Mutational profiling of acute lymphoblastic leukemia with testicular relapse

To the editor

Relapsed acute lymphoblastic leukemia (ALL) is the leading cause of deaths of childhood cancer [1‐3]. Although relapse usually occurs in the bone marrow (medullary), extramedullary relapse occasionally occurs, including either in the central nervous system or testis (<1–2%). Involvement of these organs is often associated with an inferior prognosis, perhaps because the blood-brain/blood-testis barrier hinders efficient delivery of chemotherapy, and/or the leukemic cells infiltrated in these immune-privileged sites may escape efficient immune surveillance. Currently, clonal origin and evolution of extramedullary relapse ALL remain poorly understood. To address this, we selected two pediatric ALL patients who experienced testicular relapse and interrogated their leukemic cells with exome sequencing (see Additional file 2: Supplementary Methods).

Patient D483 (5.6 years old at diagnosis) was treated as an intermediate-risk B-ALL [hyperdiploid:56,XY,+X,t(2;14)(p?13;q32),+4,+8,+9,+10,+14,+17,+18,+21,+21, absence of any well-known leukemic fusion oncogene]. He developed bone-marrow (96% blast) and testicular relapse 5 years after induction of remission. Mutations of KRAS (G12D) and CREBBP (S1436C, in histone-acetyltransferase HAT domain) were found in the founding leukemic clone at diagnosis and persisted in bone marrow and testis at relapse (Additional file 1: Table S1). CREBBP encodes a histone/non-histone acetyltransferases which is involved in regulation of glucocorticoid gene expression, its mutation contributes to prednisolone/dexamethasone (glucocorticoid) resistance [4, 5]. Mutation and copy-number deletion of CREBBP are frequent in B-cell lymphoma and ALL [3, 6] and are often associated with disease relapse. A missense mutation (R17Q) of MEF2B (MADS-box transcription enhancer factor-2) was found in both the bone marrow and testicular relapse samples. Missense mutation of MEF2B is frequently detected in diffuse large B-cell lymphoma and contributes to malignant transformation by regulating expression of the proto-oncogene BCL6 [7]. Other mutations included EVX1 (homeobox protein) and OTUD5 (regulates p53 stability by deubiquitinating p53) (Additional file 1: Table S1).

Second patient (case D727, 1.3 years old at diagnosis of B-ALL) harbored a MLL-AF9 fusion gene [t(9;11)] and was treated as a high-risk ALL (82.8% blast in peripheral blood). MLL fusion is often associated with infant-ALL and a poor prognosis. Complete remission was achieved after induction therapy; however, the patient relapsed (91% blast) after a 2.3-year remission. NT5C2 gene (encodes a 5′-nucleotidase involved in purine metabolism) had two mutations in the relapse samples, differencing in their VAF in the bone marrow (34%) and testicle (5%) for R367Q mutation; while D407V mutation was present with a VAF of 7% in bone marrow and 36% in testicular relapse. These two NT5C2 mutations occur as recurrent mutational hotspots in relapse-ALL and they have been functionally validated [8]. These mutations increase the NT5C2 inosine-5-monophosphate-nucleotidase activity; and therefore lead to resistance to one of the chemotherapeutic drugs, 6-mercaptopurine [8, 9] (part of child’s treatment). Additional mutations that occurred in this child’s ALL cells included DUSP13 (phosphatase that regulates JNK/P38 phosphorylation), MAPK8 (JNK1), PPP1R3B (protein phosphatase 1 regulatory subunit 3B), and ALPK3 (alpha-kinase 3) (Additional file 1: Table S1).

To gain insight into the evolutionary trajectories of these two ALL cases, we analyzed mutational clustering of VAF and clonal evolution based on their sequencing data (Additional file 2: Supplementary Methods). Mutations shared at leukemic diagnosis and relapse represent early mutations and constitute the founding clone, while mutations occurring only at diagnosis in the marrow or only relapse samples of testicle/bone marrow likely were acquired later. For patient D483, the relapse leukemia directly evolved from the original leukemic clone at diagnosis; all mutations at diagnosis were persistent (Fig. 1a), and four additional missense-mutations [MEF2B (R17Q), KCNG1 (L252V), AIM1 (G109R), and OTUD5 (G222D)] were acquired with different VAF at relapse in both bone marrow and testis, suggesting that both sites of relapse evolved from the same leukemic clone at diagnosis (Fig. 1b, c). In contrast, patient D727 had a proportion of mutations present at diagnosis which were absent at relapse, suggesting that the relapsed leukemia arose from an ancestral clone which existed before the overt leukemia at diagnosis. Analysis of mutational pattern and VAF suggests that relapse in patient’s testicle represents an independent subclone from the relapse in the bone marrow, albeit they share a common progenitor clone derived from the original ancestral clone (Fig. 2a–c). Of note, a fraction of mutations present at diagnosis persisted in the testicle but were absent in relapsed marrow, suggesting that the relapse ALL evolved following a parallel branching hierarchy instead of a linear acquisition path.