Neurofibromatosis, cancer, and family history of colon cancer

Hereditary non-polyposis colon cancer (HNPCC) is an autosomal dominant disorder associated with early onset of colon cancer. Heterozygous mutations in the mismatch repair (MMR) genes, especially hMSH2 and hMLH1, account for most cases. In two related articles recently published, five children from two consanguineous marriages within HNPCC families were described. [Ricciardone MD, et al., Cancer Res 59:290-3, 1999; and Wang Q, et al., Cancer Res 59:294-7, 1999]. Four of the five children had clinical features of NF1 including one patient who had only hemicorporal café au lait spots, suggesting post-zygotic somatic mutation. All five children developed malignancies by the age of six: two cases of non-Hodgkin's lymphoma, one case of presumed JMML, one case of AML with subsequent medulloblastoma, and one case of "acute leukemia." DNA sequence analysis documented homozygous MLH1 mutations in affected children from both families. These cases strongly suggest that homozygous mutations in MMR genes lead to a mutator phenotype, post-zygotic somatic "acquired" mutation in the NF1 gene, and subsequent manifestations of both neurofibromatosis and early onset of cancer. 

COMMENT: These are fascinating reports which document previous suspicions that mutator phenotypes increase the risk of subsequent genetic changes; in this case, the development of neurofibromatosis in two consanguineous families. It is known that the mutation rate in NF1 is about 10-fold higher than most other genes and hence more apt to be a secondary event in mutator phenotypes. Although NF1 mutations were not directly studied in either family, the clinical evidence is overwhelming that this was the case.

There are several interesting implications from these papers. First, with cancer at an early age in all five children, the mutations in both MMR and NF1 genes probably greatly predisposed to the development of cancer. Second, the malignant phenotype appears somewhat different from that expected, with two of the five children demonstrating non-Hodgkin's lymphoma rather than the expected myeloid malignancies and/or brain tumors. Although NHL has been reported in neurofibromatosis, it is less common than the malignancies noted above. The additional mutation in MLH1 may have predisposed to NHL and hence be important in its pathogenesis in other settings. Third, why some children with NF develop hematologic malignancies and others do not has been a matter of speculation. The studies noted above "cry out" for examination of heterozygous deletions in the MMR genes in 
patients with NF who develop heme malignancies. Finally, since 50% of all NF patients have de novo mutations without family history, many may be "predisposed" by the presence of either heterozygous or homozygous abnormalities in MMR genes. These families point out the ongoing importance of careful clinical observation in unusual patients and/or families. The role of the clinician in cancer research remains strong. William G Woods

Retinoblastoma - the genetics are getting harder all the time

Retinoblastoma is traditionally considered to be a disorder with very high penetrance; i.e., almost everybody who inherits the abnormal gene will display the disease. A small number of families are described as showing low penetrance of disease. Some of these families, linkage analysis shows that children without retinoblastoma appear to have inherited the chromosome known to carry the abnormal gene. A new report [Munier et al., Am J Hum Genet 63:1903-8, 1998] demonstrates that in some of these cases, the absence of the disease phenotype is not due to low penetrance (i.e., inherited abnormal gene but has no manifestation of disease), but to germinal somatic mosaicism in the parent. In the family described in this report, the father had one eye enucleated for retinoblastoma, and had a unifocal flat chorioretinal scar in the untreated eye; his affected son had bilateral multifocal retinomas. Linkage analysis was able to show which of the father's chromosome 13's was associated with the occurrence of retinoblastoma as it was shared with his affected son. A second unaffected son was also shown to carry the same chromosome and would have been expected to have retinoblastoma. The investigators showed direct evidence of gonadal mosaicism in the father (the presence of sperm carrying the haplotype linked with the mutated gene, some of which had a normal RB gene and some of which had a mutated RB gene). The investigators PCR amplified 394 different sperm and showed that 93% had a normal Rb allele and 7% had a mutant allele. It can thus be inferred that the mutation in the affected father occurred post-zygotically (i.e., after fertilization). Because the mutation was clearly present in at least some retinal cells (because the father had retinoblastoma) and in some spermatozoa, the authors were able to deduce the time post conception when mutation occurred. The retina is ectodermal in origin and primordial germ cells derived from blastocyst endoderm. This differentiation starts about eight days after conception, so the authors suggest that the mutational window when this event took place can be assigned to the time shortly before fertilization to approximately the 8th day of the father's embryonal development. 

COMMENT: This study shows that hereditary retinoblastoma doesn't originate exclusively from new mutations occurring in gametes, but can also occur from mutation of the embryo. This extremely accurate determination at the time the mutational event occurred is quite remarkable. Mosaicism may complicate the pattern of inheritance of a number of cancer predisposition syndromes and would clearly make invalid the normal assumptions used for genetic counseling. SM Davies 

Osteosarcoma and Ewing's sarcoma: Boning up on the latest epidemiological studies

Little is known about the etiology of childhood osteosarcoma (OS) or Ewing's sarcoma (ES), two relatively rare neoplasms of the bone. Since both malignancies affect children primarily in their pubertal years, some studies have focused on bone growth (such as attained height) or bone trauma, with inconsistent results. There is some suggestion that genetic factors may play a role in the etiology of both tumor types. Birth defects in individuals or family members have been associated with ES. Moreover, osteosarcoma is the second most common cancer in patients with hereditary retinoblastoma [McKeen et al, NEJM 1983; 309:496-7; Abramson et al, Ophthalmol 1984; 91:1351-55]. However, these familial patterns account for only a small proportion of cases. Finally, a few exploratory studies of ES have generated some modest associations with paternal occupations in the agricultural industry and maternal tranquilizer use [Holly et al, Am J Epidemiol 1992; 135:122-29; Winn et al, Can Epi Biom Prev 1992; 1:525-32]. Larger case-control studies of childhood OS and ES have been conducted recently in the United States and Canada (summarized below). 

In a study from Toronto [Hum et al, Int J Epidemiol 27:766-771, 1998], histologically confirmed cases of bone cancer in patients < 25 years of age (diagnosed during the period 1980 and 1988) were identified from the Ontario Cancer Registry. Population-based controls were selected by telephone and matched to cases on age and sex. A total of 186 case mothers and 919 control mothers completed a self-administered mailed questionnaire focusing on parental employment (overall response rate was 86% for case mothers and 81% for eligible control mothers). For each parent, 21 occupational categories were examined. For fathers, employment in the "social sciences" was positively associated with an increased risk of bone cancer (OR=2.5, 95% CI=0.7-8.4). When data were stratified on subtype (OS versus ES), this association was only apparent for ES (OR=6.2, 95%CI=1.6-24.5]. Paternal occupations involving farming and animal husbandry were also associated with an elevated risk of OS but not ES. For maternal occupations, elevated risks were most notable for managerial and administrative occupations and OS (OR=2.3, 95%CI=0.8-2.7), and farming and animal husbandry and ES (OR=7.8, 95%CI=1.9-31.7). Although the authors admit some methodological limitations, they suggest further investigation, particularly for agricultural occupations. 

From the Children's Cancer Group, Buckley et al [Cancer 1998; 83:1440-8] report findings from the analyses of two case-control studies, including 152 children with OS and 153 children with ES, respectively, and their corresponding matched control pairs (age, race) obtained through random-digit dialing. The most significant finding was a substantially lower gain in weight and height during the pubertal growth spurt for both ES and OS as compared with controls. For specific occupational exposures of either the mother or the father (including solvents, plastics, paints, oil/tar, etc), no statistically significant relationships were observed. Further, no statistically significant associations were noted for maternal or paternal height, size at birth, body size, male pubertal factors (including voice changes, development of pubic or facial hair), or body size for females. There was a suggestion that girls who developed OS had an earlier onset of breast development and menarche. Finally, other factors explored including maternal medications or illnesses at the time of pregnancy, and maternal diet showed few significant relationships. Taken together, the authors conclude that no important risk factors were identified, and that given the relative absence of any positive associations, little can be learned from case-control studies of either OS or ES at this time. 

COMMENT: In the first study by Hum et al, several employment categories were explored. One of the notable limitations of this study is the lack of adjustment for socioeconomic status (i.e., education, income), since it is unclear why occupations in the "social sciences" would be associated with an increased risk. Moreover, these authors emphasize the positive associations with the agricultural industry (which was not specifically explored in the Buckley et al study). However, it was interesting to note that the positive associations with farming and agriculture were apparent for fathers and OS (not ES) and mothers and ES (not OS). Given the thorough data analysis by the CCG group, it seems reasonable to speculate that little will be gained from another epidemiologic study of either ES or OS. However, this does not negate the fact that a future study that incorporates biological or genetic markers (yet to be determined!) may prove fruitful. In particular, both authors note the long-recognized deficit of ES among African Americans. The question, of course, is why? Julie A. Ross 

Brief commentary: Going to the dogs!

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