Taking their time in the womb: Grrl Power 

Nearly sixty years ago it was observed that while girls on average have lower birth weights than boys, they have longer gestations [Anderson & Brown, Am J Dis Child 1943; 65:523-34]. Further, de Zehgher and colleagues noted that the gestational differences between boys and girls increase with decreasing birth weight [Horm Res 51:258-59, 1999]. These observations suggest that boys have a faster growth rate in utero than girls, which could have implications for studies of birth weight and childhood and adult malignancies. In order to determine if the length of gestation is truly gender-driven, Loos et al [Lancet 358:560-61, 2001] examined data from dizygotic twin pairs of the East Flanders Prospective Twin Study in Belgium. Zygosity was determined using DNA fingerprint analysis. Birth weight and gestational age were obtained from obstetric records, and 1929 twin pairs (3858 children) were analyzed including boys of same sex twin pairs (n=936), boys of unlike-sex pairs (n=1008), girls of same-sex pairs (n=906), and girls of unlike sex pairs (n=1008).The mean gestation of unlike sex pairs was similar to that of girl same-sex pairs. However, unlike-sex pairs and girl same-sex pairs had a statistically significant longer gestation (p=0.02) than boy same-sex pairs. Moreover, the boys from unlike-sex pairs weighed 78 grams more than boys from same-sex pairs (p=0.001), even after controlling for the shorter gestation. This indicates that in unlike twin pairs it is the girl who determines the length of gestation, and that she prolongs the gestational length for her brother (which results in additional weight gain for him). 

COMMENT: These are fascinating observations and may be important in studies that evaluate the role of birth weight in the development of both childhood and adult malignancy. As indicated previously [see C3, Vol 12, No 1; Vol 10, No 3; Vol 9, No 2], high birth weight appears to be a risk factor for childhood leukemia. Loos et al speculate that since fast growing fetuses may be especially vulnerable to prenatal insults, the differences in the rate of development of males and females in the womb may help to explain some of the sex differences observed in subsequent disease.

Relapse- or maybe not? Return of the sibling clone. 

Approximately 20% of childhood ALL has a TEL-AML1 fusion gene. These cases typically have low or intermediate risk features and a good prognosis. In earlier elegant studies, Mel Greaves' laboratory has demonstrated that the initial gene fusion can occur in utero, and studies in twins suggest that this event alone is insufficient to generate overt ALL, and that some secondary postnatal event is required for leukemogenesis [Ford et al, Proc. Natl. Acad. Sci, 95, 4584, 1998; Wiemels et al, Blood, 94, 1057, 1999]. In a new study, Ford et al [Blood 98, 558, 2001] have examined the molecular changes in 3 TEL-AML1 leukemias at diagnosis and at relapse. The authors speculated that in cases of late relapse of TEL-AML1 ALL, combination chemotherapy has successfully eradicated the overt leukemic clone, but has failed to eradicate an underlying pre-malignant clone. This clone would have the potential to acquire new secondary genetic alterations, and hence generate a new TEL-AML1 leukemia. Commonly, the normal copy of TEL is deleted as one of the secondary changes associated with development of frank leukemia in TEL-AML1. The authors showed that in 2 children who relapsed late (off therapy) the leukemic cells at relapse had different TEL gene deletions than were present at diagnosis, and these changes could not be the result of evolution of the original clone. Despite this, studies of the T-cell receptor and immunoglobulin gene rearrangements showed the two leukemias to be related ("sibling clones"). Study of one child who relapsed on therapy showed that this relapse occurred in the original leukemia cells. Taken together, these data support the hypothesis that at least some TEL-AML1 cases who have a late relapse really have a new leukemia. If this is true, the leukemia should be as amenable to treatment and cure with chemotherapy as a newly diagnosed case and in clinical practice this appears to be the case. This hypothesis also predicts that it might be possible to detect TEL-AML fusion genes in the blood of at least some apparently cured cases- a testable hypothesis. 

COMMENT: This study and others from the same laboratory illustrate the power of careful studies of small numbers of unusual and highly informative cases. (See discussion of twins in adjacent column). Collection and sharing of biological material on such cases, and especially paired diagnosis and relapse samples is vital for continuing progress. Of course, the insight to recognize the importance of such cases is helpful too. 

Two is better than one? 

Although the concordance of cancer in young twins, particularly leukemia, is well-documented, several report suggest that twins are at a decreased risk of developing malignancy compared to singleton births. Murphy et al [Br J Cancer; 84:1460-1462, 2001] identified 13,009 twins born between 1963 and 1989 in the Oxford Record Linkage Study (ORLS). The ORLS is believed to be nearly population-based, and has the advantages of lack of selection, and information on individual mortality and aggregate migration. With linkage to the National Registry of Childhood Tumors (NRCT), 15 childhood malignancies were identified in this cohort, including one pair concordant for leukemia. Expected values were calculated for sex-, age-, and diagnostic groups of childhood malignancy by applying NRCT incidence rates to the live twin births. By age 15, approximately 18.9 cancers would be expected in this group, producing a non-statistically significant SIR of 0.79 (95% CI=0.39-1.20). Fewer cancers than expected were also observed for every age group as well as for each gender. These data support the results of prior prospective studies. 

COMMENT: The authors suggest that the tendency to lower birth weights of twins (perhaps because of fewer cell divisions or altered growth factor exposure) may reduce risk of malignancy. In particular, they compared the observed to expected incidence of cancer by birth weight, and reported significant deficits for babies weighing below 3000 g but excesses for those weighing 3000 g or more. This provides further evidence that birth weight may be important in malignancy. 

EDITORS' NOTE 

Twins and leukemia : How big is the risk to the second twin? 

Much has been learned about the biology of childhood leukemia from studies of twins (see above) and much of this has been led by the laboratory of Mel Greaves, which has sought out these rare cases, collecting data on 16 pairs of twins with leukemia. In a personal communication, Professors Greaves (Leukemia Research Fund Laboratories, London, UK) and Eden (Christie Hospital, Manchester, UK) estimate the leukemia concordance rate for infant twins diagnosed in < 12 months of age with ALL or AML (and generally an MLL gene rearrangement) and with a single monochorionic placenta (~ 60% of identical twins) to approach 100%, due to a shared placental circulation and blood cell chimaerism. For older children with ALL, the concordance rate is closer to 5-10%. There are no available data for older children with AML as the occurrence is so infre-quent. Professors Greaves and Eden would be interested to hear reports of other concordant pairs and particularly of any known discordant infant twin cases, and would be interested in a research collaboration (m.greaves@icr.ac.uk). They recommend the following approach when one twin is newly diagnosed with leukemia: 

• Ascertain zygosity, preferably using molecular techniques. 

• Ascertain whether there was one or two placentas at birth. Note, two placentas does not preclude twin-to-twin transfer of leukemia, but the risk in this circumstance is uncertain. 

• Advise parents of risk based on the above. Check a blood smear on the healthy twin immediately, and every 3 months for 2 years, then less frequently. While risk declines at age 10 years, leukemia can emerge in a co-twin more than a decade later [Wiemels et al Blood 1999;94:1057).

• Once the genetic features of the presenting twin's leukemia are known, the healthy twin should be assessed for the same markers. Remember that this may represent a shared fetal pre-leukemic clone and initiation of therapy is only justified if the presence of molecular markers is supported by conventional hematological evidence of disease. This approach may facilitate an early diagnosis when WBC counts are low and hence an improved prognosis. 

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