Epidemiologic, genomic, and twin studies indicate a multistep pathogenesis of B-cell ALL in infants and children, with an initial leukemogenic event(s) occurring in utero and subsequent genomic changes occurring postnatally. Most of the current thinking of the natural history of childhood ALL was based on the exquisite work of Greaves and co-workers who studied the clonal origin of leukemia in monozygotic twins 30 years ago. Greaves found that most common chromosomal translocations and their resultant gene fusions were present on neonatal blood spots or Guthrie cards. If a unique gene fusion sequence is present in at least one cell per 30,000 in the peripheral blood, it can be detected by a sensitive PCR assay. The first observations of an in utero origin of an acute leukemia were demonstrated with KMT2A rearrangements in three children with ALL, in monozygotic twins who shared the identical fusion, and in a pair of twins diagnosed at age 3 with B-precursor cell ALL. Subsequently in utero occurrence of ETV6-RUNX1, BCR-ABL1, E2A-PBX1, and loss of IKZ1, PAX5, biallelic loss of CDKN2A, and hyperdiploidy were documented prenatally on Guthrie cards. The concordance rate of twins who share a monochorionic placenta and develop leukemia is nearly 100%, whereas older twins have a discordance rate of 90%, indicating that additional postnatal leukemic events are needed. The frequency of an in utero origin of a B-cell precursor ALL in patients who are not twins as measured by clonal IGH rearrangements is reported to be 71%, supporting the notion that in the majority of infants older than 1 year, the development of ALL occurs during fetal development because they share a single placenta resulting in blood cell chimerism. The prenatal mutation occurs commonly, exceeding the actual rate of developing leukemia by some 100-fold, indicating a low rate of evolution. The acquisition of the additional genetic lesions, such as copy number alterations of ETV6, PAX5 and CDKN2A (25% to 75% of cases) are recurrent and contribute to leukemogenesis. After testing five pairs of monozygotic twins with concordant ETV6-RUNX1-positive ALL, Greaves and his colleagues demonstrated that all recurrent driver mutations detected (32 in total) were different within twin pairs, indicating that they are postnatal in origin in both twins and most likely represent secondary mutations. Single cell analyses have revealed considerable complexity within a tree-like or branching structure of genetically distinct subclones, as Greaves and col leagues first suggested. Current molecular data suggest that KMT2A fusions may be sufficient to cause leukemia since they are found in aborted fetal tissues (Fig. 1). However, formation of ETV6 RUNX1 fusion in utero is not sufficient to cause leukemia and postnatal mutations are required, such as a loss of a normal ETV6 from the other homologue. Formation of ETV6-RUNX1 has been shown, using qRT-PCR and FISH, to be present in ∼1% of new borns. Current evidence suggests the frequency of ETV6-RUNX1 incidence in a newly born is estimated to be very low, about 10−4. Prenatal origin of E2A-PBX1, as mentioned above, was also traced to Guthrie cards and was detected in cord blood of healthy new borns at the frequency of 0.6% (see box on Genetic Testing for Acute Lymphoblastic Leukemia).
Fig1. TIMELINE OF ALL DEVELOPMENT. The initiating lesions (green) occur in utero and lead to a state of preleukemia after birth. Exposure to infection leads to a dysregulated immune response in a small fraction (about 0.2% to 1%) of preleukemic children. Most preleukemic children remain healthy. The children develop ALL by acquiring secondary mutations, eventually leading to clonal expansion. In the case of KMT2A rearrangements (KMT2A-r, right), it is not completely clear whether the KMT2A-r are sufficient for ALL development (right dashed arrow). It is also possible that the KMT2A-r directly trigger secondary mutations (central dashed arrow). One case of a healthy KMT2A-r carrier has been reported leading to the possibility of a preleukemic state for those cases (left dashed arrow). The given median ages of ALL onset are 2–5 years of age for B-cell precursor subtypes (BCP, left) and below 1 year of age for pro-B-cell subtype (pro-B, right). ALL, acute lymphoblastic leukemia. (Reproduced with permission from Hein D, Borkhardt A, Fischer U. Cancer and Metastasis Review, 2020;39:161–171.)
