Fanconi anemia (FA) is a well-known autosomal recessive disorder characterized by chromosomal instability and predisposition to cancer.   While bone marrow failure and acute myelogenous leukemia in FA patients are very familiar to pediatric oncologists, it is less well-recognized that as these patients age there is a high incidence of squamous cell carcinomas (SCC).   SCCs are particularly common in the head and neck and anogenital regions.   It is estimated that Fanconi anemia patients have a 500-700 fold higher incidence of head and neck SCC than the general population and a 14% cumulative incidence of head and neck cancer by the age of forty years [Kutler DI et al Arch Otolarynology Head and Neck Surgery 2003;129:106-12; Rosenberg PS et al Blood 2003;101:822-6] .   In the general population, SCCs of the head and neck are typically associated with tobacco or alcohol use and occur at an older age. Only a minority Fanconi patients (16%) report tobacco or alcohol use. Therefore, the etiology of SCC in FA patients may be different than that of the general population.   In a new study, Kutler, DI et al [Journal of the National Cancer Institute, 2003; 95:1718-21] have performed a case- control study in which case subjects are FA patients diagnosed with SCC of the anus, vulva or head and neck and control subjects were SCC patients who did not have FA and were matched to the case subjects in a 2 to 1 ratio for age, sex and site of primary tumor.   Matching for tobacco and/or alcohol exposure status was not possible because of the small number of appropriate cases in the tumor bank.   Twenty-five tissue and blood samples were procured from 24 FA patients who had pathologically confirmed SCCs (in collaboration with the International Fanconi Anemia Registry).   The control subject's tissues were obtained from the Memorial Sloan Kettering Cancer Center Tumor Bank. The data showed that human papilloma virus (HPV) DNA was present in 84% of the SCC specimens from the case subjects and in 36% of the SCC specimens from the control subjects (P=0.001).   The prevalence of p53 mutations in SCCs from the case subjects (0%, 0 of 25) was statistically significantly lower than that of SCCs from the control subjects (36%, 12 of 33; P<.001).   In addition, a greater proportion of patients with FA and SCC were homozygous for the arg72 p53 polymorphism than an ethnically-matched cohort of FA patients without SCC (75% vs 51%; P=0.05).   The authors suggest that patients with FA have increased susceptibility to HPV-induced carcinogenesis.  

 

COMMENT: This study further expands the information about carcinogenesis from patients with DNA repair syndromes. The relatively small number of cases in each group is a limitation of this study ,   However, FA is a rare condition and identification of a larger cohort is unlikely. Unfortunately, there are no data regarding sexual behavior in cases or controls as this is a well-known route of transmission of HPV.   Despite these limitations, the data are convincing.   The authors of this study suggest that the SCCs seen in FA patients probably are caused by the inactivation of p53 by HPV-associated oncoprotein rather than by direct mutagenesis. This observation has immediate clinical relevance as strategies for immunization against infection with HPV are currently being investigated.   Patients with FA who have not been infected with HPV represent an excellent population for investigation of this strategy.

Stella M. Davies

 

More Triple Jeopardy: Breast Cancer, Brain Tumors and Fanconi Anemia

 

We have reported previously [C3 Volume 13 Number 4)] that BRCA2 is a FA gene.   The clinical implications of this finding for FA families are starting to be investigated.   In a new paper, Offit K et al [Journal of the National Cancer Institute 2003; 95: 1548-51] have described four kindred's with FA and BRCA2 mutations.   One kindred of Ashkenazi Jewish ancestry had five members who were diagnosed with breast cancer and two cousins who were BRCA2 mutation compound heterozygotes who had FA and brain tumors.   In another kindred of Ashkenazi Jewish and Lithuanian Catholic ancestry, a child with FA and medulloblastoma was shown to be a BRCA2 compound

 

heterozygote also.   Two other kindreds each contained a FA afflicted child who developed medulloblastoma; one child was of Latin American ancestry and the other was African American. The median age of the FA children with brain tumors was 3 years.   The co-occurrence of brain tumors, FA and breast cancer observed in one of these kindred's constitutes a new syndrome association. The authors recommend that individuals with a germline BRCA2 mutation who plan to have children with part Ashkenazi Jewish descent should consider undergoing genetic counseling.  

More information on this topic comes from an abstract presented in San Diego at the American Society of Hematology meeting [Auerbach AD et al Blood 2003; 102:11, 23a abstract] . This abstract reports a total of fourteen children from the International Fanconi Anemia Registry with biallelic BRCA2 mutations.   Of these 14, early-onset brain tumors (primarily medulloblastoma) occurred in 5 and Wilms' tumors occurred in 2 children. Six children in five kindreds exhibited the co-occurrence of BRCA2 mutations, FA, and early-onset acute leukemia (5 AML and 1 ALL).   Early-onset breast cancer was noted in 4 of the 5 kindreds.   The cumulative incidence of hematologic malignancy was significantly higher among FA patients with BRCA2 mutations in comparison with FA patients without BRCA2 mutation (p=0.0001) . T t he cumulative incidence of hematologic malignancy at age 5 years was 41% for BRCA2 patients compared to 0.7 % for non-BRCA2 patients. These data suggest that BRCA2 testing should be considered for all FA patients in whom the complementation group cannot be defined or in whom leukemia is diagnosed at or before 5 years of age.

 

COMMENT: These data extend our understanding of familial aggregation of cancer in families affected with FA. They also illustrate the continuing value of registries for rare diseases   that allow rapid and extensive clinical evaluation of new biological findings.  

Stella M. Davies

 

My Big Fat Greek Baby!

 

IGF-I, IGF-II, and IGF binding protein 3 (IGFBP-3) are components of the insulin-like growth factor (IGF) system, which regulates cell proliferation.   Accumulating evidence suggests that the risk of some childhood and adult cancers increases with increasing levels of IGF-I.   Since IGF-II is active mainly during gestation, it may play a similar role to IGF-I for those childhood cancers that are initiated in utero .   A recent Greek study [Skalkidoua   A et al Cancer Epi Biom Prev 2003; 12:860-865] characterized the factors that determined, or were determined by, serum IGF-1, IGF-II, and IGFBP-3 levels in 331 healthy full-term ( > 37 weeks) infants born at the University of Athens teaching hospitals.   Only babies who were clearly jaundiced (serum bilirubin > 12 mg/dl; n=209) or who were clearly not (serum bilirubin < 8 mg/dl; n=122) were included in the study, since pilot data suggested that serum IGF-I levels are inversely related to serum bilirubin levels.   Potential predictors of serum IGF-I, IGF-II, and IGFBP-3 were maternal age, maternal body mass index (BMI), maternal smoking during pregnancy, gender of baby, gestational age, birth order, birth weight, birth length, ponderal index (a measure of fetal growth derived from birth weight and height), total infant blood proteins, infant blood creatinine, and neonatal jaundice. Presence of associations between these factors and log-transformed serum levels of the IGF system components were simultaneously assessed using linear regression. Predictors of each IGF component were analyzed separately. Additionally, predictors of IGF-I and IGF-II were adjusted for serum IGFBP-3 levels to account for the latter compound's influence on the levels of unbound, and hence biologically active, growth factor.

 

IGF-1 levels were significantly inversely associated with maternal age (p = 0.01), jaundice (p = 0.06), and blood creatinine (p= 0.0001).   Birth order (p = 0.04), weight (p = 0.0001), length (p = 0.0004), and ponderal index (p=0.0003) were significantly positively associated with IGF-1 levels.   IGF-II levels were significantly inversely associated with creatinine (p = 0.003) and jaundice (p = 0.002) and were significantly positively associated with total blood proteins (p = 0.002).   Lastly, IGFBP-3 levels were significantly positively associated with gestational age (p = 0.05), total blood proteins (p = 0.02), birth weight (p = 0.03), and birth length (p = 0.04) and were significantly inversely associated with creatinine (p = 0.003) and jaundice (p = 0.01).

 

COMMENT: This study provides a window into the biology of the IGF system, which may affect the risk of cancer in children and adults. In their discussion, the investigators noted the likely direction of causality of the observed associations. Maternal age and birth order affected IGF-I levels as in previous studies, while there was no such association with IGF-II levels.   Also as in previous studies, IGF-I, but not IGF-II, levels affected perinatal measures of infant size. I GFs are produced mainly by the liver, but also by the kidneys. Thus, the strong inverse relationship between creatinine and bilirubin (as reflected by jaundice) levels and IGF system components makes sense, since high levels of the former indicate impaired liver and kidney function.   These findings may help researchers interpret the relationship between perinatal factors and cancer.  

Logan G. Spector

 

Vitamin K- Putting To Rest a Shot in the Dark

 

Following one study that raised the question as to whether intramuscular vitamin K (in comparison to an oral route) increases the risk of childhood cancer [Golding J et al BMJ 1992; 305:341-346] ,   the vast majority of subse-quent studies have found no elevated risk.   Now, a large case-control study from the United Kingdom that compared obstetric and neonatal records for 2530 children diagnosed with cancer and 4487 control children reports no difference in risk of childhood cancer overall or leukemia with route of administration of vitamin K [Fear NT et al BJC 2003; 89:1228-1231] .   This study follows a recent meta-analysis of six case-control studies that reached the same conclusion [Roman E et al; BJC 2002; 86:63-69] . While it is often difficult to rule out certain risk factors for childhood cancer, we are fairly convinced that i.m. vitamin K is unlikely to contribute.

Julie A. Ross  

 

C3 Quarterly Newsletter
Children's Cancer Research Fund
Epidemiology Research Unit
Division of Pediatric Epidemiology
Clinical Research
University of Minnesota
420 Delaware St. SE, Box 422
Minneapolis, MN 55455
pedsepi@umn.edu

Editors: 
Stella M. Davies, MD, PhD, and Julie A. Ross, PhD