Li-Fraumeni and Li-Fraumeni—like syndrome among children diagnosed with pediatric cancer in Southern Brazil

Authors

  • Juliana Giacomazzi BSc, MSc, PhD,

    Corresponding author
    1. Laboratorio de Medicina Genomica, Experimental Research Center, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    2. Post-Graduate Program in Medicine, Medical Sciences, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil
    3. Genetics Department, UFRGS, Porto Alegre, Brazil
    • Corresponding author: Juliana Giacomazzi, BSc, MSc, PhD, Laboratório de Medicina Genômica, Centro de Pesquisa Experimental, Hospital de Clínicas de Porto Alegre and Genetics Departament, Universidade Federal do Rio Grande do Sul (UFRGS), Rua Ramiro Barcelos, 2350, 90035-903 Porto Alegre, RS, Brazil; Fax: (011) 55 51 3359 7661 jugiacomazzi@gmail.com

    Search for more papers by this author
  • Simone G. Selistre MD,

    1. Post-Graduate Program in Medicine, Medical Sciences, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil
    2. Pediatric Oncology Service, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    Search for more papers by this author
  • Cristina Rossi MD,

    1. Laboratorio de Medicina Genomica, Experimental Research Center, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    2. School of Medicine, UFRGS, Porto Alegre, Brazil
    Search for more papers by this author
  • Barbara Alemar BSc, MSc,

    1. Laboratorio de Medicina Genomica, Experimental Research Center, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    2. Post-Graduate Program in Genetics and Molecular Biology, UFRGS, Porto Alegre, Brazil
    Search for more papers by this author
  • Patricia Santos-Silva,

    1. Laboratorio de Medicina Genomica, Experimental Research Center, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    2. Post-Graduate Program in Medicine, Medical Sciences, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil
    Search for more papers by this author
  • Fernando S. Pereira,

    1. Laboratorio de Medicina Genomica, Experimental Research Center, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    2. School of Medicine, UFRGS, Porto Alegre, Brazil
    Search for more papers by this author
  • Cristina B. Netto MD, PhD,

    1. Medical Genetics Service, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    Search for more papers by this author
  • Silvia L. Cossio PhD,

    1. Laboratorio de Medicina Genomica, Experimental Research Center, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    2. Post-Graduate Program in Medicine, Medical Sciences, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil
    Search for more papers by this author
  • Daniela E. Roth MD,

    1. Pediatric Oncology Service, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    Search for more papers by this author
  • Algemir L. Brunetto MD, PhD,

    1. Pediatric Oncology Service, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    Search for more papers by this author
  • Marcelo Zagonel-Oliveira BSc, MSc,

    1. Instituto Nacional de Genetica Medica Populacional (INaGeMP), UFRGS, Porto Alegre, Brazil
    Search for more papers by this author
  • Ghyslaine Martel-Planche,

    1. International Agency for Research on Cancer, Lyon, France
    Search for more papers by this author
  • Jose R. Goldim BSc, PhD,

    1. Bioethics Research Laboratory, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    Search for more papers by this author
  • Pierre Hainaut BSc, PhD,

    1. International Prevention Research Institute, Lyon, France
    Search for more papers by this author
  • Suzi A. Camey PhD,

    1. Department of Statistics, Institute of Mathematics, UFRGS, Porto Alegre, Brazil
    Search for more papers by this author
  • Patricia Ashton-Prolla MD, PhD

    1. Laboratorio de Medicina Genomica, Experimental Research Center, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    2. Post-Graduate Program in Medicine, Medical Sciences, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil
    3. Genetics Department, UFRGS, Porto Alegre, Brazil
    4. Post-Graduate Program in Genetics and Molecular Biology, UFRGS, Porto Alegre, Brazil
    5. Medical Genetics Service, Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil
    Search for more papers by this author

  • We thank Cláudio Galvão, Lauro Greggianin, Camila Matzenbacher Bittar, and Filippo Vairo for their collaboration with patient recruitment; Diego D'Avila Paskulin, Marcia da Silveira Graudenz, and Luise Meurer for laboratory support; and Lavinia Schuler-Faccini and INaGeMP for their help with geomapping.

Abstract

BACKGROUND

Pediatric cancers are a feature in patients with Li-Fraumeni syndrome and its variant Li-Fraumeni–like syndrome (LFS/LFL). To the best of the authors' knowledge, TP53 germline mutations are currently the only molecular defect known to be associated with this disease. Recently, a specific germline mutation in this gene, p.R337H, has been reported at a high prevalence in Brazil.

METHODS

The prevalence of LFS/LFL was investigated in children with cancer who were diagnosed with tumors on the LFS/LFL spectrum and in a small consecutive series of controls without cancer. The prevalence of the germline p.R337H mutation and of other germline TP53 mutations was investigated in a general group of children with cancer and exclusively in children fulfilling the clinical criteria for LFS/LFL, respectively.

RESULTS

Among the 65 children without cancer, 1.5% had a family history of LFL whereas of the 292 children with cancer, 25.3% had a family history of LFL (P < .001). Screening for the p.R337H mutation identified 11 carriers (3.7%), 9 of whom were diagnosed with adrenocortical carcinomas (ACC) and 2 of whom were diagnosed with choroid plexus carcinomas. One of the ACC probands was homozygous mutant. The Brazilian founder haplotype and loss of heterozygosity at the p.R337H locus were present in all carriers. In addition, direct sequencing of the entire TP53 coding region and gene rearrangement analysis of probands fulfilling the criteria for LFL (Eeles 2 criteria, Birch and/or Chompret criteria) and who were negative for the p.R337H mutation revealed a DNA-binding domain pathogenic mutation, p.G245S, in 1 child.

CONCLUSIONS

TP53 p.R337H testing should be offered to Brazilian children diagnosed with ACC and choroid plexus carcinoma. A significant percentage of children with cancer in southern Brazil fulfill the criteria for LFL and should be referred for genetic risk assessment. Cancer 2013;119:4341–4349. © 2013 American Cancer Society.

INTRODUCTION

Pediatric cancer is one of the more frequent causes of nontraumatic deaths in children, representing 1% to 3% of all cancer diagnoses worldwide.[1-5] The Brazilian National Cancer Institute estimates that approximately 384,340 new cancer diagnoses will have been made in Brazil in 2012, including 11,530 (3%) diagnosed in children and adolescents aged < 19 years.[5] Childhood cancers that arise within the context of a genetic predisposition are rare events, and are estimated to account for approximately 5% to 10% of all pediatric cancer cases.[5] However, with increased awareness of cancer family history and advanced technologies in mutation detection, this percentage will likely increase,[6] facilitating the presymptomatic identification of high-risk individuals and ultimately decreasing cancer-related mortality in this setting.

Childhood cancers are a common feature in patients with Li-Fraumeni syndrome and its variant Li-Fraumeni–like syndrome (LFS/LFL), which are inherited autosomal dominant disorders associated with multiple malignancies in childhood and in young adults. These include, among other tumors, sarcomas, breast cancers, leukemias, adrenocortical carcinomas (ACCs), and central nervous system tumors. To the best of our knowledge, germline mutations in the TP53 gene are currently the only molecular defect known to be associated with the disease. In the general population, such mutations are estimated to occur in 1 of 2000 to 5000 individuals.[7-11] The most common childhood tumors observed in mutation carriers are soft-tissue and bone sarcomas, brain tumors, leukemia, and ACC.[8, 10, 12] Other tumor types that are less frequently associated with the syndrome include Wilms tumors and germ cell tumors.[10]

Several clinical criteria based on cancer family history have been developed to identify families with suspected LFS/LFL and are used to indicate TP53 mutation testing. Criteria for classic LFS were proposed in 1988 based on the cancer phenotype observed in 24 families.[13] Less-stringent criteria were proposed by Birch and Eeles in 1994 and 1995, by respectively, as a result of the observation that not all mutation-positive families display the classic LFS phenotype.[10, 14, 15] More recently, Chompret et al and Tinat et al proposed additional criteria for the identification of germline TP53 mutation carriers.[16-18] The Chompret criteria are more restrictive regarding the cancer types and age of onset in the proband than the previously published LFL criteria, but allow for the possibility of testing a cancer-affected individual with a negative family history. The probability of detecting a germline TP53 mutation is directly related to the stringency of the phenotypic criteria. However, even in classic LFS families, germline TP53 mutations are observed in approximately 56% to 80% of the probands.[19-23] Germline TP53 mutations are estimated to occur in 11% to 40%, 20% to 40%, and 7% to 14% of the families meeting the criteria of Eeles, Birch et al, and Chompret et al, respectively.[14-16, 22-24]

Recently, a specific germline mutation occurring in the tetramerization domain (exon 10) of the TP53 gene, p.R337H, has been reported at a high prevalence in southern and southeastern Brazil.[25-29] Initial studies on this mutation claimed that its main, if not exclusive, cancer-related risk was childhood ACC in carriers.[25, 27] However, other recent studies identified p.R377H carriers using the less-strict LFL criteria,[28] and in patients with tumors other than ACC, such as choroid plexus carcinoma (CPC), osteosarcoma, and breast cancer.[30-32] Although a few mutation carriers have been identified in Europe,[17, 33, 34] the majority of p.R337H-positive individuals have been described in Brazil, especially in the southern and southeastern regions, where the mutation is estimated to occur in 1 in every 300 individuals in the general population.[29, 35, 36]

In the current study, to better understand the contribution of germline TP53 mutations in Brazil, we performed an in-depth assessment of a large pediatric series from the southern region of the country among individuals diagnosed with tumors on the LFS/LFL spectrum. We assessed clinical criteria that would allow for the identification of the LFS/LFL syndrome as well as the frequency of the common Brazilian founder mutation p.R337H and other germline TP53 mutations and performed a detailed familial, clinical, and molecular characterization of mutation-positive individuals.

MATERIALS AND METHODS

Participant Recruitment

All parents or legal guardians of children diagnosed with and/or treated for tumors on the LFS/LFL spectrum (ACC, sarcoma, central nervous system tumors, leukemia, germ cell tumors, and Wilms tumor) at the Pediatric Oncology Service of the Hospital de Clinicas de Porto Alegre, Brazil (a reference public and university hospital) between 1998 and 2011 were contacted and invited to participate in the study, regardless of their family history of cancer. Contact was attempted by telephone at least 3 times and if no contact was possible, an attempt was made by mail to reach all patients with a definitive histopathological diagnosis of the above-mentioned cancers. Those who volunteered to participate were included after providing informed consent. Parents and other relatives of proband mutation carriers identified in the study were also invited to participate. In addition, we also interviewed a consecutive sample of the parents of 65 children without cancer who were admitted to the pediatric ward of the same hospital during a period of 3 months for the treatment of non–cancer-related disorders to verify their cancer family history.

Ethical Aspects

The study was approved by the Institutional Research and Ethics Committees (protocol 08-080) and by the National Commission for Ethics in Research (protocol 14821). All individuals/families recruited for the current study provided written informed consent and authorized the publication of pedigrees.

Clinical Data and Pedigrees

Clinical data regarding all recruited patients and families were obtained by interviews, specific questionnaires, and review of medical records. The family history was recorded in detailed pedigrees with information traced as far back and laterally as possible, extending to paternal and maternal lines and including a minimum of 3 generations. Confirmation of a family history of cancer was attempted in all cases with pathology reports, medical records, and/or death certificates and was obtained whenever possible. The presence of criteria for LFS/LFL was assessed through independent review of the pedigree by 3 of the authors (P.A.P., C.B.O.N., and J.G.).

Genetic Analyses

Peripheral blood samples were collected in ethylenediamine tetraacetic acid from all patients and formalin-fixed, paraffin-embedded (FFPE) tumor and adjacent normal tissues were retrieved whenever possible. DNA was extracted by standard methods using the Illustra blood genomicPrep Mini Spin Kit (GE Healthcare Life Sciences, Sao Paulo, Brazil) and QIAmp DNA FFPE Tissue Kit (Qiagen, Valencia, Calif). Screening for the germline TP53 p.R337H mutation (either genomic DNA from peripheral blood or FFPE nontumoral tissue) was performed by allelic discrimination using a TaqMan assay (Applied Biosystems, Foster City, Calif) in duplicate, and using positive and negative internal controls in all experiments. All mutation-positive results were confirmed by a second independent TaqMan analysis and by direct exon 10 sequencing using primers and conditions described previously.[37] In probands fulfilling LFS or LFL criteria and for whom initial p.R337H testing was negative, the entire coding sequence of TP53 (exons 2-11) was sequenced using the protocols described,[37] and screening for gene rearrangements was performed by multiplex ligation-dependent probe amplification according to the manufacturer's instructions (SALSA MLPA probemix P056-B1 TP53; MRC-Holland, Amsterdam, The Netherlands). Tumor DNA was sequenced for TP53 mutations to assess loss of heterozygosity (LOH) for p.R337H or the presence of additional alterations in all germline mutation carriers. The presence of the Brazilian founder p.R337H haplotype (A3) was verified in mutation-positive samples as previously described by Garritano et al using allele-specific oligonucleotide–polymerase chain reaction and nested polymerase chain reaction analyzing single nucleotide polymorphism (SNP) 28 (rs9894946) from a panel of 29 intragenic SNPs.[35] Because this SNP is identical in haplotypes A1 and A3, all cases were further genotyped for SNP15 (rs1642785) and SNP18 (rs1800370) by direct sequencing.

Statistical Analysis

SPSS statistical software (version 14.0; SPSS Inc, Chicago, Ill) was used for data handling and statistical analyses. Continuous variables were expressed as the median and range; categorical variables were expressed by their absolute and/or relative frequencies. P values < .05 were considered to be statistically significant.

RESULTS

The families of all 717 patients diagnosed and treated for cancers on the LFS/LFL spectrum at the study institution between 1998 and 2011 were invited to participate in the study; 292 (40.7%) volunteered and were recruited. The majority of the families who agreed to participate (n = 288; 98.6%) were families of cancer survivors. A total of 48 families (6.7%) refused to participate in the current study and the remainder did not respond to the invitation. The average recruitment rate per year was 46.3% (range, 21.6%-83.8%), with recruitment rates of > 50% in the years 2005 and 2009 through 2011. The birthplace of the children with cancer who were included in the study is shown in Figure 1, demonstrating that the vast majority of patients admitted to the study institution are born in the state of Rio Grande do Sul.

Figure 1.

The birthplaces of the 292 children with cancer who were included in the current study are shown. Black and red triangles and blue circles represent the city of birth of the homozygous mutant, heterozygous, and homozygous normal individuals, respectively. (A) Distribution in the Brazilian territory is shown. (B) Details of the distribution of patients originating in the southern region, including the states of Rio Grande do Sul, Santa Catarina, and Parana, are shown.

Clinical information regarding the patients included in the current study is summarized in Table 1. The mean age of the group of 65 children without cancer was 9 years. Although a positive first- and/or second-degree family history of cancer in general and breast cancer in particular was reported in 23 (35.9%) and 4 (6.3%) of the children without cancer (excluding 1 child who was adopted), respectively, only 1 child (1.6%) had a family history that was consistent with LFL criteria (Chompret, Chompret modified, Birch, and Eeles 1),[14-16] which was significantly different from the frequency observed among the children with cancer (P < .001).

Table 1. Clinical Characteristics of the Children With Cancer Included in the Current Study (N = 292)
FeatureNo.%Median Age of Patient at Cancer Diagnosis, (Range), Years
  1. Abbreviations: CNS, central nervous system; LFL, Li-Fraumeni-like syndrome; LFS, Li-Fraumeni syndrome.

  2. a

    Other sarcomas in the probands included Ewing (n = 17), synovial (n = 4), primitive neuroectodermal (n = 3), clear cell (n = 3), epithelioid (n = 1), fibromyxoid (n = 1), primary hepatic (n = 1), and myofibroblastic (n = 1).

  3. b

    Other CNS tumors in the probands included craniopharyngioma (n = 5), glioblastoma (n = 3), optic nerve glioma (n = 4), ependymoma (n = 2), germinoma (n = 2), pinealoblastoma (n = 2), meningioma (n = 1), neuroblastoma (n = 1), primitive neuroectodermal (n = 1), and unspecified type (n = 6).

  4. c

    Leukemias in the probands included acute lymphocytic leukemia (n = 60), acute myelocytic leukemia (n = 13), and chronic myelocytic leukemia (n = 4).

  5. d

    Germ cell tumors in the probands included nongerminomatous-endodermal sinus tumor (n = 5); mature teratoma (n = 3); choriocarcinoma (n = 1); cystic teratoma (n = 1); teratoma with malignant transformation (n = 1); germinomatous-ovarian dysgerminoma (n = 1); mixed germ cell tumor-sex cord and stromal (n = 1); mixed germ cell tumor-mature teratoma, embryonal carcinoma, and endodermal sinus tumor (n = 1); and mature teratoma and choriocarcinoma (n = 1).

  6. e

    Wilms tumor: 1 case with bilateral Wilms tumor.

  7. f

    Multiple tumors: the median ages refer to the age at the time of diagnosis of the first and second primary tumors. Diagnoses included, in temporal order of diagnosis, acute lymphocytic leukemia and colorectal cancer, ependymoma and Burkitt lymphoma, Ewing sarcoma and Hodgkin lymphoma, osteosarcoma and retinoblastoma, pleomorphic undifferentiated sarcoma and retinoblastoma, rhabdomyosarcoma and bilateral retinoblastoma, testicular germ cell tumor and astrocytoma, and extramedullary Schwannoma and Wilms tumor.

  8. g

    Four probands were adopted and their cancer family history was therefore unavailable; the biological parents of 3 probands were unable to provide their cancer family history.

  9. h

    Some families fulfilled more than one LFL criterion.

  10. i

    A total of 49 families (16.8%) fulfilled Chompret or modified Chompret criteria.

  11. j

    A total of 21 families (7.4%) fulfilled only Eeles 1 criteria.

Sex   
Male17058.26.0 (0.16–17)
Female12241.86.5 (0.16–18)
Cancer diagnoses   
Adrenocortical carcinoma113.82.0 (0.33–17)
Sarcoma8127.79.0 (0.25–18)
Osteosarcoma217.212.0 (1–16)
Rhabdomyosarcoma299.94.0 (0.91–17)
Other sarcomasa3110.69.0 (0.25–18)
CNS6622.66.0 (0–17)
Choroid plexus carcinoma20.71.0 (0)
Astrocytoma175.86.0 (0.16–12)
Medulloblastoma206.86.0 (1–15)
Other CNS tumorsb279.36.5 (0–17)
Leukemiac7726.46.0 (1–17)
Germ cell tumord155.15.0 (0.5–17)
Wilms tumore3411.63.0 (0.4–16)
Multiple primary tumorsf82.75.0 (0.16–13) and 15.5 (11–22)
Family history of cancer (n = 285)g   
First- or second-degree relative, any cancer14550.78.0 (0.25–18)
First- or second-degree relative, breast cancer3311.58.0 (0.25–16)
Relative(s) with cancer aged <45 y10538.640 (19–45)
Relative(s) with cancer aged <18 y299.911 (0.5–18)
Family history criteria for LFS/LFL (n = 285)g   
Absent21374.75 (0–18)
Present (at least one)h7225.37.5 (0.25–17)
Chompret criteriai3411.98.5 (0.33–17)
Chompret modified criteriai4114.48.0 (0.33–17)
Birch criteria134.68.0 (1–16)
Eeles 2 criteria10.316 (0)
Eeles 1 criteriaj6221.76.0 (0.25–17)

Screening for the germline TP53 p.R337H mutation among patients with cancer identified 11 mutation-positive probands (3.8% of the total sample), including 9 patients who were diagnosed with ACC and 2 patients who were diagnosed with CPC (Fig. 2). One of the 11 ACC-affected probands (Fig. 2, pedigree 10) was homozygous for the p.R337H mutation.

Figure 2.

Pedigrees and genotyping results of mutation-positive probands are shown.

Figure 2 (Continued).

Genotypes for the proband and close relatives tested in each family are provided below each individual. Cancer-affected relatives are represented by blackened symbols. Arrows indicate the proband. Dx indicates age at diagnosis; WT, wild-type; CNS, central nervous system. Genotype results are given for each individual tested.

Loss of heterozygosity (LOH) at the p.R337H locus was observed in all tumors from the mutation-positive probands for which tumor tissues were available (Table 2). The mutation status of the parents was assessed in 10 of the 11 carrier probands. In 5 probands, the mutation carrier was the father, in 4 probands it was the mother, and in 1 proband (homozygous affected patient) both parents were carriers (Fig. 2). In 1 family, the parents were unavailable for testing. Among the 47 relatives tested, the p.R337H mutation was encountered in 2 of 2 relatives of the children with cancer (100%) and in 25 of 45 relatives of the children without cancer (55.5%). The Brazilian founder haplotype was assessed in at least 1 member of each of the p.R337H-positive families, for a total of 37 individuals, and was found to be present in all of them.

Table 2. Features of Pathogenic Germline TP53 Mutation Carriers Identified in the Current Study
PedigreeSexMutationCancer DiagnosisaAge, YearsStage of DiseaseLOH in Tumorp53 in Tumor, %bTreatmentMetastatic Disease/RecurrenceFollow-Up, Years Outcomec
  1. Abbreviations: BMT, bone marrow transplantation; CT, chemotherapy; LOH, loss of heterozygosity; M, metastatic; NA, not available; R, disease recurrence; RT, radiotherapy; S, surgery.

  2. a

    Adrenocortical carcinoma presenting symptoms: Cushing syndrome and virilization.

  3. b

    p53 expression by immunohistochemistry, results are shown as the percentage positive.

  4. c

    Outcome; dead indicates patient was deceased as of June 30, 2012.

1Malep.R337HChoroid plexus carcinoma1IIINANAS/CT/BMTM/R1Dead
2Malep.R337HChoroid plexus carcinoma1IIINANAS/CT11Alive
3Femalep.R337HAdrenocortical carcinoma0.50IVYes20S/CTM7Alive
4Malep.R337HAdrenocortical carcinoma1IYes40S2Alive
5Malep.R337HAdrenocortical carcinoma17IIINA0CTM1Dead
6Femalep.R337HAdrenocortical carcinoma1IIIYes40S/CT5Alive
7Femalep.R337HAdrenocortical carcinoma0.33IYes40S4Alive
8Femalep.R337HAdrenocortical carcinoma11IVYes70S/CTM2Dead
9Malep.R337HAdrenocortical carcinoma2IYes30S6Alive
10Femalep.R337HAdrenocortical carcinoma1IYes30S8Alive
NAFemalep.R337HAdrenocortical carcinoma5IIIYes0S/CTM/R1 
11Femalep.G245SMedulloblastoma10IVNA90S/CT/RTM/R3.3 

Direct TP53 sequencing of 48 probands who met LFS/LFL criteria (Birch and Chompret)[15, 16] and who did not carry the p.R337H mutation revealed a classic (pathogenic) DNA-binding domain mutation, p.G245S, in only 1 child (2.0%) who was diagnosed with medulloblastoma at age 10 years (Fig. 2, pedigree 11). In the remaining 47 probands that fulfilled LFL criteria (including the more relaxed modified Chompret criteria)[18], sequencing of exons 2 to 11 and multiplex ligation-dependent probe amplification of the TP53 gene did not identify deleterious germline mutations or gene rearrangements. In 3 probands fulfilling the Chompret criteria,[16] the available biological material was insufficient for full gene sequencing.

DISCUSSION

In the current study, we describe for the first time the TP53 p.R337H mutation frequency in a series of children diagnosed with tumors on the LFS/LFL spectrum in southern Brazil, who were recruited regardless of their cancer family history. Although the overall pathogenic mutation frequency was relatively low (< 5%) in this series, these results confirm the strong association between the mutation and ACC and CPC, as reported in other regions of the country.[25, 26, 28, 30, 32] We were unable to demonstrate an association between the mutation and pediatric sarcomas in this large series of cases (n = 81). Even among patients with osteosarcoma (n = 21), a tumor previously associated with the p.R337H mutation, no carriers were identified.[32] In addition, the mutation was also not encountered among patients with leukemia (n = 77) and Wilms tumor (n = 34), reinforcing a very specific cancer phenotype in carriers children that appears to be restricted to only a few tumor types. Furthermore, among the 11 mutation-positive probands, only 7 (64%) had a family history of cancer in first- or second-degree relatives. The absence of a family history of cancer in 4 mutation-positive probands reinforces the belief that germline TP53 mutation screening should be undertaken in any child with a diagnosis of ACC or CPC,[18, 38, 39] especially in Brazil, where one could initiate an investigation using p.R337H testing. When analyzing the tumors from probands who carried the p.R337H mutation identified herein, we observed a LOH and abnormal nuclear accumulation of p53, findings that have been commonly described in patients with LFS harboring classic DNA-binding mutations. Thus, the findings of the current study reinforce an important role for this specific mutation in the development of ACC and CPC, which is consistent with the classic model of loss of function mutations in a tumor suppressor gene.[40, 41]

We also demonstrated in all p.R337H-positive families available for testing that the mutation is present in the germline of at least 1 parent. It is interesting to note that all carrier parents were unaffected by cancer, confirming that partial penetrance is an important feature of this mutation. Furthermore, in all the mutation carriers identified, either probands or their relatives, the mutation occurred on the Brazilian founder haplotype.[35] Although a report from Europe has suggested the occurrence of the mutation on a different haplotype,[33, 34] to the best of our knowledge all Brazilian carriers identified to date, including the 37 from the carriers from the 11 families studied herein, apparently derive from the same common ancestor.

Another important result of the current study was the identification of criteria for LFL in a significant percentage (25.3%) of the probands, a finding that was not apparent in a small group of children who were unaffected by cancer and recruited from the same institution. The high percentage of children with cancer who had family histories consistent with LFL emphasizes the importance of educating health care professionals in obtaining a detailed cancer family history for every child diagnosed with cancer, especially those with cancers on the LFS disease spectrum.

Conversely, one of the most intriguing findings in this subset of patients meeting LFL criteria was the low frequency of germline TP53 mutations or rearrangements in the probands fulfilling Birch and/or Chompret LFL criteria.[15, 16] The mutation prevalence in this group (23.5%) was lower than expected based on observations published for other similar clinical series. When we consider that among the 12 mutation-positive cases, only 1 was not the founder Brazilian p.R337H mutation, the prevalence of classic DNA-binding domain mutations in this series is exceedingly low.[14-16, 22-24] In these families, the molecular mechanisms to explain the autosomal dominant cancer predisposition phenotype remains to be determined, posing a significant challenge for genetic counseling and management.

Conclusions

Genetic counseling and TP53 p.R337H testing should be offered as the first diagnostic approach to all Brazilian children diagnosed with ACC and CPC, and when negative, testing should be expanded to comprehensive germline TP53 mutation screening. Furthermore, the LFL phenotype is common among children with cancer who reside in southern Brazil and the proper identification and referral of at-risk families will require education and compromise among health care professionals directly involved with their care. Efforts to understand the molecular basis of the autosomal dominant predisposition of early–onset tumors observed in the syndrome, and of its genetic and nongenetic modifiers as well, should be undertaken to provide adequate management to affected children and their families.

FUNDING SUPPORT

Supported by an Ethnic Research Initiative Grant from GlaxoSmithKline, UK; Fundo de Incentivo a Pesquisa do Hospital de Clínicas de Porto Alegre, Brazil (FIPE-HCPA); and FAPERGS (PPSUS 2009), Brazil.

CONFLICT OF INTEREST DISCLOSURES

Dr. Giacomazzi has received grant and travel support from GlaxoSmithKline (Ethnic Research Initiative Grant), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior (CAPES), and Fundacao de Amparo a Pesquisa do Estado do Rio Grande do Sul (FAPERGS). Dr. Ashton-Prolla has received grant and travel support from GlaxoSmithKline (Ethnic Research Initiative Grant), is employed by the Federal University of Rio Grande do Sul, and has received academic research grants from Brazilian government funding agencies.

Ancillary