LOH down on MLL 

The involvement of the MLL gene at 11q23 in infant and therapy-related leukemias has been well-described [see C3 Vol 9, No 1; Vol 6, No 5].  Although the function of the MLL gene product has not yet been determined, it is thought to be a crucial regulator of a developmental (homeobox) gene. Recent studies describe the loss of heterozygosity (LOH) at MLL in childhood leukemia. LOH (2 alleles in normal tissue, but only one allele seen in tumor tissue) is a marker of loss of genetic material in malignant cells. Since the remaining allele is usually mutated, LOH often suggests a gene is acting as a tumor suppressor gene.

Takeuchi et al [Oncogene 18: 7387-7388, 1999] evaluated the occurrence of loss of heterozygosity (LOH) at 11q in 113 primary childhood ALL samples using 14 microsatellite markers. They found that 18 (16%) patients had LOH of 11q. Further analyses revealed three distinct regions of deletion. The first, at 11q22-23, contained the ATM gene (not MLL), although mutational analysis suggested that the ATM gene was not altered. The second and third regions affected contained the MLL gene. Interestingly, all of the patient samples with LOH at MLL lacked detectable MLL gene rearrangements. Further, 20 children were studied at initial diagnosis and relapse; none of these children who relapsed acquired LOH at 11q, possibly suggesting that 11q deletions are infrequently involved in the progression of childhood ALL. 

Webb et al [Blood 94: 283-290, 1999] describe LOH at MLL in children with leukemia (both ALL and AML). In this study two groups were evaluated, 18 infants (< 18 months of age)  and 38 randomly selected older children who presented with leukemia. Cytogenetic analysis revealed that 9/18 infants demonstrated MLL rearrangements with other chromosomes; however, none of the 18 infants had LOH at 11q. In contrast, for the older children, no translocations involving MLL were observed, although LOH at MLL was found in 9 of 20 patients informative for the microsatellite markers used. Of the informative patients, 4 had AML, of which 3 (75%) demonstrated LOH at MLL. The authors suggest that LOH at MLL might be a common event in childhood leukemia.

COMMENT: It is interesting to speculate that  MLL may affect (or be associated with) the development of leukemia differently in children of various ages. Infants with ALL have the highest proportion (80%) of translocations involving MLL (usually partnered with chromosome 4). For infants with AML, the frequency of MLL translocations are lower (50-60%), and typically involves chromosomes 9 and or 19 (similar to therapy-associated leukemias). Trans-locations are often hallmarks of oncogene activation (e.g., bcr-abl, cerb-B2). It has been speculated that MLL rearrangements are the initiating event for infant leukemias as they have been shown to occur in utero. In contrast to translocations, the above two studies observed LOH at MLL in older children with leukemia. LOH suggests the loss of a tumor suppressor gene (e.g., Rb1 (retino-blastoma), BRCA1,2 (breast cancer), p53 (several malig-nancies). It is possible that MLL functions in two different capacities depending on the differentiation stage of the hematopoietic cell type affected. The RET proto-oncogene has some interesting properties depending on the type of genetic involvement and the tissue affected [see C3 Vol 5, No 2].  Germline deletions and mutations of RET in some tissues are associated with a loss of function, whereas translocations in other tissues are associated with a gain of function. Further studies of MLL involvement in leukemias, as well as other hematologic and non-hematologic malignancies may shed further light on this issue.  Julie A. Ross

The Double Whammy? Germ cell tumor and leukemia

The occurrence of a mediastinal germ cell tumor and leukemia in the same patient has been described in case reports and small series since 1985 [Nichols et al, Ann Int Med 102:603-9].  In an analysis of 635 patients with extragonadal germ cell tumor, Hartmann et al [J Natl Cancer Inst 2000; 92:54-61] define this entity more clearly.  Of note, the cases of leukemia all occurred in patients with a primary non-seminomatous mediastinal germ cell tumor.  No cases were recorded in 283 patients with retroperitoneal germ cell tumors or 51 patients with seminomatous mediastinal tumors.  The authors calculate that approximately 1 in 17 patients with a non-seminoma-tous germ cell tumor will develop leukemia. The leukemias reported typically occurred in the first year after diagnosis of the germ cell tumor (median 6 months, range 0-47 months), and bore none of the characteristics of a therapy-related leukemia. Importantly, in five leukemias the germ cell tumor marker chromosome i(12p) was found.  The leukemias were mainly characterized as megakaryocytic (FAB M7). Prognosis was extremely poor, with none of the reported patients surviving.  Seven of the 17 cases were so fulminant they received no therapy and died rapidly. Eight received combination chemotherapy and only one achieved a brief remission.  Two received bone marrow transplant; one relapsed early and the other died of treatment-related complications.

COMMENT: These data suggest that non-seminomatous germ cell tumors and leukemias can arise from a common progenitor cell.  The authors note that the hematologic malignancy might be arising in areas of yolk sac tumor, then disseminating to the bone marrow, recapitulating the normal embryonal onset of hematopoiesis within the yolk sac followed by migration to the liver and spleen then bone marrow. The reason for the preponderance of megakaryocytic differentiation remains unclear.  A number of cases also had trisomy 8, a common abnormality in myeloid leukemia, which may support a multi-step carcinogenesis pathway.   Stella M. Davies

Malignancy in individuals with Down syndrome

The increased risk of leukemia in children with Down syndrome has been well-described, with an estimate 10-20 fold increased risk compared to children without Down syndrome. However, little data exist with respect to the risk of other malignancies or the risk of leukemia later in life for individuals with Down syndrome. Hasle et al [Lancet 355: 165-169; 2000] describe the incidence of cancer in 2814 individuals with Down syndrome registered in the Danish Cytogenetic Registry between 1968 and 1995.  Malignancies were identified by linkage with the Danish Cancer Registry, which has received data from all clinical and pathological departments in the country since 1943. Expected numbers of cancers were derived using rates calculated from all Danish inhabitants specific for age, sex, and period. Overall, there were 60 cancers found and 49.8 expected, which was not statistically significant (standardized incidence ratio (SIR)=1.20; 95% CI=0.92-1.55)). When site-specific malignancies were evaluated, however, leukemias accounted for 60% of cases overall, and 97% of cancers in children under 15 years of age. The rate of leukemia was highest for children under 5 years of age (SIR=56.4;95% CI=37.8-81.0) but declined substantially thereafter; the cumulative incidence by the age of 5 years was 2.1% and by the age of 30 years, 2.7%. There was almost a four-fold higher incidence of AML compared to ALL before the age of 5; whereas the rate ratio was similar for individuals older than 5. Interestingly, with the exception of germ cell tumors and retinoblastomas, there was a notable lack of other malignancies in both children and adults with Down syndrome. In particular, there were no cases of breast cancer observed when about 7 cases would have been expected. The authors note that these data provide further evidence for a role of specific genes on chromosome 21 in the development of leukemia in children. Moreover, with the observed lack of epithelial cancers found in this group, they suggest that either exposures, genes on chromo-some 21 (such as an extra copy of copper-zinc superoxide dismutase to help decrease free oxygen radicals), or a combination of the two may decrease the risk of adult-type malignancies.

COMMENT: Questions often arise regarding the risk of leukemia in older individuals with Down syndrome, as well as the risk of other malignancies; this study has nicely addressed these issues. As noted previously [see C3, Vol 9, No 3], there has been an observed lack of neuro-blastoma in children with Down syndrome. It is interesting to speculate that the lack of certain malignancies in individuals with hereditary conditions may point to a potential protective influence of the specific genes involved. These types of observations underscore the ability of epidemiologic studies to help us understand the biological nature of disease. Julie A. Ross 

More Chernobyl fall-out

An increased frequency of pediatric papillary thyroid carcinomas has been reported in children exposed to radiation fall-out after the Chernobyl accident.  A specific type of RET gene rearrangement, RET/PTC3, is highly prevalent (50-58%) in these tumors but is infrequent in tumors from unexposed children [See C3, Vol 7 No 1].  The histology of these tumors is also specific, with a particular type of solid growth papillary carcinoma that is rare in sporadic adult or pediatric populations. Nikiforav et al [Oncogene 18:6330-6335, 1999] have characterized the breakpoints in these gene rearrangements in Minsk, Belarus. The study incorporated 22 post-Chernobyl pituitary carcinomas with RET/PTC3 rearrangement. Characteristics of these breakpoints suggest that the rearrangements were not caused by homologous recombination and they carried none of the molecular features of rearrangements associated with topo-isomerase-II activity or recombinases.  In contrast, the authors suggest that by analysis of the position of the breakpoints in the two affected genes, their data show signs of the passage of a single radiation particle as the cause of the concerted double strand DNA breaks in the two chromosomal fragments, which are closely spaced in the thyroid cell nucleus. 

COMMENT: This study demonstrates quite directly a potential cause of carcinogenic genetic damage.  How generalizable this mechanism of recombination is to the cancers caused by background radiation is unclear, but further biological studies may allow this question to be addressed.  Stella M. Davies

Ultrasounds and leukemia: It’s time to stop looking

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