Subclonality and prenatal origin of RAS mutations in KMT2A (MLL)‐rearranged infant acute lymphoblastic leukaemia

This is the authors' abstract. We don't add key points for this paper.

KMT2A (previously termed MLL) rearrangements (KMT2A-r) constitute the most frequent aberration in infant lymphoblastic leukaemia (iALL). Mutations in the RAS genes have previously been recognized as an important somatic abnormality in KMT2A-r-iALL (Taketani et al, 2004; Liang et al, 2006; Driessen et al, 2013; Prelle et al, 2013). Certainly, KMT2A-r in iALL are present at birth (Greaves et al, 2003). On the other hand, RAS mutations (RASmut) are suggested to be secondary and subclonal hits in the leukaemogenic pathway. In this context, our study aimed to (i) quantitatively evaluate the association between RASmut at diagnosis and the main clinical features of KMT2A-r-iALL, and (ii) analyse the mutations in matched samples [neonatal blood spots, also known as Guthrie cards (Gc), remission and/or relapse]. KRAS/NRASmut were screened by direct sequencing, and mutated cases were also subjected to pyrosequencing in order to quantify the percentage of blasts harbouring the mutation. The detailed methods performed in this study are described in the Supporting Information (Data S1). Ninety-two KMT2A-r-iALL cases were included. RASmut were found in 21·7% of cases [KRAS (14·1%) and NRAS (7·6%)]. In childhood ALL, RASmut have been diagnosed in 6–21% of patients (Liang et al, 2006; Wiemels et al, 2010). Recent studies in large cohorts of KMT2A-r cases have reported frequencies of 14–15% RASmut, with an even higher frequency in t(4;11) infants (c. 25%) (Driessen et al, 2013; Prelle et al, 2013). These previous data are therefore concordant with our study. The clinical-laboratory features of KMT2A-r-iALL according to RAS status are shown in Table SI. Sixty percent of RASmut patients were ≤6 months old, 60% harboured AFF1 as the KMT2A-translocation partner gene (TPG), and 80% presented both the KMT2A-AFF1 and AFF1-KMT2A transcripts at mRNA level. Although these differences were not statistically significant, these analyses are in agreement with previous data, such as a higher frequency of RASmut among younger and t(4;11) infants (Driessen et al, 2013; Prelle et al, 2013). Furthermore, KRAS (P = 0·033) but not NRAS (P = 0·415) mutations were associated with high white blood cell (WBC) counts. The quantitative association between RASmut at diagnosis and clinical features was analysed by pyrosequencing the mutated clones. The median percentage of mutated clones did not differ according to age group (31% vs. 27%; Fig S1A; P = 0·43) or KMT2A-TPG (28% vs. 33%; Fig S1C; P = 0·89). Significantly higher percentages of mutated clones were found in cases with WBC counts >300 x 109/l (Fig S1B; P = 0·01). The correlation of RAS-mutated clones with higher WBC count may indicate that RASmut play a role in aggressiveness. Accordingly, the presence of RASmut was an independent prognostic factor in a previous cohort (Driessen et al, 2013). Despite recent attempts, neither the role of a second-hit nor the subclonality of RASmut in KMT2A-riALL has been demonstrated (Driessen et al, 2013; Prelle et al, 2013). Nevertheless, according to pyrosequencing analyses, we checked for a putative subclonality of RASmut and found that, in 19/20 cases, the percentage of mutated alleles at diagnosis was lower than expected for a clonal alteration. The RASmut were monoallelic (50% in a single cell). Taking the percentage of blast cells into account, if the alteration is clonal, the percentage of mutated alleles is expected to be half of the blast count for each patient (Table SII). Then, subclonality was observed in this present study. Regarding the analysis of matched samples, 17 samples of patients who achieved remission were analysed, but no mutations were found. Among the relapse samples (n = 12), five patients lost RASmut at relapse and seven patients were wild-type at both diagnosis and relapse. Concerning the two Gc samples, in both cases the RAS-mutated allele ‒ present at diagnosis ‒ was also detected at birth (Table SII). One of these infants (Patient 13) was diagnosed with ALL 17 d after birth, with high WBC, KMT2A-MLLT10 gene fusion (Fig 1A) and KRASmut (Fig 1B) detected at birth. The intra-patient's quantitative comparison between samples showed that the Gc sample had a higher percentage of KRAS-mutated clones than the diagnostic sample (42% vs. 33%, Fig 1C, D). Finding the KRASmut in this infant's Gc sample was an expected result as the baby was presumably born with the disease and presented with 97% blasts in the peripheral blood. The other infant (Patient 15) had an NRASmut (Fig 2A) at 7-months old, with 96% peripheral blood blasts and KMT2A-AFF1. The percentage of NRAS mutated clones was 50% in both the Gc and the diagnostic sample (Fig 2C, D). This high percentage of mutated cells within the Gc sample may explain the backtracking success. The presence of RASmut at birth clearly argues against a secondary role in a two-hit model for this subgroup of leukaemias. Although former studies performed in hyperdiploid childhood leukaemia strongly indicate that RASmut are secondary hits (Case et al, 2008; Davidsson et al, 2010), this might not be a general rule for childhood ALL. Recently, an increased risk of RASmut was associated with maternal tobacco smoking during pregnancy [odds ratio (OR) 3·06, 95% confidence interval (CI), 1·03–9·07], and this risk was modulated by the NQO1 609CT polymorphism (OR, 4·24, 95% CI, 1·24–14·50) (Andrade et al, 2014). Additionally, the quantification of matched Gc/diagnosis samples that showed a higher percentage of mutant alleles in the Gc sample compared to the diagnostic sample, suggests that RAS-mutated subclones are very important for accelerated disease onset, but may not be crucial for disease perpetuation. This hypothesis is compatible with a transgenic mouse model in which a KRASmut promoted KMT2A-AFF1 disease onset with a short latency (Tamai et al, 2011), and also with our finding that five patients lost their RAS-mutated clones at relapse. In conclusion, RASmut are subclonal and can also arise in the prenatal period in KMT2A-r-iALL. However, RASmut do not seem to be driver mutations, rather a passenger mutation that may aid disease onset by accelerating the initial expansion of cells. RAS-mutated cells will be certainly negatively selected by current chemotherapy regimens, due to the growth promoting activity of active RAS. On the other hand, relapsed leukaemia cells infrequently display RASmut, indicating that it is not necessary for disease maintenance. We are grateful to the children and their parents for participating in the study. We thank Alessandra de Jesus Faro, Camilla Fernandes Gomes Andrade, Caroline Barbieri Blunck, and all physicians of the Brazilian Collaborative Study Group of Infant Acute Leukaemia who contributed with laboratory diagnosis, clinical and demographical data. This investigation was supported by the Brazilian National Research Council (CNPq) and Instituto Nacional de Câncer (INCA). The project was partially funded by grants from INCT-Controle do Cancer (CNPq #573806/2008-0, FAPERJ#E026/170.026/2008 and FAPERJ#E-26/110.509/2010; FAPERJ#E-26/110.823/2012). MSPO has been supported by CNPq research scholarships (#302423/2010-9) and FAPERJ (#E026/101.562/2010). ME has been supported by Brazilian Ministry of Health through the Institutional Development Program Scholarship. ME and MSPO designed the research study. ME wrote the manuscript. ME, TCB and BAL performed the research and analysed the data. TCB, BAL and CM contributed to drafting of the manuscript. CM analysed the KMT2A breakpoint. MSPO and RM revised the final version critically. The authors have no competing interests. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.