. Clinicians take these into account when prescribing this treatment.
If these types of chemotherapy were shown to impact sperm in some patients, this could have clinical implications on treatment plans and family planning.
Further research is required to investigate this at a deeper level before changing treatment for cancer in men. It is currently unclear why these types of chemotherapies seem to impact the sperm more than the egg cells.
Dr. Joanna Kaplanis, first author and Post-Doctoral Fellow at the Wellcome Sanger Institute, said: “Hypermutation in children, where they have between two and seven times more random mutations than the general population, is rare and therefore cannot be caused by common carcinogens or exposures.
“Our research analyses over 20,000 families and highlights new causes of these mutations, linking them back to germline mutations in the father’s sperm as well as identifying a new mutational signature.
“Understanding the impact of these germline mutations in the sperm could help us uncover why some people are more likely to have children with these high rates of random mutations, and help protect against these if they cause disease.”
John Danesh, Director of HDR UK Cambridge, who supported the research, said, “Hypermutation in children is an uncommon but important phenomenon that increases the risk of life-altering genetic diseases. By bringing together genetic data at scale, and linking this with routine clinical data like the hospital records of parents, the team has identified new risk factors that may influence future healthcare decisions.
“This work elegantly demonstrates how work in Health Data Research UK’s Understanding the Causes of Disease Programme is helping to link nationwide genetic data and clinical records in secure, anonymised and trustworthy ways that provide unique insights into unanticipated, but important questions.”
Sir Mark Caulfield, from Queen Mary University of London, and former Chief Scientist at Genomics England, said: “These findings were only possible due to access to whole genomes and linked health record data on the family members from the 100,000 Genomes Project. These findings could really help people with cancer consider family planning.”
Professor Matthew Hurles, senior author and Head of Human Genetics at the Wellcome Sanger Institute, said: “Chemotherapy is an incredibly effective treatment for many cancers, but unfortunately it can have some damaging side effects. Our research found a plausible link between two types of chemotherapy and their impact on sperm in a very small number of men.
“These results require further systematic studies to see if there is a causal link between chemotherapy and sperm mutations, and if there is a way of identifying individuals at risk prior to treatment so they could take family planning measures, such as freezing their sperm prior to treatment.
“I would also like to thank the families that donated their genetic and health information to make this research possible.”
Germline hypermutation is an uncommon but important phenomenon. We identified 12 hypermutated individuals from over 20,000 parent offspring sequenced trios in the DDD and 100kGP cohorts with a 2-7 fold increased number of dnSNVs. It is likely that there are additional, currently undetected, germline hypermutated individuals in the DDD cohort. The stringent strategy we adopted to screen this exome-sequenced cohort for potential hypermutated individuals for subsequent confirmation by genome sequencing will have missed some individuals with hypermutation of 2-7 fold.
In two of the 12 hypermutated individuals, the excess mutations appeared to have occurred post-zygotically, however for the majority (n=8) of these hypermutated individuals, the excess dnSNVs phased paternally implicating the father as the source of this hypermutation. For five of these fathers, characteristic mutational signatures and clinical records of cancer treatment prior to conception strongly implicated the mutagenic influence of two different classes of chemotherapeutics: platinum-based drugs (3 families) and mustard-derived alkylating agents (2 families). We also identified likely paternal mutator variants in two hypermutated families. These were rare homozygous missense variants in two known DNA repair genes: XPC and MPG. Functional and clinical data strongly supported the mutagenic nature of these variants.
It is well established that defects in DNA repair genes can increase somatic mutation rates and elevate cancer risk56. Our findings imply that germline mutation rates can be similarly affected. However, defects in DNA repair pathways do not always behave similarly in the soma and the germline. We interrogated PTVs in an established somatic mutator gene, MBD4, and found they did not have a detectable effect in the germline57. We also examined the impact of parental rare nonsynonymous variants in DNA repair genes on the number of DNMs in offspring and did not find a significant difference. To detect more subtle effects of these variants other analytical approaches will need to be explored. Paternal variants that have previously been associated with a cancer phenotype were nominally significant but having one of these variants only amounted to an estimated average increase of ~2 DNMs in the child. If only a subset of these variants have an impact in the germline this would dilute the power to detect a mutagenic effect and it is likely that both larger sample sizes and additional variant curation will be needed to investigate this further. There may also be genes and pathways that impact mutation in the germline more than the soma; uncovering the genes and associated variants in these genes will be more challenging.
Germline hypermutation accounted for 7% of the variance in germline mutation rate in the 100kGP rare disease cohort. The ascertainment in this cohort for rare disease in the offspring, together with the causal contribution that germline mutation plays in rare diseases, means that germline hypermutated individuals are likely enriched in this cohort relative to the general population. As a consequence, our estimate of the contribution of germline hypermutation to the variance in numbers of dnSNVs per individual is likely inflated. However, the absolute risk of a germline hypermutator having a child with a genetic disease is still low. The population average risk for having a child with a severe developmental disorder caused by a de novo mutation has been estimated to be 1 in 300 births11 and so a 4-fold increase in DNMs in a child would only elevate this absolute risk to just over 1%. Therefore, we anticipate that most germline hypermutated individuals will not have a rare genetic disease, and germline hypermutation will also be observed in healthy population cohorts.
The two genetic causes of germline hypermutation that we identified were both recessive in action. Similarly, most DNA repair disorders act recessively in their cellular mutagenic effects. This implies that genetic causes of germline hypermutation are likely to arise at substantially higher frequencies in populations with high rates of parental consanguinity. In such populations, the overall incidence of germline hypermutation may be higher and the proportion of the variance in the number of dnSNVs per offspring accounted for genetic effects will be higher. We anticipate that studies focused on these populations are likely to identify additional mutations that affect germline mutation rate.
We found that, among 7,700 100kGP families, parental age only explained ~70% of the variance in numbers of dnSNVs per offspring, which is substantially smaller than a previous estimate of 95% based on a sample of 78 families3. Repeated sampling of 78 trios from the 100kGP showed that estimates of the variance explained by parental age can vary dramatically stochastically and we regard our estimate based on two orders of magnitude more trios to be more reliable, although other differences between the studies such as measurement error and criteria for ascertainment of families might be having a subtle influence.
The residual ~20% of variation in numbers of germline dnSNVs per individual remains unexplained by parental age, data quality and hypermutation. We found that rare variants in known DNA repair genes are unlikely to account for a large proportion of this unexplained variance. Heritability analyses suggested that polygenic contributions from common variants (MAF>1%) are unlikely to make a substantive contribution to this variance; however, we observed some evidence that the polygenic contribution of intermediate frequency paternal variants (0.001<MAF<0.01) could be more substantial although larger sample sizes are required to confirm this observation. A limitation to these heritability analyses is that we use DNMs in offspring as a proxy for individual germline mutation rates. Measuring germline mutation rates more directly by, for example, sequencing hundreds of single gametes per individual, should facilitate better powered association studies and heritability analyses.
Environmental exposures are also likely to contribute to germline mutation rate variation. We have observed evidence that certain chemotherapeutics can affect germline mutation rate and targeted studies on the germline mutagenic effects of different chemotherapeutics (e.g. in cancer survivor cohorts) will be crucial in understanding this further. We anticipate that these studies will identify considerable heterogeneity in the germline mutagenic effects of different chemotherapeutics, in part due to differences in the pemeability of the blood-testis barrier to different agents58, as well as variation in the vulnerability to chemotherapeutic germline mutagenesis by sex and age. As so few individuals are treated for cancer prior to reproduction, chemotherapeutic exposures will not explain a large proportion of the remaining variation in germline mutation rates however chemotherapeutic mutagenesis has important implications for cancer patients who plan to have children, especially in whether they decide to store unexposed gametes for future use of assisted reproductive technologies.
Unexplained hypermutation and additional variance in germline mutation rate may be explained by other environmental exposures. A limitation of this study was the lack of data on non-therapeutic environmental exposures. However, and somewhat reassuringly, the relatively tight distribution of DNMs per person in 100kGP suggests that there are unlikely to be common environmental mutagen exposures in the UK (e.g. cigarette smoking) that causes a substantive (e.g, >1.5 times) fold increase in mutation rates and concomitant disease risk. The germline generally appears to be well protected from large increases in mutation rate. However, including a broader spectrum of environmental exposures in future studies would help to identify more subtle effects and may reveal gene-by-environment interactions.
reference link https://www.biorxiv.org/content/10.1101/2021.06.01.446180v2.full
Original Research: Open access.
“Genetic and chemotherapeutic influences on germline hypermutation” by Matthew Hurles et al. Nature
