The Ups and Downs of Leukemia in Children With Trisomy 21

Patients with constitutional trisomy 21 (Down syndrome [DS]) are at high risk of developing both acute lymphoblastic leukemia and myeloid leukemia (DS-ALL and ML-DS, respectively). These malignant conditions are associated with distinctive epidemiology, genetic signatures, and clinical outcomes compared to those of non-DS leukemias.^1-3  DS-ALL is nearly exclusively of the B-cell lineage and occurs in children and adolescents/young adults, with extremely few cases reported in infants younger than one year. Many DS-ALL cases are cytogenetically normal, suggesting a role for activating mutations as leukemic drivers.³  Almost all ML-DS cases are of the acute megakaryoblastic subtype and occur in children younger than four years. ML-DS commonly ensues following the canonical preleukemic transient myeloproliferative disorder (TMD) characterized by somatic GATA1 mutations and is hypothesized to arise from a residual TMD population that subsequently acquires additional driver mutation(s).^{1, 4, 5}  Patients with DS-ALL have inferior clinical outcomes versus those with non-DS-ALL, which is attributed to higher rates of treatment-associated mortality (TRM) and relapse.^{6, 7}  Conversely, children with ML-DS have excellent chemosensitivity and high overall survival (OS), although survival of those with relapsed ML-DS is dismal.⁵ 

Attempts to reduce the treatment toxicities for pediatric patients with DS-associated leukemias have shown mixed success to date. TRM is typically due to infection and can occur during any phase of treatment, including maintenance.⁶  In addition to mortality resulting from the toxicity itself, individual treatment modifications in response to toxicities seem to be an important cause of lower event-free survival (EFS) in children with DS-ALL.⁷  Prior strategies in the Children’s Oncology Group (COG) AALL0932 and AALL1131 clinical trials to reduce anthracycline exposure, modify methotrexate dosing, decrease vincristine/glucocorticoid pulses, and intensify supportive care for patients with DS-ALL did not appreciably decrease TRM.⁸  The current COG AALL1731 trial further reduces chemotherapy toxicity by non-randomly assigning children with DS-ALL to three cycles of blinatumomab, a bispecific CD19×CD3 antibody immunotherapy, and omitting daunomycin in induction (all DS-ALL) and the cyclophosphamide/cytarabine-based second month of delayed intensification (high-risk DS-ALL only). Together, these two phases of treatment represent almost all of the TRM in high-risk DS-ALL.⁹  Preliminary analysis of interim AALL1731 data showed similar TRM rates (approximately 8%) among patients with DS-ALL on AALL1131 and AALL1731 however, and an increased blinatumomab-associated seizure risk that can be mitigated with antiepileptic drug prophylaxis.¹⁰ 

The reasons for the significantly higher rate of relapse in DS-ALL are likely multifactorial, including underlying leukemia biology, decreased immune surveillance compared to non-DS patients, and under-reported treatment reductions in response to adverse events.^{3, 6}  A specific association between the distinct genetics of DS-ALL and poorer outcomes has not been established. For example, CRLF2 rearrangements with frequent JAK2 co-mutations that are associated with poor outcomes in patients with non-DS-ALL¹¹  occur in 50 to 60 percent of DS-ALL and seem to have limited prognostic significance.^{6, 12, 13}  Conversely, favorable genetic alterations in non-DS-ALL, such as ETV6-RUNX1 fusions, are also associated with good prognosis in DS-ALL.^{6, 7} 

Given the superior EFS of ML-DS compared to non-ML-DS, a major focus of recent clinical trials has been to reduce treatment intensity and toxicity.^{5, 14, 15}  The COG AAML0431 trial aimed to improve EFS and reduce TRM by administering high-dose cytarabine (HiDAC) during induction II instead of intensification, decreasing total anthracycline exposure, and reducing prophylactic intrathecal chemotherapy.¹⁴  While the adjustments to anthracyclines and intrathecal chemotherapy did not affect EFS or OS, the earlier HiDAC exposure improved both metrics, though adverse events occurred more frequently during induction II.^{5, 14}  The COG AAML1531 trial subsequently had the goal of omitting HiDAC in standard-risk patients with ML-DS who achieved measurable residual disease (MRD) less than 0.05 percent at the end of induction I.⁵  Unfortunately, children who did not receive HiDAC-based induction II had significantly higher relapse rates and worse EFS compared to MRD-negative patients treated on AAML0431, demonstrating that HiDAC inclusion is an essential component of ML-DS therapy.⁵  Interestingly, a Japanese Childhood AML Cooperative Study Group trial investigating a similar cytarabine dose-reduction reported much lower TRM and no significant differences in three-year EFS and OS,¹⁶  concordant with earlier studies.^{15, 17}  Differences in anthracycline pharmacodynamics and other small differences among treatment protocols have been hypothesized to account for these discrepant results,¹⁸  but these effects remain difficult to isolate, and HiDAC remains essential in ML-DS therapy in the United States.

The small numbers of children with relapsed/refractory ML-DS have limited the ability to identify prognostic molecular characteristics and other biomarkers. In addition to MRD positivity, older age at diagnosis (≥4 years) has been suggested to portend an inferior prognosis,^{17, 19}  though such ML-DS cases are usually more genetically similar to non-DS-associated AML.²⁰  The genetic characteristics of ML-DS blasts are clearly distinct from non-ML-DS cells,⁵  which is consistent with the unique cellular origins of ML-DS, but the exact significance of the pathogenic GATA1 mutations remains unclear. In a phenomenon termed “silent TMD,” as many as 30 percent of neonates with DS have been reported to have GATA1 mutations without detectable TMD in peripheral blood.^{17, 21, 22}  Interestingly, no association has been identified between specific GATA1 mutations and risk of ultimately developing ML-DS.²³ 

Patients with relapsed DS-ALL^{24, 25}  or ML-DS^{5, 19}  have poor prognoses, with successful salvage (including allogeneic hematopoietic stem cell transplantation) often limited by TRM and subsequent relapses.^{24, 25}  Exclusion of children with DS-associated leukemias from many early-phase clinical trials has hampered both access to innovative new therapies and tailored protocols focused upon DS-specific toxicity reduction.²⁶  Emerging data encouragingly suggest that patients with relapsed/refractory DS-ALL tolerate CD19-directed cellular therapy well, which can perhaps be used as definitive therapy without hematopoietic stem cell transplantation in some patients.²⁷  Similar advances are still needed for children with relapsed/refractory ML-DS.