Background: Li–Fraumeni syndrome is an autosomal dominant cancer predisposition syndrome caused by germline mutations in the TP53 gene. The frequency of germline de novo TP53 mutations is largely unknown; few unequivocal de novo mutations have been reported.
Methods and results: Of 341 patients with early onset cancer sent for clinical testing to a national reference laboratory, 75 patients had TP53 germline mutations. Five (7%) de novo mutations were identified, as well as an additional 10 TP53 germline mutations likely to be de novo by family history. The frequency of de novo TP53 mutations in this patient sample is at least 7% and may be as high as 20%.
Conclusions: The possibility that de novo germline TP53 mutations are relatively common has implications for testing and the identification of potential Li–Fraumeni syndrome in patients with little or no family history of cancer.
Statistics from Altmetric.com
Li–Fraumeni syndrome (LFS) (OMIM 151623) is an autosomal dominant cancer predisposition syndrome characterised by early onset tumours. The syndrome was initially described in 1969 by Li and Fraumeni1 and eventually characterised by five main cancers including sarcoma, breast cancer, leukaemia, brain tumours, and adrenocortical carcinoma (ACC).2 In some families, the underlying cause of LFS was later discovered to be due to a germline mutation in the TP53 gene.3 The lifetime penetrance of LFS is high4; females have an overall higher risk of developing cancer (93% in women vs 68% in men by the age of 50) and have an earlier average age of onset (29 years of age vs 40 years of age in men).5
The frequency of germline TP53 de novo mutations is unclear. TP53 de novo mutations are reported to be frequent (up to 20%),6 uncertain (www.geneclincs.org),7 and very low (1%).4 8 Only seven de novo mutations have been documented by: (1) absent family history of early cancer, (2) absent mutation in both parents, and (3) DNA based confirmation of paternity.6 7 9
We previously analysed patients from 525 families with family histories sent for clinical testing to the Clinical Molecular Diagnostic Laboratory (CMDL) at The City of Hope. Full sequencing of the TP53 gene identified 91 families (17.3%) with TP53 mutations. Sufficient family history information was available for 341 (75 TP53 mutant and 266 TP53 wild type) families. At least 7% of the 75 TP53 mutant probands had de novo germline mutations.
Materials and methods
All coding exons (2–11) and associated splice junctions of the TP53 gene were analysed in both directions by direct DNA sequencing analysis using the ABI 377 or the ABI 3730 DNA automated fluorescent sequencer (Applied Biosystems, Foster City, California, USA). Sequence analysis was performed in both directions.
Five hundred and twenty-five probands, predominantly from the USA and Canada, were sent to the Clinical Molecular Diagnostic Laboratory (CMDL) at The City of Hope for TP53 testing. Sufficient family history information was available for 341 families—that is, at least a three generation pedigree with ages of cancer onset. The 341 probands with pedigree data were analysed further. The average age of first cancer diagnosis in these patients was 22 years. Thirty-six per cent (122/341) of probands had two or more cancers, and in 30% (37/122) of these patients a TP53 mutation was detected.
Classification as origin or potential origin
For the five confirmed de novo mutations, the mother and the father were both tested for, and found not to carry, the germline TP53 mutation identified in their child. Mutations from 10 other families were considered probable de novo mutations if one or both parents were unavailable for testing and if the patient had no first or second degree relatives (other than offspring) with cancer under the age of 70 years. The AmpFISTR Identifiler PCR Amplification Kit (Applied Biosystems, P/N 4322288) was used to confirm maternity and paternity among the five de novo mutation families. In this well validated, commercially available assay, 15 tetranucleotide repeat loci and a gender determining marker are amplified in a single, robust polymerase chain reaction (PCR) amplification.10 The data analysis and allele designations were carried out using GeneMapper Software (Version 4.0; Applied Biosystems).
A mutation in the TP53 gene was detected in 75 of 341 probands with sufficient pedigree information submitted for molecular testing. Among the 75 TP53 mutations, 15 were possibly de novo, based on a negative family history. Among five of these possible de novo mutations, both parents were available, confirmed by DNA analysis to be the biological parents, and found not to carry the mutation. Therefore, the frequency of confirmed de novo TP53 mutations in this sample is 7% (5/75).
The average age of first cancer diagnosis for the confirmed de novo mutations was 5.6 years. Four of the five (80%) patients with de novo mutations had multiple primary cancers (table 1). Four of the five patients also had a childhood cancer, including four with either ACC or choroid plexus carcinoma and two with sarcoma. One patient had bilateral breast cancer in her 20s. The mutations included one splicing mutation and four missense mutations. The missense changes were at amino acids evolutionarily conserved in TP53 from mammals, chicken, trout, Xenopus, invertebrates, and p73 from human and rat. All five mutations have been reported in the International Agency for Research on Cancer (IARC) database of either germline or somatic changes in cancer.11
In addition to the five patients with confirmed de novo mutations, 10 patients had no first or second degree relatives with cancer (other than children or grandchildren) under 70 years of age. None of these 10 patients had a first or second degree relative with an LFS “core” cancer at any age (core cancers are defined as cancers of the breast and brain, adrenocortical carcinoma, and sarcoma, which together account for >77% of cancers seen in LFS probands).12 However, the parents of these 10 patients were unavailable for testing. The average age of onset of the first cancer was 13.7 years of age (table 2), which is well below the overall average age of onset for all 75 mutation carriers of 21.9 years of age (data not shown). If these patients are included, the frequency of de novo TP53 mutations could be as high as 20% (15/75).
An analysis of 75 patients with TP53 germline mutations reveals that the de novo mutation frequency is at least 7% and could be as high as 20%. The five patients with demonstrated de novo mutations had a first cancer diagnosis at an average age of 5.6 years, ranging from 4 months to 24 years of age. Four of the five had additional primary cancers. This may reflect a high threshold for sample submission in the absence of family history. The 10 additional patients with candidate de novo mutations had an average of first cancer onset of 13.6 years and 60% had two or more primaries. It follows that patients who develop cancers associated with LFS, especially if they have multiple primary cancers, should be considered for testing even without a family history of cancer.
In the IARC germline TP53 database (www-p53.iarc.fr), the mode of inheritance is denoted as “new” for several patients. Among a total of seven papers6 7 8 13 14 15 16 which report these “new mutations,” only two papers6 7 report confirmed de novo mutations (in which both the mother and the father are tested and paternity testing is performed) (tables 3 and 4).
Reflecting the criteria for Li-Fraumeni syndrome, studies in the literature tend to be biased against patients with a de novo TP53 mutation. A total of three papers report seven de novo mutations in which both the mother and the father were tested for the mutation and parental identity testing was performed (table 3).6 7 9 18 Among these seven de novo mutations, three are quite intriguing. Chompret et al9 examined a cohort of 268 patients who were either affected with multiple cancers or had at least one first or second degree relative or first cousin with any cancer before the age of 46.9 They identified 17 (6.3%) TP53 mutations, four of which were de novo mutations, as both parents tested negative for the identified mutation and parental identity testing was confirmed in each case. However, probands with three of the four de novo mutations had significant family histories. One patient had a maternal cousin with breast cancer at 29, another had a maternal aunt with ovarian cancer at 30, and one had a maternal cousin with neuroblastoma at 7 years of age.
The 10 patients in this study with putative TP53 origin mutations had either early onset cancers or multiple cancers, suggestive of a cancer syndrome. The absence of family history in these patients and in the five with confirmed origin mutations, despite the high penetrance of TP53 mutations in the individual, is in striking contrast to the other 60 families with TP53 mutations in this study. Unfortunately, the parents of the 10 patients with putative origin mutations are not available for testing. While it is our clinical impression that the cancers in these patients result from de novo mutations in the parent (or perhaps the grandparent), some of these cases may not represent origins of mutations, due to the caveats described below.
In this context, we note that two additional, ultimately spurious, putative de novo mutations were found during the course of our clinical testing. In one family, there was a history of cancer on the paternal side. Neither the mother nor the father carried the mutation; the affected child had a TP53 germline mutation. Testing indicated non-paternity, revealing a pattern consistent with a sibling of the father. In another family, a child had an LFS core cancer at 2 years of age and both parents were negative for the mutation. However, testing of the father revealed an ambiguous pattern inconsistent with first degree relation, resulting from a past bone marrow transplantation from a donor brother. Analysis of sperm confirmed that the father carried the germline mutation, although, in his 30s, he had never had cancer (his bone marrow transplant was for a non-cancer condition). It is worth noting that, because of the high frequency of female breast cancer, the penetrance of TP53 mutations in males is considerably lower than that in females (93% vs 68%, respectively, by age 50).4 This can result in an over-estimation of de novo mutations if the father appears to be a non-carrier and has not been tested for the mutation. Note also that identity testing with only a few polymorphisms might have spuriously confirmed parentage by chance in both of the cases described above.
There are a number of caveats that could affect the measured frequency of de novo mutations. The classic criteria for LFS or LFS-like syndrome (Birch and Eeles) cannot be made in a patient with sporadic disease.20 21 22 However, recent criteria advocated by Chompret et al do account for sporadic disease.9 23 Incomplete pedigrees which are missing relatives with cancer could result in an overestimation of the number of possible de novo mutations. More likely, families with de novo mutations are underrepresented because of the general sense that de novo mutations are uncommon in LFS. Thus, one affected individual in a family is unlikely to be sent for testing unless that individual is very young or has multiple cancers. Of the 15 documented or possible de novo mutations described in this study, 11 affected patients had cancer in childhood. Three of the four patients older than 18 had multiple cancers. One had an LFS core cancer in the late 20s and a son with an LFS core cancer at age <10 years. In addition, all but three with childhood cancer and a TP53 mutation had a second cancer. Since the majority of patients with germline TP53 mutations do not have cancer in childhood and do not have multiple cancers before 30, the sporadic patients submitted for analyses are likely to represent the minority of the phenotypes displayed by a minority of all the patients with de novo TP53 germline mutation. Thus, it is prudent to recognise that a patient with sporadic cancer of onset <50 may have a de novo TP53 mutation.
A high frequency of de novo mutations is not surprising given that fitness is limited. Recognition of families with TP53 germline mutations can be difficult given the wide spectrum of cancer types that can occur in LFS families. Several clinical criteria have been constructed in order to diagnose LFS families. Chompret is the only criterion for TP53 testing which accounts for the possibility of de novo mutations: (1) proband with multiple primary tumours, two of which are sarcoma, brain tumour, breast cancer, and/or ACC, with the initial cancer occurring before 36 years of age, regardless of family history; (2) proband with ACC regardless of age of onset or family history. Four of the five confirmed de novo mutation carriers and nine of the 10 probable de novo mutation carriers meet Chompret criteria. The patients who did not meet Chompret criteria were diagnosed with early onset breast cancers.
The frequency of de novo TP53 mutations is at least 7% and may be has high as 20%.
The average age of first cancer diagnosis for the confirmed de novo TP53 mutations was ∼6 years of age.
80% of patients with de novo TP53 mutations had multiple primary cancers.
A high frequency of de novo TP53 mutations has major clinical implications for recognition of and screening for LFS patients.
It may be prudent to consider TP53 germline mutation testing for any individual with cancer <30 (especially breast cancer, sarcomas, brain tumours, leukaemia, adrenocortical carcinoma, or choroid plexus carcinoma, or multiple primaries)
In conclusion, herein we report a frequency of de novo mutations of 7–20%. One caveat of the study is that the authors did not know the ages of patients’ relatives without cancer, making it possible that some of the “potential” de novo mutations may actually be mutations inherited from relatives who have not yet developed cancer. However, it is possible that the true frequency of de novo TP53 mutations is even higher than 20% due to the ascertainment bias of a clinical sample set (that is, patients with a negative family history are less likely to be sent for TP53 testing). A high frequency of de novo TP53 mutations has major clinical implications for recognition of and screening for LFS patients. It may be prudent at least to consider genetic testing for TP53 germline mutations in individuals who have cancer below age 30, especially breast cancer, sarcomas, brain tumours, leukaemia, ACC, and choroid plexus carcinoma, or multiple cancers at any age, especially if at least one is of the above types.
Detail has been removed from these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.
The authors thank Cindy Wen-Fong, Vu Nguyen, Jennifer Han, and Hernanie Dominguez for technical laboratory work.
Funding Supported in part from a grant from the National Cancer Institute of Canada (to DM) with funds from the Canadian Cancer Society.
Competing interests None.
Patient consent Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.