NOTCH1 mutation can be an early, prenatal genetic event in T-ALL

There are substantial evidence-based studies of twins with leukemia, retrospective scrutiny of archived neonatal blood spots, and cord blood screening that most pediatric leukemias are initiated in utero.^1,2  This conclusion applies to B-cell–precursor acute lymphoblastic leukemia (ALL) with TEL-AML1^3,4  or hyperdiploidy,^5,6  infant pro-B-lineage ALL (B ALL) with MLL-AF4 fusion,^7,8  acute myeloid leukemia (AML) with AML1-ETO⁹ as well as to Down syndrome patients with acute megakaryocytic leukemia (AMKL) and GATA1 mutations.¹⁰  In contrast, there are currently little data on the developmental timing of T-lineage acute lymphoblastic leukemia (T-ALL)/non-Hodgkin lymphoma (NHL)

NOTCH1 mutations occur in more than 50% of patients with T-ALL, in concert with other genetic changes.^11,12  From these data on clinical samples alone, one cannot deduce the temporal sequence of genetic events to decipher whether NOTCH1 mutation was initiating or secondary. The high prevalence of these mutations provided an opportunity to ask if NOTCH1 mutations occur prenatally. We have screened matched diagnostic, leukemic DNA from patients with T-ALL with their archived neonatal blood spots for concordant NOTCH1 mutations.

Eleven pediatric patients diagnosed with T-ALL were entered into the study, with ethical approval and with matched blood spots. Release of diagnostic samples (leukemia) and archived neonatal blood spots of patients were both approved by the standing Ethics Committee of the Institute of Child Health and Hospital for Sick Children (Great Ormond Street), and informed consent was obtained in accordance with the Declaration of Helsinki.

Genomic regions including the heterodimerization (HD-N and HD-C) and PEST domains of NOTCH1 were amplified by polymerase chain reaction (PCR) using leukemic DNA samples as previously described,¹¹  purified from agarose gel and directly sequenced. Mutations were subcloned into plasmid vectors and confirmed by sequencing.

A forward primer for the SIL gene and 4 reverse primers for different regions of the TAL1 gene were mixed and used in single PCRs as reported previously.¹³  Positive PCR products were directly sequenced to determine the type of SIL-TAL1 fusion.¹³ 

PCR primers amplifying NOTCH1 mutations in the leukemic samples were designed from the sequence data and optimal PCR conditions were established for each specific primer pair using leukemic DNA samples.

Small pieces (∼1mm² ) were cut out from each blood spot, mixed with PCR reagents containing Ampdirect Plus (Shimadzu, Kyoto, Japan). and NovaTaq (Novagen, Madison, WI) at 50 μL final volume and amplified.^4,9 

The NOTCH1 normal genomic sequences for HD-N (for patients MGT01 and MGT09) or PEST domain (for MGT11) were amplified as an internal control for blood spot DNA: HDN-f and HDN-r, 5′-GCTTGATGGGGTGCTTGCGCAG-3′ for amplification of HD-N region and PEST-f, 5′-ACCACCACAGCCGCACCTTG-3′ for PEST region.

The presence of NOTCH1 mutations was first assessed in the eleven leukemic DNA samples. Heterodimerization (HD) and PEST domains were screened only because they are the domains in which most mutations are found.^11,14,15  Four patients (MGT01, 09, 10, and 11) had mutations (confirmed by cloning and sequencing) in either HD (MGT01 and 09) or PEST domains (MGT10 and 11). Three involved insertions of unrelated short sequences and one had a duplication of a relatively long sequence from within NOTCH1 (Table 1). Only one patient sample (MGT01) of the eleven had a SIL-TAL1 fusion. This was a type 1 fusion⁶  (data not shown).

PCR primers were designed based on the sequences of patient-specific mutations and specificity and sensitivity were assayed with diagnostic DNA samples and cloning vectors with mutated sequences by serial dilutions. In one patient sample (MTG10), no primer set worked well, therefore samples from only 3 (MTG01, MTG09 and MTG11) with NOTCH1 mutations were further analyzed. With the patient-specific primer sets, positive PCR products were obtained only in corresponding patient DNA showing specificity of the primers (Figure 1A). Sensitivity of mutation-specific primers was assayed by serially diluting diagnostic DNA and cloned plasmid DNA. Variable and modest sensitivity was obtained. Down to 10⁻⁵ pg of mutated DNA could be detected with the primer set for MGT01 and 10⁻³ pg for MGT09 and MGT11 when plasmids with mutated NOTCH1 were used as a screen (data not shown). Using the diagnostic leukemic DNA as a screen for sensitivity, the threshold for detection was 1 ng for MGT01 (∼100 cells) and 10 ng for MGT11 (∼1000 cells). No specific mutation sequence could be detected in MGT09 up to 100 ng (Figure 1B), which may reflect the fact that the DNA was derived from a bone marrow sample that was substantially hemodiluted.

Excised blood spot segments were placed directly into PCR reactions with Ampdirect Plus to maximize the opportunity to amplify rare sequences⁴  and at least 16 segments of each spot analyzed. Normal NOTCH1 genomic sequences corresponding to the site of mutation on the region of HD-N (MGT01 and MGT09) and PEST (MGT11) were amplified as PCR controls and all segments (at least 8 segments in each patient) examined showed positive amplification (data not shown), indicating DNA on the Guthrie cards was intact.

Of the 3 patient samples studied, one, MGT01, showed unambiguous amplification of specific mutated sequence on several segments of blood spots (Figure 1C). In total, 20 pieces of the Guthrie card were subjected to PCR amplification and 11 were positive. Several of the positive bands were purified and sequenced and all had the patient-specific HD mutation (Table 1). A total of 20 pieces of control blood spots from the Guthrie cards of nonleukemic children were examined as additional controls with the MGT01 mutation-specific primer set. No amplification bands were observed. Patient MGT01 also had the SIL-TAL1 gene fusion. Sixteen fragments of blood spots for patient MGT01 were analyzed for the presence of the specific SIL-TAL1 fusion sequence but no PCR product was observed after 2 rounds of PCR amplification (70 cycles of PCR in total; Figure 1D).

Patient MGT01's blood spots, positive for NOTCH1 mutation, were the only ones for which we obtained sensitivity comparable with that achieved and required with leukemia fusion genes that register positive in blood spots – approximately 100 cells.^2,4,7 

Negative results in neonatal blood spot screening are uninterpretable but tend to underestimate the frequency of prenatally-initiated leukemias.²  This is particularly so when, as in the present study with NOTCH1 mutation, levels of sensitivity achievable with subtle sequence mutations, while maintaining specificity, are very modest compared with fusion genes.^4,7 

No general conclusion can be drawn about the usual timing of this common genetic abnormality in T-ALL, but these data indicate that NOTCH1 mutation can occur prenatally as an early or possibly initiating event in T-cell leukemogenesis. This interpretation accords with the view that NOTCH1 regulates self-renewal properties of stem cells or progenitors^16,17  and can initiate T-cell leukemogenesis as a transgene in mice¹⁸  or zebra fish.¹⁹  In animal models of T-ALL, secondary NOTCH1 mutations have been detected in leukemias initiated by other genetic changes.^20,21  As in this particular positive case, MGT01, the concurrent diagnostic SIL-TAL1 fusion sequence was absent from the neonatal blood spot (with a sensitivity threshold of ∼10 cells/mutation copies; Figure 1B), then it is likely that the latter was a secondary, postnatal event complementing the functional impact of prenatal NOTCH1 mutation.

There are only 2 other studies addressing the possible prenatal origins of T-ALL. The first was our previous report of a pair of monozygotic twins with T-NHL/ALL.²²  Malignant cells from the pair shared an identical TCRβ rearrangement (including an 11 bp N region) indicative of a single cell clonal origin in utero. Cells from these twins did not have a NOTCH1 mutation (A.M. Ford and M.G., unpublished observation, March 2007). Fasching et al²³  reported that 2 patients with T-ALL had detectable TCR gamma clonal rearrangements in their neonatal blood spots, identical to those in the leukemic cells at diagnosis. Collectively, these data argue that pediatric T-ALL can be initiated prenatally.

We thank Drs Ian Hann and Jane Chalker (Great Ormond Street Hospital) for access to clinical samples and Prof Kinuko Mitani for support (M.E. and M.E.-I.).

This work was supported by the Leukaemia Research Fund, United Kingdom.

Contribution: M.E.-I. and M.E. carried out all experimental work. H.K. provided clinical samples and diagnostic details plus Guthrie cards. M.G. conceived and planned the study and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Present address for M.E.-I. and M.E.: Department of Hematology, Dokkyo University School of Medicine, Tochigi, Japan.

Correspondence: Prof Mel Greaves, Section of Haemato-Oncology, Institute of Cancer Research, Brookes Lawley Building, 15 Cotswold Road, Belmont, Sutton, Surrey SM2 5NG, United Kingdom; e-mail: mel.greaves@icr.ac.uk.