Evolution of Sub-Clones with KRAS Mutations In Pediatric Patients with MLL-AF4 Rearrangements

Abstract 2454

Mixed Lineage Leukemia (MLL) gene rearrangements are common genetic translocations occurring in human leukemia. Numerous MLL-partner genes have been identified so far; of these MLL-AF4 was found to characterize the largest subgroup of fusion proteins in pediatric acute lymphoblastic leukemia (ALL). A recent study reported that mutations in kras accelerate leukemo-lymphogenesis in a MLL-AF4 transgenic mouse model. Activating mutations in RAS genes, frequently occuring in codons 12, 13 and 61, prevent the hydrolysis of RAS-GTP and result in the constitutive activation of the RAS proteins enhancing cell survival and proliferation. Mutations in RAS genes were found in several types of human cancers, including leukemia and ALL with MLL rearrangements.

We analyzed the presence of KRAS mutations in pediatric MLL-AF4 rearranged ALL samples at diagnosis and relapse with the aim to establish the frequency of KRAS mutations in MLL-AF4 leukemia and to tentatively associate RAS mutations to the leukemogenic process.

In this study 40 ALL patients were included (23 infant and 17 non infant); of these, 14 pairs at diagnosis and relapse, 22 at diagnosis only and 4 samples at relapse only. The study was approved by the institutional ethical committee and informed consent was obtained in accordance with the Declaration of Helsinki.

Mutation screening was performed for exon 2 and exon 3 of KRAS by 454 technology (Roche Applied Science) with GS Junior Sequencing Instruments; to recognize each specific patient, primer pairs included a 10-base molecular identifier barcode sequence (MID). We generated amplicon ultra-deep sequencing of the 2 amplicons with a median coverage per amplicon of 1800 reads per sample.

Twenty patients (55,6%) out of 36 with t(4;11) at diagnosis harbored KRAS mutations (range of penetrance of 0,4–29,7% for mutated sequences). Interestingly, in some patients more than one clone with different mutations at difference penetrance in KRAS codons were present. Of the 18 relapse samples 4 (22%) harbor KRAS mutations (penetrance of 1.3–38% of reads).

Sanger sequencing was used to validate the presence of mutations in patients with more of 20% of penetrance identified with 454 technology.

To evaluate the role of KRAS mutations in the leukemogenic process we analyzed matched paired diagnostic and relapse samples; in patients with mutations at diagnosis, 3 distinct situations were identified in the relapsed samples, pointing different changes between diagnosis and relapse: (1) presence of KRAS mutations in a specific codon at diagnosis and disappearance of this clone at relapse (6 of 9 cases, 67%); (2) decrease of the penetrance of a mutation between diagnosis and relapse and appearance of a new different mutation at relapse (2 of 9 cases, 22%); (3) clonal evolution of the same mutation and increase of the percentage of the mutated sequence from diagnosis to relapse (1 of 9 cases, 11%).

Apparently, patients with MLL-AF4 rearrangements at diagnosis and relapse presented with subclones and during the relapse phase previous clones disappeared while new clones appeared and only in one case we can assume a selective advantage of the clone with the KRAS mutation.

To further explore the propagation of KRAS mutated subclones in MLL-AF4, we transplanted 2 patients at diagnosis into a NOD/SCID xenograft mouse model. Xenografted samples obtained from leukemia bearing mice transplanted further onto subsequent recipients in serial passages. KRAS mutations were analyzed for each of the passages as mentioned above. One of the diagnostic samples used for transplantation harbored KRAS G12S mutation. The human cells of this patient isolated after the first passage in the mouse carried the same mutation of the diagnosis but with a very low penetration, suggesting a minority clone that decreased further in the second passage. No trace of mutations was found in the third passage. The other diagnostic sample for which serial transplanted samples were analyzed had no KRAS mutation at diagnosis and no new mutations appeared during further passages in the mice.

The disappearance of the mutated KRAS subclone during passages in the recipient mice seems to reflect the negative selection for mutated clones observed in patients between diagnosis and relapse.

Overall, these data showed that KRAS mutations are frequently occurring in MLL-AF4 leukemias at diagnosis however these mutations do not seem to add a selective advantage to the blast cells.