Origins of iAMP21 in children who later developed pediatric acute lymphoblastic leukemia: an investigation of neonatal blood spots

Certain genetic subtypes of pediatric acute lymphoblastic leukemia (ALL) such as KMT2A rearrangements (KMT2A-r), ETV6::RUNX1 fusion and high hyperdiploidy, can originate in utero.^1-3 Recurrent chromosomal abnormalities correlate strongly with prognosis and therapeutic response and high-risk lesions such as KMT2A-r, BCR::ABL1, hypodiplody and intrachromosomal amplification of chromosome 21 (iAMP21) confer adverse outcomes and guide treatment intensification.^1,4-6 Interactions among inherited predisposition, early-life infectious exposures, and abnormal cytokine signaling have been proposed to contribute to the emergence and expansion of leukemic clones.^7-9 The developmental timing and cellular origins of iAMP21-ALL however remain incompletely defined.¹⁰ Clarifying whether these changes arise prenatally or postnatally is essential for understanding disease initiation and the subsequent clonal evolution of this subtype of pediatric ALL.

iAMP21 is observed in 2% of pediatric B-cell precursor (BCP) ALL cases. Clinically, iAMP21 is associated with older age at diagnosis (median 9 years) and low white blood cell counts.^10,11 Constitutional aberrations of chromosome 21 translocations, particularly the Robertsonian translocation, rob(15;21) (q10;q10)c, and ring chromosome 21, r(21)c, are associated with a markedly increased risk of developing iAMP21-ALL, often due to genomic instability and chromothripsis initiated at the dicentric Robertsonian or ring chromosome.^12,13 However, in a significant number of patients, the iAMP21 structure arises sporadically without a known underlying constitutional abnormality.¹⁰ Generation of the iAMP21 chromosome is an early clonal event,¹⁰ however the important question of how protracted the preclinical natural history can be in these older children remains unresolved.

In the present study, using patient-matched neonatal blood spots (NBS) collected at birth, we retrospectively investigated the presence of iAMP21 aberrations in two cases of pediatric ALL diagnosed at ages 8.3 and 14.4 years, respectively. At diagnosis, both cases presented with >60% blasts and thrombocytopenia but a Robertsonian translocation or isodicentric chromosome was not detected (Online Supplementary Table S1). Initial treatment was administered according to the NOPHO intermediate-risk protocol. Patient 1 experienced a relapse 4 years after diagnosis and was subsequently treated under the IntReALL 2010 protocol. Following a second remission, the patient underwent hematopoietic cell transplantation (HCT) and remains alive 10 years after the initial diagnosis. In contrast, patient 2 also underwent HCT but died due to post-transplant complications (Online Supplementary Table S1). This study was approved by the regional ethical review board in Stockholm (DNR2020-00774). Clinical and genetic data at diagnosis (Online Supplementary Table S1) were obtained from the Swedish Childhood Cancer Registry, a nationwide quality registry.

In the first patient (iAMP21 patient 1), using whole-genome sequencing (WGS) followed by integrative genomics viewer analysis (IGV), we randomly identified several complex chromosome 21 aberrations at diagnosis, including a small deletion at chr21:40585321-40585440 (-119 bp) and a complex translocation detected at high-depth and predicted between chr21:16859181-17010117. Using polymerase chain reaction (PCR) across the fusion regions, we were able to fully backtrack the small 119 bp deletion detected at diagnosis to each NBS punch of the individual (Figure 1A, B). Possible sequence microhomology at the breakpoints suggests involvement of microhomology-mediated end joining (MMEJ) in the repair process.¹⁴ The lack of an expected germline band for this region suggests either a germline homozygous deletion, patient polymorphism or a clonal event acquired prenatally. The breakpoint sequences of the diagnostic translocation (~141 kb apart) suggested an intergenic inversion or complex rearrangement, which was confirmed by PCR of the diagnostic DNA (Figure 1C, D). The variant allele fraction (~12-13%) further suggested either a subclonal population at diagnosis or a low-level mosaic germline event. Notably, only one of three NBS punches tested, produced a faint PCR product that had an identical DNA sequence to the diagnostic rearrangement (Figure 1E, F). This result is consistent with the existence of the rearrangement in only a subclonal population of cells prior to birth.

Following IGV analysis of diagnostic DNA from patient 2, we again randomly identified a chromosome 21 translocation (chr21:26001836-26002391) predicted to represent a 540 bp inverted segment. PCR amplification and sequencing confirmed this inversion at diagnosis (Figure 2A). All three NBS punches contained the same rearrangement, indicating that this region was broken, inverted and rejoined before birth. IGV data at diagnosis suggests that the translocation was present in all blood cells and could be a germline event (Online Supplementary Figure S1A). Strikingly, a 69 bp duplication event also observed in patient 2 at diagnosis (chr21:21:43292455-43292524) had likely progressed postnatally from a small deletion present in the NBS before birth and at the same location (Figure 2 B), confirming clonal evolution. We similarly backtracked a small deletion at chr21:41540011-41540460 and confirmed its presence at birth in each NBS (Online Supplementary Figure S1B). Sequence microhomology at the breakpoints of both these aberrations was again suggestive of MMEJ repair. Taken together, these results suggest that early genetic variants predispose to the acquisition of more dominant rearrangements and confirm that additional postnatal events and clonal evolution serve to drive overt iAMP21 leukemia (Online Supplementary Figure S2).

Figure 1.Complex translocations and deletions within chromosome 21 of intrachromosomal amplification of chromosome 21 patient 1 at diagnosis backtrack to birth and show clonal evolution. (A) A selected deletion at chr21:40585321-40585440 (-119bp) in intrachromosomal amplification of chromosome 21 (iAMP) patient 1 at diagnosis is present in the neonatal blood spot (NBS) at birth. Primers in grey blocks, prospective deleted sequences in yellow. (B) Identical fusion sequence to the deletion identified in patient 1 at diagnosis is observed in DNA extracted from the NBS punch. Boxed sequences show possible microhomology. (C) DNA sequence acquired from whole-genome sequencing (WGS) of diagnostic DNA shows a complex translocation between (chr21:16859181-17010117). Primers in grey blocks. (D) DNA sequence confirms the expected translocation after polymerase chain reaction of diagnostic DNA from the patient. (E) Interrogation of 3 separate NBS punches suggests the presence of the translocation to be only in a subclone of cells within a single NBS. All samples (except diagnosis [1.dx]) were isolated from NBS DNA. The MLL gene acts as a loading control. (F) Identical fusion sequence to the diagnostic translocation of patient 1 is observed as a subclone within a single NBS punch. con: control.

Using WGS data and the MiXCR tool for immune repertoire analysis,¹⁵ we next assessed the clonality of IGH rearrangements in these two patients. Both cases showed clonal rearrangements at diagnosis involving one or both components of the IGH (VDJ) or IGL (VL) repertoires. In patient 1, after WGA of the NBS followed by patient-specific quantitative PCR (Q-PCR), we identified a matching, patient-specific VDJ rearrangement sequence identical to that observed at diagnosis, including the non-templated nucleotide regions inserted between the VH and DH and DH and JH segments respectively (Figure 3A). Backtracking of this clonotypic IGH rearrangement to the NBS further indicates the presence of a pre-leukemic clone at birth. In contrast, for patient 2 we did not detect the diagnostic patient-specific clonal VDJ rearrangement in any of the three NBS punches analyzed (Figure 3B). Taken together, these findings demonstrate that selected iAMP21-associated chromosomal aberrations and clonal IGH rearrangement can be traced back to NBS, providing evidence of a prenatal origin for at least some pre-leukemic clones.

Backtracking of both iAMP21 patients, confirmed that some chromosome aberrations were detectable in both the diagnostic sample and the NBS, indicating either a germline origin or a somatic origin in utero, where blood cells are rapidly proliferating. These microdeletions and simple inversions on chromosome 21 can be germline only if they are small, simple, and do not involve essential 21q22 genes. We note that the small deletions and insertions we have analyzed here, do occur in introns. Larger deletions, inverted duplications, and complex 21q rearrangements (such as iAMP21) are incompatible with germline events and must be somatic aberrations and, given the extensive latency to overt leukemia, they are likely to be postnatal. In patient 1, the complex translocation identified at diagnosis was not present in all NBS punches examined, perhaps indicative of a secondary ‘hit’ rather than a germline event. An identical, patient-specific VDJ rearrangement to that observed at diagnosis was also detected in the NBS of this patient, providing more evidence for the presence of a pre-leukemic clone at birth. In the second patient a small deletion detected in the NBS had progressed to a duplication event by diagnosis of iAMP21-ALL. These findings indicate that iAMP21 can arise through multiple developmental routes: either as a predisposing early embryonic (potentially germline) event, or as a clonally restricted prenatal lesion that later expands to clinical leukemia.

Figure 2.Inversions and deletions within intrachromosomal amplification of chromosome 21 patient 2 at diagnosis backtrack to birth and show clonal evolution. (A) The break and inversion of a translocation within chromosome 21 identified at diagnosis for patient 2 (Pt2) (chr21:26001836–26002391) is already present in the neonatal blood spot (NBS) at birth. Boxed sequences show possible microhomology. (B) The duplication event present in Pt2 diagnosis (chr21:21:43292455-43292524) likely arises postnatally and subsequent to a small deletion already present in the NBS.

A prenatal mutation in an early hematopoietic stem cell alone is insufficient to induce leukemia. iAMP21 development is instead driven by chromosomal instability,¹⁰ often leading to amplification of the RUNX1 region on chromosome 21, disruption of normal B-cell maturation, and emergence of pre-B-cell leukemic features. Accumulation of secondary mutations confers a proliferative advantage, resulting in aggressive clonal expansion and eventual dominance of the leukemic clone in the bone marrow.¹⁰

The dual patterns observed for iAMP21 raise the possibility that this lesion does not fit neatly into a single developmental category. Instead, iAMP21 may represent a biologically heterogeneous subtype, with prenatal initiation occurring in some patients and postnatal acquisition in others, but with clonal evolution occurring in both scenarios. The positive backtracking results for these older children with iAMP21 ALL, show that a very protracted covert natural history of iAMP21-ALL can occur prior to disease appearance and provides evidence that the postnatal latency in this subtype of ALL can be a decade or longer.

Figure 3.IGH clonality by VDJ rearrangement analysis and patient-specific quantitative polymerase chain reaction. (A) Presence of a IGH VDJ pre-leukemic clone in the neonatal blood spot (NBS) of patient 1 with DNA sequence identical to that identified at diagnosis (NBS3, including random base insertion, N2). (B) No clonal IGH sequences were found in the NBS of patient 2 at birth. dx: diagnosis; RFU: relative fluorescence units.

This study is limited by the small sample size, reflecting the rarity of both iAMP21 and the availability of suitable neonatal samples. While the sensitivity of our assays was high, extremely small pre-leukemic clones may remain below the limit of detection. Larger population-based studies, ideally incorporating single-cell sequencing or ultra-deep targeted assays,¹⁰ family studies and sampling from different tissues, such as skin or saliva will all be essential to determine the full spectrum of developmental origins for these lesions. Nevertheless, our data provide new and direct molecular evidence that contributes to the longstanding question of when specific pediatric leukemia begins. By integrating neonatal sampling with modern genomic tools, we demonstrate that iAMP21 can initiate prenatally as a germline or early clonal event, so highlighting distinct developmental trajectories among ALL subtypes.

Footnotes

• Received February 28, 2026
• Accepted April 7, 2026

Correspondence

*GL and VL contributed equally as first authors.

^#AMF and BG contributed equally as senior authors

G. LEIJONHUFVUD - gustaf.ab.leijonhufvud@uu.se

Disclosures

No conflicts of interest to disclose.

Contributions

Funding

The authors gratefully acknowledge financial support from the ALF agreement (Avtal om Läkarutbildning och Forskning) at Uppsala University, The Dickens Family Leukaemia Fund and The Institute of Cancer Research, London. We also thank the Mary Béves Foundation, the Anna Lisa and Arne Gustafsson Foundation, and the Swedish Childhood Cancer Fund for their generous support, which made this study possible.

Acknowledgments

The authors thank Clinical Genomics Uppsala, Science for Life Laboratory, Uppsala University, for their expertise in sequencing and analysis. Clinical and genetic data at diagnosis were kindly provided by the Swedish Childhood Cancer Registry. We are grateful to the Swedish National Board of Health and Welfare (Socialstyrelsen) for access to the maternal health register, and to the Center for Medical Mass Screening (CMMS) for facilitating access to neonatal blood spot (PKU) samples. Finally, we thank Professor Christine Harrison for her very helpful discussions.