CDKN2 deletions have no prognostic value in childhood precursor-B acute lymphoblastic leukaemia Leukemia (2005) 19, 1281–1284. doi:10.1038/sj.leu.2403769 Published online 21 April 2005 TO THE EDITOR In acute leukaemia, p15/p16 (CDKN2) deletions have been found with high frequencies in both adult and childhood acute lymphoblastic leukaemia (ALL) (430%). In acute myeloid leukaemia (AML), chronic myeloid leukaemia (CML) and myelodysplastic syndromes (MDS), very low frequencies of p15/p16 deletions have been observed (0–1%). However, the rates of p15 promoter hypermethylation, another frequent mechanism of p15/p16 inactivation, are very high in AML (70%), MDS (40%) and also precursor B-ALL and T-ALL (40%). 1 Therefore, inactivation of this locus is thought to play an important role in leukaemogenesis and could have important clinical significance. The publications on the prognostic value of CDKN2 inactivation in acute leukaemia show contradictory results. 1 In most of these studies, however, no difference was made between childhood ALL cases with B- and T-cell phenotype. The reported poor prognostic value of CDKN2 deletions in some of these studies may result from differences in patient groups, as T-ALL in general has a worse outcome and a high incidence of CDKN2 deletions. 1 As the clinical relevance of CDKN2 deletions is still a subject of debate, we investigated, in a single centre study, the prognostic significance of CDKN2 deletions in 109 diagnostic childhood common/precursor B (c/pre-B)-ALL cases from the Erasmus MC, Sophia Children’s Hospital using FISH analysis. A summary of the patients’ characteristics is given in Table 1. At diagnosis, informed consent of the patients and/or parents/guardians was obtained to use left-over material for research purposes. Loss of CDKN2 was analysed using P1.1063 (53 kb), which spans most of the locus, covering the region from upstream exon 2 of p15 to distal to exon 2 of p16, including the coding sequence of p14. 2 If a deletion was observed, it was further characterized using a combination of P1.1062 and cos5. These probes cover both p15 and p16, but differ from P1.1063 in that P1.1062 (51 kb) starts upstream of p15 and ends immediately downstream of p16. Cos5 encompasses p15 and exons 1–2 of p16 and p14, and because of its smaller size (23 kb), smaller deletions may be detected. 2 All cases had also been analysed for the presence of cytogenetic aberrations. Cytogenetic abnormalities were observed in 96/109 (88%) c/ pre-B-ALL cases. The remainder had apparently normal kar- yotypes, except for one case with karyotype failure due to lack of metaphases. The cytogenetic findings are summarized in Table 2. The known poor prognostic aberrations, t(9;22) and 11q23 rearrangements, were grouped together (Table 1), be- cause the numbers were too few to use the individual translocations as a factor in survival analysis. Loss of CDKN2 (both p15 and p16) was observed in 37/109 (34%) c/pre-B-ALL cases using P1.1063 and the combination of P1.1062 and cos5, whereas the other 72/109 (66%) were normal for this locus. Of these cases with CDKN2 loss, 7/37 (19%) had hemizygous deletions (loss of the CDKN2 FISH signal of one chromosome 9, with the signal on the other chromosome 9 being retained), whereas homozygous deletions (loss of CDKN2 FISH signals of both chromosomes 9) were observed in 15/37 (40%) cases. Additionally, we observed the presence of both hemi- and homozygous deletions in 15/37 (40%) cases (Table 2). In seven cases, both types of CDKN2 deletions were observed in 20–25% of the nuclei each. In three cases, the percentage of cells with a homozygous deletion was much higher than those with a hemizygous deletion (58–61 vs 18–20%, respectively), whereas in three other cases, hemizygous deletions were observed in higher percentages than homozygous deletions (50–74 vs 17–21%, respectively). In two cases, both hemi- and homozygous deletions were observed in 40–50% of the nuclei. The group with mixed deletions shows that different clones occur next to each other. This also demonstrates the heterogeneity in leukaemic samples. The percentage of deletions (34%) observed in our study is higher than that reviewed in Drexler 1 (23% for both p15 and p16), but comparable to other studies in which CDKN2 deletions have been observed in approximately 30% of B-lineage ALLs. 3,4 The FISH results were compared to the corresponding karyotypes (Table 3). In total, only 18 cases had cytogenetically visible aberrations of the short arm of chromosome 9 (Table 4). Five of these cases had breakpoints in 9p21 (cases 1–5), whereas the remaining 13 cases (cases 6–18) showed breakpoints in other bands of 9p. Only 2/5 cases with 9p21 breakpoints showed CDKN2 deletions with FISH (cases 4 and 5), resulting either from deletions or unbalanced translocations as observed in the karyotype. The other 3/5 cases, having cytogenetically visible deletions or a translocation involving 9p21, did not have CDKN2 deletions (cases 1–3). In addition, 12/13 cases with abnormalities of the short arm of chromosome 9 other than 9p21 showed CDKN2 deletions (cases 7–18). Among the cytogeneti- cally visible aberrations were dic(9;20)(p11–13;q11) in four cases (cases 7, 12, 15 and 16), a dic(9;10)(p13;p1?1) (case 8), unbalanced translocations with breakpoint in 9p13 or 9p2? in two cases (cases 13 and 18), additions to 9p1?2, p2?2 or p1 in three cases (cases 9–11), a del(9)(p?) in case 14 and a 9?p in case 17. In 6/12 cases, homozygous CDKN2 deletions were observed. Deletion of the second allele has occurred, which was not expected based on the karyotypic information. The remaining case in this category with a dic(9;10)(p?13;q?21) had no CDKN2 deletion (case 6). The other 23/37 cases with CDKN2 deletions showed no aberrations of 9p, indicating that CDKN2 deletions mostly are cryptic abnormalities. Thus, CDKN2 deletions occurred in patients with abnormal and normal karyotypes with and without cytogenetic 9p abnormalities. Received 14 December 2004; accepted 15 March 2005; Published online 21 April 2005 Correspondence: Dr H Berna Beverloo, Department of Clinical Genetics, Dr Molewaterplein 50, 3015 GE, Rotterdam, The Nether- lands; Fax: þ 31 10 40 89492; E-mail: b.beverloo@erasmusmc.nl 4 Current address: Centre for Human Genetics, Catholic University of Leuven, Leuven, Belgium Correspondence 1281 Leukemia