Genomic Alterations of the p19^ARF Encoding Exons in T-Cell Acute Lymphoblastic Leukemia

WE SHOWED 3 YEARS ago that deletions of the MTS1/p16^INK4a andMTS2/p15^INK4b genes occur frequently in T-cell acute lymphoblastic leukemias (T-ALL) cells.1 This has been confirmed by a number of studies,2-5 and it is now established that chromosomal rearrangements deleting or disrupting theMTS locus represent a genetic event characteristic of T-ALL.6 In the majority of cases, the whole MTSlocus is deleted on both chromosomes. In about a third of alleles, the breakpoints fall within the locus and most of them are located in two major clusters (MTS1^bcrα andMTS1^bcrβ).5,7 We have recently cloned and sequenced several breakpoints falling in either cluster and have shown that these recombinations are due to V(D)J recombinase activity, which, active in thymocytes, is involved in the generation of antigen T-cell receptor diversity.7 Although the incidence and the mechanism of these rearrangements are now well-defined in T-ALL, their functional targets remain to be identified.

The MTS locus8 extends over 50-kb at the 9p21 chromosomal band and includes two genesMTS1/CDKN2/p16^INK4a andMTS2/p15^INK4b. MTS1 acts as a tumor suppressor.9 The gene has a complex structure, as it encodes two unrelated proteins, p16^INK4a10 and p19^ARF11 from two types of RNAs,11-14,MTS1α and MTS1β, respectively. These RNAs are transcribed from distinct promotors located 5′ to alternative first exons (MTS1 exon 1α and exon 1β, respectively) and share the MTS1 exons 2 and 3. The MTS1 exon 2 is read in two distinct frames to code for p16^INK4a and p19^ARF. MTS2 includes two exons and encodes p15^INK4b.15 p16^INK4a and p15^INK4b are two of the INK4 family members demonstrating cyclin-dependent kinase (cdk) 4 and 6 binding and inhibiting activities.16 Their expression blocks the cell cycle at late G1 by inhibiting retinoblastoma gene protein product (Rb) phosphorylation by the CyclinDs-cdk4/6 complexes. p19^ARF is also a cell cycle inhibitor, but its biochemical function, which is distinct from that of the INK4 family members, remains unknown.11 A wide body of evidence supports the view that p16^INK4a is a tumor suppressor protein,6 and it is thought to represent the major functional target of genetic events occurring in the MTS locus. However, there is, to date, no definitive argument for discarding p19^ARF and p15^INK4b as functional targets of inactivating genetic events in cancer. Moreover, the demonstration that the MTS2 promotor is hypermethylated in various malignant hematologic disorders, including T-ALL, supports the view that MTS2is a tumor suppressor gene.17-19 

In this work, we have searched for evidence of p19^ARFinactivation in T-ALL. First, by analyzing the consequences onMTS1 expression of a particular rearrangement in three cases, we have excluded p16^INK4a as the target of the recombination and pinpointed p19^ARF and p15^INK4b as likely candidates. Second, by studyingMTS locus configuration in a large series of T-ALLs, we have shown that the p19^ARF encoding gene is always deleted/disrupted when chromosomal rearrangements occur in the locus, whereas p15^INK4b and p16^INK4a encoding exons are spared in a significant number of cases, suggesting that p19^ARF is targeted by the inactivating events.

Bone marrow or peripheral blood samples were obtained from 86 patients with T-ALL. Mononuclear cells from bone marrow or peripheral blood samples were isolated by Ficoll-Hypaque centrifugation and viably frozen in liquid nitrogen until use. A high percentage (usually more than 90%) of tumor cells was present in all samples. Patients were identified by unique numbers used in all reports from our group. Hela, RS4;11, Molt4, K562, and Saos2 cell lines were used in selected experiments.

Southern blots were performed as previously described.5Analysis was performed in the Molt4 cell line and in 87 samples from 86 T-ALLs including 58 previously studied patients (in one case, T39, theMTS locus configuration was different at presentation [T39P] and at relapse [T39R] and both samples have been included in the study). An MTS1 exon 1β probe7 was used in addition to the panel of probes used in our previous work.5Data were analyzed by comparison with the restriction map of the locus. The localization of the MTS1 exon 1β indicated in our first version of the MTS map5 was that published by Duro et al12 and has been subsequently found to be incorrect.7 We have precisely localized this exon and a corrected map is now available.7 

Total RNA was extracted from cryopreserved cells according to Chomczynski and Sacchi.20 Quality and quantity of RNA was controlled on an ethidium bromide stained 1% agarose gel containing 2.2 mol/L formaldehyde. RT-PCR analysis was performed as previously described.21 Amplification of MTS1α andMTS1β cDNA was performed with 2 or 2.5 mmol/L MgCl₂, respectively, and 5% formamide. After an initial denaturation step of 3 minutes at 94°C, a limited number of cycles (the number of cycles for each RT-PCR is given in the figure legends) consisting of 1 minute at 94°C, 1 minute at 60°C, 1 minute at 72°C followed by a final extension of 10 minutes at 72°C were performed on a DNA thermal cycler (Stratagene, La Jolla, CA). For all RNA samples, amplification of Abelson mRNA was performed as a control from a common cDNA. MTS1αtranscripts were amplified using primers OL18⁷: 5′-GCCCAACGCACCGAATAGT-3′ and OL29: 5′-GCTACCGGGTCGAGGAGTC-3′¹⁹ located on exon 1α and exon 2, respectively. MTS1β transcripts were amplified using primers OL27: 5′-ATGGTGCGCAGGTTCTTGGT-3′¹² and OL28: 5′-TGCACGGGTCGGGTGAGAGT-3′ located on exon 1β and exon 2, respectively. Abelson transcripts: abl1 (2nd exon): 5′-CCTTCAGCGGCCAGTAGCATCTGAC-3′ and abl2 (3rd exon): 5′-GGACACAGGCCCATGGTACCAGGAG-3′.

PCR products were run on 2% NuSieve (FMC BioProducts, Rockland, ME) 1% agarose gel, transferred to a Hybond N+ membrane (Amersham International plc, Les Ulis, France) and hybridized to 30 ng of a specific oligonucleotide probe labeled with 10 μCi γ-³²P–adenosine triphosphate (ATP) by T4 Polynucleotide Kinase (New England Biolabs Inc, Beverly, MA). MTS1αtranscripts: 5′-CGATCCAGGTCATGATGAT-3′. MTS1βtranscripts: 5′-GAAGACCAGGTCATGATGAT-3′. Abl transcripts: Abl3: 5′-TCCCAATTGTGATTATAGCC-3′.

Cell proteins were extracted with Triple Detergent Lysis Buffer, quantified using the BCA kit (Pierce, Rockford, IL), size fractionated on either 15% Tris-glycine sodium dodecyl sulfate-polyacrylamide gels (SDS-PAGE) for p16^INK4a, or on 8% Tris-glycine SDS-PAGE for Rb. After electrotransfer onto nitrocellulose, the membranes were blocked 1 hour in 10% nonfat dry milk in PBST (phosphate-buffered saline; Tween 20, 0.1%) and probed with antibodies. After a 1-hour incubation, the membranes were washed with PBST and incubated with peroxidase-conjugated secondary antibodies for 1 hour. After washes in PBST, membranes were revealed by chemiluminescence according to the manufacturer's instructions (ECL, Amersham). An antiactin antibody (Oncogene Science, Uniondale, NY) was used to verify protein loading and integrity. The RS4.11 cell line (which demonstrates a biallelicMTS1 deletion, unpublished data, 1995) and the Hela cell line (which demonstrates a high level of MTS1 protein expression) were used as negative and positive controls forMTS1 protein expression, respectively. The K562 myeloid leukemia cells and SAOS2 osteosarcoma cells were used as positive and negative controls for Rb protein expression, respectively. Antibodies used and their sources were as follows: anti-MTS1/p16^INK4a monoclonal antibody (Pharmingen, San Diego, CA), antibodies against Rb (Santa Cruz Biotechnology, Santa Cruz, CA), goat antirabbit IgG-peroxidase conjugate and goat antimouse IgG-peroxidase conjugate (Boehringer Mannheim, Reylan, France).

SSCP analysis of MTS1 exon 1α and MTS1 exon 2 was performed as described previously.5 Oligonucleotides used for MTS1 exon 3 SSCP analysis were: OL30: 5′-TTTTCTTTCTGCCCTCTGC-3′ located immediately upstream of the splice acceptor site, and OL31: 5′-TTGTGGCCCTGTAGGACCTT-3′ located 73 nucleotides downstream of the stop codon. SSCP analysis of MTS1exon 1β was performed on PCR products amplified using the following oligonucleotides: OL2: 5′-GGGGTGGGGGTGAAGGTG-3′ located 113 nt upstream of the putative transcriptional start site and OL4: 5′-ACGATTGAGGGCTGTGTGAAG-3′ located 204 nt downstream of the donor splice site. PCR products were digested by the NarI restriction endonuclease resulting in two fragments of 281 and 303 bp. Running conditions were identical to those described in Cayuela et al.5 

In four cases, a rearrangement sparing MTS1 exon 1α, exon 2, and exon 3 had occurred and was characterized in three cases (T39P, T117, and T127). An identical 36-kb interstitial deletion due to illegitimate V(D)J recombinase activity had occurred, juxtaposing the 5′ MTS2 exon 1 and 5′ MTS1 exon 1α regions. All MTS2 exons and the MTS1 exon 1β had been deleted.7 A schematic representation of the rearrangement is shown in Fig 1A and relevant clinical and hematologic data are summarized in Table 1.

Because the rearrangements have spared the MTS1 promoter located 5′ of the MTS1 exon 1α (see Fig 1 legend), we wondered if MTS1α transcripts were expressed and performed an RT-PCR analysis in T39p, T117, and T127 samples and in the Hela and RS4;11, positive and negative controls, respectively (Fig 1B). Clear expression was detected in the three primary leukemia blood samples, which contained more than 80% tumor cells. This expression reflects transcription from the rearranged allele, as the other is deleted7 and no expression was found in normal peripheral blood mononuclear cell (PBMC) under these experimental conditions (Fig1B).

To verify that the MTS1α transcripts encode proteins, p16^INK4a protein expression was studied by Western blotting. As shown in Fig 1C, clear expression was found in the T39p and T117 cases, but was barely detectable in the third case (T127). This expression parallels that of the MTS1α transcripts. SSCP analysis of MTS1 exons 1α, 2, and 3 was performed in all three cases (data not shown). No mutations were detected.

It has previously been suggested that MTS1α transcript expression is enhanced when Rb is physically or functionally inactivated (eg, by phosphorylation).22 We analyzed whether this correlation had been modified by the recombinational event. Rb phosphorylation was studied by Western blotting in the T39p, T117, and T127 samples and in the K562 and SAOS2 cell lines used as positive and negative controls, respectively. As shown in Fig 1D, hyperphosphorylated forms were found in the T39p and T117 cases, but were not detectable in T127, correlating with MTS1α transcript expression.

These data show that the interstitial deletion had spared the possibility to express MTS1α and its protein product p16^INK4a, suggesting that other proteins had been targeted by the chromosomal event. Because the rearrangement had disrupted the p19^ARF encoding exons and deletedMTS2/p15^INK4b, the sole gene identified in the 36-kb DNA region located between the two breakpoints, these results suggest that p15^INK4b and/or p19^ARFmight be the functional target(s) of the deletion in these cases.

The hypothesis that the gene encoding p19^ARF could be inactivated by genetic events in other cases of T-ALL was further studied.

The MTS locus configuration, as defined by Southern blot analysis in 87 T-ALL samples and in the Molt4 cell line, was evaluated with respect to the integrity of p16^INK4a, p19^ARF, and p15^INK4b encoding exons. A total of 176 alleles were analyzed.

The various allelic configurations are shown in Fig 2. Deletion and/or rearrangement had occurred in 149 alleles. Complete deletion/disruption of the locus was documented in 124 alleles (Type D alleles). The other rearrangements can be functionally classified into two types, A and B (Fig 2). In 21 cases, the recombinations have deleted the p16^INK4a encoding exons and disrupted the p19^ARF encoding exons, but spared the p15^INK4bencoding exons, defining the type A alleles. In four other cases (Type B alleles), and as detailed above, a rearrangement sparing MTS1exon 1α, exon 2, and exon 3, but deleting MTS1 exon 1β and the p15^INK4b encoding exons had occurred. Interestingly, the p19^ARF encoding exons were deleted or disrupted in all cases (149/149), in which a deletion or a chromosomal rearrangement had occurred in the MTS locus, suggesting that the product of these exons might be a functional target of these recombinational events.

The hypothesis that p19^ARF inactivation is targeted by the recombinational events imply that normal and/or neoplastic immature T cells have the potential to express MTS1βtranscripts. Previous studies have shown that these RNAs are expressed in most murine tissues,11 but expression in the thymus has not been studied to date. MTS1β transcript expression was therefore analyzed by RT-PCR in one human thymus. A low level of expression was detected, which is similar to the level found in bone marrow (Fig 3A). MTS1βtranscript expression was also studied in 10 T-ALL cases in which no genomic alteration of the corresponding exons was found by Southern blotting on at least one of the two alleles. Clear expression was detected in all cases, contrasting with an expected lack of expression in the T39P, T117, and T127 cases in which the MTS1 exon1β had been deleted. SSCP analysis of MTS1 exon1β was performed in eight of the cases expressing the corresponding transcripts (T81, T99R, T108, T115, T118, T121, T21, and T96). Normal SSCP patterns were found in all cases (data not shown). These results show that immature T cells can express p19^ARFencoding RNAs.

To date, no convincing data supporting the view that p19^ARFmight be a tumor suppressor protein have been published and p16^INK4a has been considered to be a more likely candidate, based on two arguments: (1) mutations have not been found in MTS1exon 1β (p19^ARF specific) whereas MTS1 exon 1α (p16^INK4a specific) mutations have been documented (reviewed in Quelle et al23); (2) cancer-associated inactivating mutations of MTS1 target p16^INK4a and not p19^ARF.23 Our data, however suggest that the possibility that the p19^ARF inactivation may be important in oncogenesis should be reconsidered. Indeed, we have shown that p19^ARF encoding exons are always deleted or disrupted when recombinational events occur in the MTS locus in T-ALL. Although we cannot exclude that these data may be due to the particular structure of the locus and to the localization of the V(D)J recognition signal sequences, which target the chromosomal rearrangements,7 we favor the hypothesis that p19^ARF may be one of the functional targets of the genetic events, which involve the MTS locus in the majority of T-ALLs and that its inactivation may be relevant to the pathogenesis of the disease.

We thank E. Macintyre for helpful comments. We thank the Laboratoire de photographie de l'Institut d'Hématologiéfor photographs. We also thank all G.B.B.M. members for discussions.