Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Acute Leukemias

PAX5 mutations occur frequently in adult B-cell progenitor acute lymphoblastic leukemia and PAX5 haploinsufficiency is associated with BCR-ABL1 and TCF3-PBX1 fusion genes: a GRAALL study


Adult and child B-cell progenitor acute lymphoblastic leukemia (BCP-ALL) differ in terms of incidence and prognosis. These disparities are mainly due to the molecular abnormalities associated with these two clinical entities. A genome-wide analysis using oligo SNP arrays recently demonstrated that PAX5 (paired-box domain 5) is the main target of somatic mutations in childhood BCP-ALL being altered in 38.9% of the cases. We report here the most extensive analysis of alterations of PAX5 coding sequence in 117 adult BCP-ALL patients in the unique clinical protocol GRAALL-2003/GRAAPH-2003. Our study demonstrates that PAX5 is mutated in 34% of adult BCP-ALL, mutations being partial or complete deletion, partial or complete amplification, point mutation or fusion gene. PAX5 alterations are heterogeneous consisting in complete loss in 17%, focal deletions in 10%, point mutations in 7% and translocations in 1% of the cases. PAX5 complete loss and PAX5 point mutations differ. PAX5 complete loss seems to be a secondary event and is significantly associated with BCR-ABL1 or TCF3-PBX1 fusion genes and a lower white blood cell count.


PAX5 (paired-box domain 5) is the guardian of the B-cell identity as stated in a recent review.1 This transcription factor belongs to the family of paired-box domain transcription factors.2 Its expression is initiated during early stages of B-cell differentiation beginning at the pro-B stage,3 and is turned off to allow terminal B-cell differentiation.4 The Pax5 homozygous deletion in murine models leads to the trans- or de-differentiation of B-cells into several other hematopoietic cell lineages.5, 6, 7 PAX5 is thus involved both in the maintenance of B-cell identity and in the control of terminal B-cell differentiation.

Deregulations and mutations of key differentiation factors are frequently found in lymphomas and leukemias. Translocations associated with hematologic malignancies involving PAX5 exemplify PAX5 dual function. On one hand, the t(9;14)(p13;q32) translocation brings the potent enhancer of the IGH gene close to the PAX5 promoter leading to an aberrant expression of a normal PAX5 protein.8, 9 This translocation is recurrent in small plasmacytoid B-cell lymphocytic lymphomas and diffuse large B-cell lymphomas.10 It emphasizes the importance of PAX5 downregulation during terminal B-cell differentiation.9 On the other hand, PAX5 translocations have also been associated with a block of early B-cell differentiation because the PAX5-ETV6 chimeric protein, product of the dic(9;12)(p13;p13), is associated with B-cell progenitor acute lymphoblastic leukemia (BCP-ALL).11 Additional PAX5 fusion partner genes have been identified as HIPK1 (chromosomal band 1p13),12, 13, 14 LOC392027 (7p12.1),15 AUTS2 (7q11.1),13 POM121 (7q11),14 ELN (7q11),16 JAK2 (9p24),14 SLCO1B3 (12p12),15 DACH1 (13q21),14 PML (15q24),17 ZNF521 (18q11.2),12 ASXL1 (20q11.1),15 C20orf112 (20q11.1),13, 14, 15 KIF3B (20q11.21)15 and BRD1 (22q13).14 PAX5-ELN,16 PAX5-FOXP112 and PAX5-ETV612 act as constitutive repressors of the remaining PAX5 allele product, explaining the block of B-cell differentiation.

To further emphasize the function of PAX5 in B-cell differentiation and oncogenesis, it has recently been reported that the PAX5 gene is the most frequent target of somatic mutations in childhood BCP-ALL, being altered in 38.9% of the cases.12 These mutations consist of partial or complete hemizygous deletions, homozygous deletions, partial or complete amplifications, point mutations or fusion genes.12 Some of these mutants have a dominant negative role on wild-type PAX5.12

Adult and childhood BCP-ALL can be considered as two distinct pathological entities in terms of pathogenesis and prognosis.18 Although 80% of the children with ALL can be cured, only 30% of the adults achieve long-term disease-free survival (DFS).18 Chromosomal abnormalities associated with these pathologies are different in term of occurrence. For example, BCR-ABL1 fusion gene is associated with a poor long-term response to chemotherapy and is the most common rearrangement associated with adult BCP-ALL accounting for 25% of the cases whereas it is very rarely found in pediatric BCP-ALL. Conversely, ETV6-RUNX1 fusion gene and high hyperdiploidy are associated with a good prognosis and account for half of the child BCP-ALL but rarely occur in adult BCP-ALL.18 Apart from these three biological entities, other categories such as TCF3-PBX1 fusion gene (also called E2A-PBX1) or MLL rearrangements or normal karyotypes occur with similar frequencies.18

Despite these discrepancies between childhood and adult BCP-ALL, we report here the high frequency of PAX5 mutations in a unique cohort of adult BCP-ALL treated according to the protocols of the GRAALL Intergroup (Group of Research on Adult Acute Lymphoblastic Leukemia), GRAALL-2003 (BCR-ABL1-negative BCP-ALL19) and GRAAPH-2003 (BCR-ABL1-positive BCP-ALL20). The deletion of one copy of PAX5 was found to be significantly associated with BCR-ABL1 or TCF3-PBX1 fusion genes and a lower white blood cell (WBC) count.

Material and methods

GRAALL-2003 and GRAAPH-2003 clinical protocols

The GRAALL includes the former LALA (Leucémies aiguës lymphoblastiques de l'adulte), the GOELAL (Groupe Ouest et Est des Leucémies Aigües Lymphoblastiques), and the Swiss Group for Clinical Cancer Research. The GRAALL-2003/GRAAPH-2003 study was a risk-adapted prospective phase 2 trial, conducted in 70 centers in France, Belgium and Switzerland ( Identifier: NCT00222027). Patients aged 15- to 60-year old with a newly diagnosed ALL were eligible. Between November 2003 and November 2005, 300 patients entered the study. Written informed consent was obtained from all patients or from the parents of those aged less than 18-year old before enrollment. The study was approved in March 2003 by the institutional review board of Purpan hospital, Toulouse, France, and was conducted in accordance with the Declaration of Helsinki Principles. All patients first received a common steroid prephase. Corticoresistance was defined as a peripheral blood blast cell count higher than 1.0 × 109 per liter at the end of this 7-day prephase. Chemoresistance was defined as a percentage of blasts higher than 5% at day 8 of the induction course. Patients were eligible for the GRAAPH-2003 study if they were diagnosed with a BCR-ABL1-positive ALL defined as ALL carrying the t(9;22) translocation on standard karyotype and/or fluorescent in situ hybridization (FISH) analysis and/or positivity for BCR-ABL1 fusion transcript detected by PCR analysis. The GRAAPH-2003 study evaluated the efficacy of imatinib mesylate combined to chemotherapy.20

PAX5 exon copy number

Quantification of PAX5 copy number was performed by quantitative PCR on genomic DNA from cells of 117 BCP-ALL included in the GRAALL-2003/GRAAPH-2003 trial and for which DNA material was available. Quantitative PCR was performed for each of the 10 exons in triplicate using Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA), specific PAX5 exon primers (Supplementary Table S1) and normal tonsil DNA as calibrator, each point of the dilution series being tested six times. The same measurements were performed with two reference genes located on chromosomes 15 (B2M gene) and 16 (CNGB1 gene), because these two chromosomes are rarely modified in number in ALL,21 using Ref15F and Ref15R primers for B2M and Ref16F and Ref16R primers for CNGB1 (Supplementary Table S1). The percentage of each PAX5 exon to the mean of Ref15 and Ref16 copy number was calculated. A value of 100% means that the number of PAX5 copies is identical to that of the references, and corresponds to two copies.

PAX5 point mutations

PAX5 DNA mutation screening was performed by sequence analysis on both strands (PAX5 exons 2, 3, 7, 8 and 9) and high-resolution melting PCR (HRM-PCR) for the remaining exons (Supplementary Table S1). Sequencing was performed using BigDye dideoxynucleotides and the products were separated on a 3130 XL sequencing apparatus (Applied Biosystems). Electrophoregrams were analyzed using the Sequencher software (version 4.1.2; Gene Codes Corporation, Ann Arbor, MI, USA) with a secondary peak threshold of 20% followed by manual verification. The presence of polymorphisms was evaluated using dbSNP and remission genomic DNA when available. HRM-PCR was performed using 1 × LightCycler 480 HRM Master Mix (Roche Applied Science, Mannheim, Germany) with 10 ng of genomic DNA, primers 0.1 μM and 25 mM MgCl2. HRM-PCR cycling conditions were initial denaturation at 95 °C during 10 min followed by 50 cycles at 95 °C for 10 s, at 63 °C for 15 s and at 72 °C for 25 s. Melting curve was measured from 70 to 95 °C with 25 acquisitions per °C.

PAX5 partial deletions

Full-length cDNA amplification of partial deletion mutants was performed by PCR on a 2720 Thermal Cycler (Applied Biosystems) with the Advantage 2 Polymerase Mix (Clontech, Mountain View, CA, USA). The amplification program was 5 cycles at 72 °C, 5 cycles at 70 °C and 25 cycles at 68 °C. Primers used were PAXQ and PCR5 (Supplementary Table S1). Reverse transcriptase (RT)–PCR products were analyzed by electrophoresis on an agarose gel, cut out and sequenced.

PAX5 FISH analysis

Fluorescent in situ hybridization analyses were performed on cases presenting either a chromosomal 9p13 breakpoint or deletion using a PAX5 commercial probe (PAX5 FISH DNA Probe, Split Signal, Y5413; Dako, Carpinteria, CA, USA) and/or a combination of two fluorescent-labeled bacterial artificial chromosome clones RP11-243F8 hybridizing PAX5 from exon 1 to 9 and RP11-344B23 hybridizing from exon 7 to 10 and to an extended telomeric region. The TCF3 split signal FISH probe (TCF3 FISH DNA Probe, Split Signal, Y5402; Dako) was also used.

IGH/TCR rearrangements analysis

IGH and TCR rearrangements were evaluated by the four centers performing the minimal residual disease follow-up of the clinical trial (Paris-Necker, Paris-Saint-Louis, Paris-Robert-Debré and Lille) according to the BIOMED-2 Concerted Action BMH4-CT98-3936-2 protocol.22 Immunoglobulin heavy chain specific amplification of the proximal VH6-1 segment rearrangement with JH was performed according to the BIOMED-2 protocol,22 using the VH6 FR2 primer along with the JH consensus primer (Supplementary Table S1).

Statistical analysis

Binary variables were compared with the two-sided Fisher's exact test. The Mann–Whitney test was used for median comparisons. The median follow-up of surviving patients for the entire protocol was 37 months. Event-free survival (EFS) was calculated from the date of prephase initiation. Events accounting for EFS were failure of remission induction, relapse and death in first complete remission (CR). Failure time data were estimated by the Kaplan–Meier method,23 then compared by the log-rank test.24 Cumulative incidence estimations took into account competing risks and were compared by the Gray test.25 A P-value less than 0.05 was considered to indicate statistical significance. All calculations were performed using the STATA/SE software, version 9.0 (Stata Corporation, College Station, TX, USA) and the R software, version 1.5.1 (The R Development Core Team, A Language and Environment Copyright, 2002).


Representativeness of the samples tested

We screened the occurrence of PAX5 mutations in a series of 117 BCP-ALL prospectively treated in the GRAALL-2003/GRAAPH-2003 study, with a median follow-up of 22 months (Table 1). The GRAALL-2003/GRAAPH-2003 included 300 adults with ALL. The 117 BCP-ALL cases analyzed for the occurrence of PAX5 mutations were similar to the 107 other BCP-ALL cases of the trial for whom no material remained available for analysis. We confirmed the absence of difference regarding main patient characteristics, including age and WBC count, as well as corticosensitivity, chemosensitivity, first CR rate, overall survival and BCR-ABL1 status (data not shown). Therefore, the population of 117 BCP-ALL analyzed for the presence of PAX5 mutations is representative of the whole GRAALL-2003/GRAAPH-2003 protocol.

Table 1 Statistical analysis of clinical and biological data

Genomic PAX5 copy number

Normal PAX5 content was found in 81 patients (69%), 20 had a complete hemizygous deletion (17%), 12 a partial hemizygous deletion (10%), 2 a partial amplification (2%) and 2 a complete amplification (2%) (Figure 1; Supplementary Table S2). PAX5 copy number results were validated by FISH analysis (Supplementary Table S3), karyotype (Supplementary Table S4) and by sequencing the partially deleted cDNA (Figure 2). These results indicate that alteration of PAX5 exon copy number is a frequent event in adult BCP-ALL occurring in 31% of the cases.

Figure 1

PAX5 (paired-box domain 5) mutations in adult B-cell progenitor acute lymphoblastic leukemia (BCP-ALL). Genomic DNA content for each of the 10 exons (X1–X10) was determined by quantitative PCR for each patient (GRAALL-2003/GRAAPH-2003 identification number given in the first column). Results (detailed in Supplementary Table S2) are expressed using color code (first frame): less than 100% of the control in green (deletion), 100% (=2n copies) in black (normal) and more than 100% in red (amplification). Patients were classified into four groups according to the PAX5 status: normal PAX5, patients with two PAX5 copies for each of the 10 exons; deleted PAX5, complete loss of one of the two PAX5 alleles or deletion from PAX5 exon 1; structural mutant PAX5, patients with partial deletion, partial amplification, point mutation or fusion gene, associated in some cases with deletion of the second allele; amplified PAX5, patients with a global amplification of PAX5. The second frame detailed the copy-number alterations. The third frame shows point mutations, N as normal. A star indicates when PAX5 exons 1, 4, 5, 6 and 10 were not investigated by high-resolution melting (HRM) analysis due to absence of remaining material. The fourth frame indicates fusion genes identified. The fifth frame provides the European Group for the Immunological Characterization of Leukemias (EGIL) status for each patient (BAL, bi-lineage acute leukemia; BL, B-cell lineage).

Figure 2

Validation of PAX5 (paired-box domain 5) exon copy number quantification. (a) PAX5 copy number quantification using quantitative PCR, 100% as two copies of a PAX5 exon. Standard deviations calculated on the normal PAX5 group are shown. Partial deletion of patients 66590 and 259440 are indicated. (b) PAX5 mRNA expression of patients 66590 and 259440 showing a partial deletion of PAX5. cDNA were amplified by reverse transcriptase (RT)–PCR between PAX5 exons 1 and 10. PCR products were analyzed on agarose gel. Lane 1 shows amplification of a 1505 bp normal RT–PCR product corresponding to the normal PAX5 mRNA, lane 2 shows a smaller 770 bp product corresponding to a PAX5 product with deletion from exon 2 to 6 and lane 3 a smaller 538 bp product corresponding to a PAX5 product with deletion from exon 2 to 8. (c) Sequence chromatographs of the amplified PCR products. The upper panel shows the direct fusion between exons 1 and 7 and the lower panel shows direct fusion between exons 1 and 9 confirming the deletion.

PAX5 genomic deletion as a secondary event

In two cases (patients 462837 and 119339), blasts carrying the TCF3-PBX1 fusion gene and deletion of one PAX5 allele were further analyzed using the relevant FISH probes together on the same slide. TCF3 rearrangement was detected in 90% (patient 462837) and 80% (patient 119339) of the nuclei. Of these TCF3-rearranged nuclei, only 60% (patient 462837) and 20% (patient 119339) of the nuclei, respectively, had a PAX5 deletion suggesting that, at least in these cases, PAX5 deletion is a secondary event (Supplementary Figure S1).

PAX5 point mutations

PAX5 point mutations were investigated by sequence and HRM-PCR analyses in the same 117 BCP-ALL samples (except for 8 patients for whom no material remained for HRM-PCR analysis). Eight mutations were identified among the 117 cases thus representing 7% of the cases (Supplementary Figures S2 and S3). Among them, the P80R was the most frequent point mutation, occurring in five out of eight cases (patients 21540, 114611, 115046, 165040 and 391099). The three other mutations were one point mutation S67N (patient 528234) and two frameshift mutations (insertion of one G at the base 2 of exon 4 (codon I138, patient 142309) and insertion of two G at the base 97 of exon 8 (codon G336, patient 520209)). Five point mutations were associated with a complete deletion of the second PAX5 allele (patients 114611, 115046, 391099, 520209 and 528234) (Figure 1).

PAX5 chromosomal rearrangement

Fluorescent in situ hybridization analysis was performed when a chromosomal abnormality involving the 9p chromosomal band was identified by standard karyotype. Eleven patients showed a rearrangement of the 9p13 chromosomal band (patients 49490, 49833; 41756; 62902; 66590; 114117, 144051, 144948, 196712, 301228 and 391099) and in a single case (patient 41756: 46,XY,t(7;9)(q21;p21)[17]) a translocation which fused PAX5 and ELN without loss of the remaining normal allele, was detected.16 Thus PAX5 translocations occur in adult BCP-ALL but are a rare event (<1%).

PAX5 alterations in BCP-ALL

Globally, PAX5 alterations were identified in 40 cases out of 117 adult BCP-ALL (34%). These alterations are heterogeneous consisting in complete loss in 17%, focal deletions in 10%, point mutations in 7% and translocations in 1% of the cases. To further analyze these mutations, we apportioned the patients according to the probable consequences of mutations (Figure 1). The PAX5 mutated BCP-ALL were subdivided in three groups. A first group called deleted PAX5, pooling 17 cases (15%), characterized by a probable lower expression of the normal PAX5 protein due to a complete deletion of one PAX5 allele or a partial deletion removing PAX5 exon 1 with no mutation of the second allele. A second group called structural mutant PAX5 bringing together 21 cases (18%) characterized by the expression of a PAX5 mutant allele, that is, having an altered activity regarding its transcriptional function consisting either of PAX5 partial deletion conserving PAX5 exon 1 (10 cases), partial amplification (2 cases), point mutations (8 cases) and fusion gene (1 PAX5-ELN case). A third group called amplified PAX5 consisting of two cases (2%) that could lead to a higher expression of the normal PAX5 protein as a consequence of a whole PAX5 genomic amplification. Owing to the small number of cases, this group was not further analyzed.

Characteristics of adult PAX5 mutants BCP-ALL

We compared the two main groups of PAX5 mutations (deleted PAX5 and structural mutant PAX5) to the largest category composed of patients without evidence of PAX5 mutation (77 cases, 66%, labeled as normal PAX5).


We analyzed the membrane expression of major B-cell differentiation markers. We first analyzed CD19 as it is a direct target of PAX5.26 However, no difference according to the PAX5 status was observed (Table 1). CD20-positive cells were not significantly different in deleted PAX5 (8 of 16, 50%), structural mutant PAX5 (10 of 20, 50%) and normal PAX5 groups (24 of 70, 34%) (Table 1). In contrast, no CD10-negative cases were detected in the deleted PAX5 group (0 of 17) compared to 21 of the 71 (30%) normal PAX5 cases (P=0.018) (Table 1). This difference is probably related to the fact that CD10-negative BCP-ALL are associated with MLL translocations27 and PAX5 alterations are not associated with MLL translocations (to the unique exception of patient 467102) The structural mutant PAX5 group showed an intermediate result because 2 of 21 BCP-ALL (10%) lacked CD10 surface expression.

European Group for the Immunological Characterization of Leukemias (EGIL) classification allows a subdivision of BCP-ALL according to the stage of maturation block of the leukemic cells.28 B-I represents the earliest stage and B-IV the latest stage of differentiation of BCP-ALL. Because the EGIL classification used the detection of CD10, B-I being CD10-negative BCP-ALL and B-II to B-IV being CD10 positive, we found an imbalanced repartition in EGIL subset between deleted and normal PAX5 groups (P=0.002; Table 1).

Cytogenetics and fusion genes

Fusion gene rearrangements are frequent in BCP-ALL, especially BCR-ABL1 in adult BCP-ALL. We evaluated the association of these fusion genes or other karyotypic features regarding PAX5 status (Tables 1 and 2). We found a significant association between the presence of BCR-ABL1 (53%) or TCF3-PBX1 (18%) and the deleted PAX5 group compared to the normal group (19% BCR-ABL1 and 3% TCF3-PBX1) or the structural mutant PAX5 group (29% BCR-ABL1 and 0% TCF3-PBX1) (P=0.009; Table 1). No MLL rearrangement and no normal karyotype were detected in the deleted PAX5 group as compared to 10 (13%) and 20 cases (26%), respectively, in the normal PAX5 group.

Table 2 Cytogenetic data

IGH VHDHJH and TCR rearrangement

Because in mice Pax5 homozygous deletion impairs B-cell transition from DHJH to VHDHJH IGH rearrangement,5 we investigated the rearrangement status of IGH in the context of PAX5 mutational status. Complete VHDHJH IGH rearrangements were detected in 76% (42 of 55), 91% (10 of 11) and 69% (9 of 13) of normal, deleted and mutant PAX5 subgroups respectively. These differences were not statistically significant (Table 1). Although VHDHJH rearrangement occurs, VH segment proximal vs distal accessibility could be impaired in a similar way to the IL7R homozygous deletion model.29 VHDHJH rearrangements were sequenced in a unique center (Paris-Necker) in 19 BCP-ALL (12 normal PAX5, 3 deleted PAX5 and 4 structural mutant PAX5). Intriguingly, the two PAX5 P80R mutants analyzed showed a rearrangement involving the proximal VH6-1 segment, the closest VH segment to JH segments (only 1 of 12 in normal PAX5 BCP-ALL). This VH segment is usually rarely used in physiological situations and this bias suggests a possible impairment of accessibility to the IGH locus with this PAX5 P80R mutant. We therefore analyzed the use of the VH6-1 segment and found that P80R (two of three analyzed) and S67N mutants used this VH6-1 (data not shown). We also investigated the occurrence of illegitimate rearrangements of the TCRG and TCRD loci but no differences were detected according to PAX5 mutational status.

Prognostic significance

There was no difference in median age between the normal, deleted and mutant PAX5 subgroups (Supplementary Tables 1 and S5). Of note, median WBC was significantly lower in the deleted PAX5 subgroups as compared to the two other subgroups (3.4 vs 13.8 (normal) and 11.4 G/L (structural mutations); P=0.01; Table 1).

Two patients eventually were not treated according to the protocol and the two patients showing an amplification of PAX5 were not included in this analysis. In the 113 remaining patients, CR rate was 68 of 76 (89%), 16 of 17 (94%) and 18 of 20 (90%) in the normal, deleted and structural mutant PAX5 subgroups respectively (P=0.03; Table 1). The percentage of corticoresistant leukemia was 24, 18 and 20%, respectively, in these three subgroups (not significant; Table 1). A total of 42 patients were allografted in first CR (28 from the normal PAX5, 5 from the deleted PAX5 and 9 from the structural mutant PAX5 subgroup, respectively). Cumulative incidence of relapse and DFS are shown in Figure 3, according to the PAX5 status. As indicated, there were no significant differences in outcome among the three subsets (Figure 3a). However, it is worth noting that a trend toward a higher incidence of relapse in the structural mutant PAX5 subgroup was observed (Figure 3b). Results were similar when patients allografted in first CR were censored at transplant time (not shown). Of note, patients with BCR-ABL1-positive ALL (treated here with imatinib combined to conventional chemotherapy) displayed a similar outcome to those with BCR-ABL1-negative ALL in this series (not shown).

Figure 3

Disease-free survival (DFS) and cumulative incidence of relapse according to PAX5 (paired-box domain 5) status (a) DFS. At 3 years, estimated DFS was 57% (95%; CI 44–69%), 64% (95%; CI 38–84%) and 43% (95%; CI 20–64%), in the normal, deleted and structural mutant PAX5 group, respectively (P=0.33, by the log-rank test). (b) Cumulative incidence of relapse. At 3 years, estimated cumulative incidence of relapse was 29, 22 and 48%, in the normal, deleted and structural mutant PAX5 subgroup, respectively (P=0.22, by the Gray test).


We report here that PAX5 is frequently mutated in adult BCP-ALL. PAX5 mutations have been extensively reported in pediatric BCP-ALL.12 In adult BCP-ALL so far, only PAX5 deletion has been reported in small series of 26 adult and adolescent BCP-ALL30 and 22 BCR-ABL1 positive adult BCP-ALL.31 We investigated in this study, PAX5 deletions but also PAX5 structural mutations in a large series of 117 adult BCP-ALL in a unique clinical protocol. Out of the patients analyzed from the GRAALL-2003/GRAAPH-2003 clinical trial, 40 carried a mutation of the PAX5 gene (34%). PAX5 mutations occur in different ways, as complete or partial deletions, complete or partial amplifications, point mutations or fusion genes.12 We analyzed PAX5 mutants according to their probable functional consequences according to the study by Mullighan et al.12 Isolated complete deletion of one PAX5 allele is found in 15% of adult BCP-ALL (deleted PAX5 group) and 11% of child BCP-ALL.12 Mutant alleles are found in 18% of adult BCP-ALL (mutant PAX5 group) and 22% of children BCP-ALL.12 PAX5 point mutations are detected in 7% of adult BCP-ALL and 7% of child BCP-ALL.12 PAX5 P80R mutation is a frequent event in BCP-ALL detected in 5 of 117 cases in adult BCP-ALL and 4 of 192 cases in child12 occurring therefore in 3% of BCP-ALL. We identified three new mutations such as PAX5 S67N, one G insertion in codon 138 (exon 4) and one insertion of two G in codon 336 (exon 8) leading both to a frameshift of PAX5 last exons encoding the transactivation and inhibitory domains. Therefore, PAX5 mutations appear to be similar, in nature and frequency, between child and adult BCP-ALL.

PAX5 maintains the identity of B cells.7 Its homozygous deletion in mice blocks B-cell differentiation at an early stage.5, 6 In our series, out of the 17 cases of hemizygous deletion (deleted PAX5 group), no early stage CD10-negative BCP-ALL were identified in contrast to the 15 of 74 cases (20%) found in the normal PAX5 group. The proportion of these BCP-ALLs in the PAX5 structural mutant group was intermediate between normal PAX5 and deleted PAX5 (2 of 21 cases, 10%). This result suggests that the hemizygous deletion, that is, a lower dose of normal PAX5, is either a later event or blocks the leukemic cells at a later stage during B-cell differentiation. An alternative explanation is linked to the absence of PAX5 deletions in MLL-rearranged cases, and consequently the absence of CD10-negative BCP-ALL, the hallmark of MLL rearrangements,27 is merely a consequence of this exclusion.

Fusion genes BCR-ABL1, TCF3-PBX1 or involving MLL are frequently detected in adult BCP-ALL. We identified these fusion genes in 26, 4 and 9% of our cases respectively. PAX5 deletions are highly skewed toward BCR-ABL1 and TCF3-PBX1 fusion genes, occurring in 71% of patients in deleted PAX5 compared to 22% in normal PAX5 and 29% in mutant PAX5, suggesting a very important role of PAX5 dosage during the transformation process of these two oncogenes. It is of note that the frequency of PAX5 deletions is similar in child and adult BCP-ALL, even if associated events such as BCR-ABL1 are not similarly distributed. In addition, PAX5 deletion seemed to be a secondary event as suggested by our FISH analysis performed on blasts carrying both TCF3-PBX1 fusion gene and PAX5 deletion.

Structural mutations are heterogeneous, consisting of partial deletions, partial amplifications, point mutations and fusion genes. The complete deletion of one PAX5 allele is frequently associated with deletion of CDKN2A located telomeric of PAX5 in the short arm of chromosome. Transduction using a BCR-ABL1 retrovirus in bone marrow cells of mice with only one copy of PAX5 shorten drastically the survival (median from 60 to 36 days).32 Moreover the concomitant haploinsufficiencies of PAX5 and CDKN2A reduce again significantly the survival to a median of 21 days.32 This demonstrates that loss of one copy of CDKN2A and PAX5 is synergistic during the BCR-ABL1 transformation. PAX5 point mutations, such as P80R, and PAX5 fusion genes are isolated events, not associated with BCR-ABL1. Partial deletion mutants target either the DNA-binding domain or the transactivation domain. Deletion of the DNA-binding domain only is associated with BCR-ABL1 (two of the four cases). Frameshift structural mutants (I138 and G336) are also associated with BCR-ABL1 (for the two cases).

PAX5 point mutations are frequently associated with the deletion of the remaining allele (10 of 13 pediatric cases;12 5 of 8 cases in our series) suggesting that the mechanism of action of these point mutations during the oncogenic process may not be associated to a dominant-negative effect. By contrast, a dominant-negative action was established for PAX5-ELN,16 PAX5-ETV611, 12 and PAX5-FOXP112 fusions.

It has been reported that the most proximal VH segment, VH6-1, is overused in adult B-ALL patients.33 Furthermore, Pax5 homozygous deletion in mice showed that Pax5 is crucial for the transition from DHJH to VHDHJH IGH rearrangement, and especially for distal rearrangement.34 Although we were unable to detect a significant frequency difference regarding IGH VHDHJH rearrangement between each group of BCP-ALL, we have therefore looked at the rearrangement at VH6-1. Our results confirm that in average 11.4% (12 of 105) of the adult BCP-ALL tested overuse the VH6-1 segment, suggesting a modification of the accessibility of the other VH segments in this pathology, with a bias toward PAX5 DNA-binding mutants such as S67N or P80R.

In conclusion, PAX5 mutations are frequent in adult BCP-ALL in accordance with the pediatric BCP-ALL cases.12 Our data clearly show a difference regarding the type of PAX5 mutations (deleted PAX5 or structural mutant PAX5) in term of association to fusion genes, EGIL classification and WBC counts. The loss of an entire allele of PAX5 seems to be a rather late event and might be considered as a secondary event.


  1. 1

    Cobaleda C, Schebesta A, Delogu A, Busslinger M . Pax5: the guardian of B cell identity and function. Nat Immunol 2007; 8: 463–470.

    CAS  Article  Google Scholar 

  2. 2

    Robson EJ, He SJ, Eccles MR . A PANorama of PAX genes in cancer and development. Nat Rev Cancer 2006; 6: 52–62.

    CAS  Article  Google Scholar 

  3. 3

    Adams B, Dorfler P, Aguzzi A, Kozmik Z, Urbanek P, Maurer-Fogy I et al. Pax-5 encodes the transcription factor BSAP and is expressed in B lymphocytes, the developing CNS, and adult testis. Genes Dev 1992; 6: 1589–1607.

    CAS  Article  Google Scholar 

  4. 4

    Fuxa M, Busslinger M . Reporter gene insertions reveal a strictly B lymphoid-specific expression pattern of pax5 in support of its B cell identity function. J Immunol 2007; 178: 3031–3037.

    CAS  Article  Google Scholar 

  5. 5

    Nutt SL, Heavey B, Rolink AG, Busslinger M . Commitment to the B-lymphoid lineage depends on the transcription factor Pax5. Nature 1999; 401: 556–562.

    CAS  Article  Google Scholar 

  6. 6

    Rolink AG, Nutt SL, Melchers F, Busslinger M . Long-term in vivo reconstitution of T-cell development by Pax5-deficient B-cell progenitors. Nature 1999; 401: 603–606.

    CAS  Article  Google Scholar 

  7. 7

    Cobaleda C, Jochum W, Busslinger M . Conversion of mature B cells into T cells by dedifferentiation to uncommitted progenitors. Nature 2007; 449: 473–477.

    CAS  Article  Google Scholar 

  8. 8

    Iida S, Rao PH, Nallasivam P, Hibshoosh H, Butler M, Louie DC et al. The t(9;14)(p13;q32) chromosomal translocation associated with lymphoplasmacytoid lymphoma involves the PAX-5 gene. Blood 1996; 88: 4110–4117.

    CAS  PubMed  Google Scholar 

  9. 9

    Busslinger M, Klix N, Pfeffer P, Graninger PG, Kozmik Z . Deregulation of PAX-5 by translocation of the Emu enhancer of the IgH locus adjacent to two alternative PAX-5 promoters in a diffuse large-cell lymphoma. Proc Natl Acad Sci USA 1996; 93: 6129–6134.

    CAS  Article  Google Scholar 

  10. 10

    Poppe B, De Paepe P, Michaux L, Dastugue N, Bastard C, Herens C et al. PAX5/IGH rearrangement is a recurrent finding in a subset of aggressive B-NHL with complex chromosomal rearrangements. Genes Chromosomes Cancer 2005; 44: 218–223.

    CAS  Article  Google Scholar 

  11. 11

    Cazzaniga G, Daniotti M, Tosi S, Giudici G, Aloisi A, Pogliani E et al. The paired box domain gene PAX5 is fused to ETV6/TEL in an acute lymphoblastic leukemia case. Cancer Res 2001; 61: 4666–4670.

    CAS  PubMed  Google Scholar 

  12. 12

    Mullighan CG, Goorha S, Radtke I, Miller CB, Coustan-Smith E, Dalton JD et al. Genome-wide analysis of genetic alterations in acute lymphoblastic leukaemia. Nature 2007; 446: 758–764.

    CAS  Article  Google Scholar 

  13. 13

    Kawamata N, Ogawa S, Zimmermann M, Niebuhr B, Stocking C, Sanada M et al. Cloning of genes involved in chromosomal translocations by high-resolution single nucleotide polymorphism genomic microarray. Proc Natl Acad Sci USA 2008; 105: 11921–11926.

    CAS  Article  Google Scholar 

  14. 14

    Nebral K, Denk D, Attarbaschi A, Konig M, Mann G, Haas OA et al. Incidence and diversity of PAX5 fusion genes in childhood acute lymphoblastic leukemia. Leukemia 2009; 23: 134–143.

    CAS  Article  Google Scholar 

  15. 15

    An Q, Wright SL, Konn ZJ, Matheson E, Minto L, Moorman AV et al. Variable breakpoints target PAX5 in patients with dicentric chromosomes: a model for the basis of unbalanced translocations in cancer. Proc Natl Acad Sci USA 2008; 105: 17050–17054.

    CAS  Article  Google Scholar 

  16. 16

    Bousquet M, Broccardo C, Quelen C, Meggetto F, Kuhlein E, Delsol G et al. A novel PAX5-ELN fusion protein identified in B-cell acute lymphoblastic leukemia acts as a dominant negative on wild-type PAX5. Blood 2007; 109: 3417–3423.

    CAS  Article  Google Scholar 

  17. 17

    Nebral K, Konig M, Harder L, Siebert R, Haas OA, Strehl S . Identification of PML as novel PAX5 fusion partner in childhood acute lymphoblastic leukaemia. Br J Haematol 2007; 139: 269–274.

    CAS  Article  Google Scholar 

  18. 18

    Armstrong SA, Look AT . Molecular genetics of acute lymphoblastic leukemia. J Clin Oncol 2005; 23: 6306–6315.

    CAS  Article  Google Scholar 

  19. 19

    Huguet F, Leguay T, Raffoux E, Thomas X, Beldjord K, Delabesse E et al. Pediatric-inspired therapy in adults with Philadelphia chromosome-negative acute lymphoblastic leukemia: the GRAALL-2003 Study. J Clin Oncol 2009; 27: 911–918.

    CAS  Article  Google Scholar 

  20. 20

    de Labarthe A, Rousselot P, Huguet-Rigal F, Delabesse E, Witz F, Maury S et al. Imatinib combined with induction or consolidation chemotherapy in patients with de novo Philadelphia chromosome-positive acute lymphoblastic leukemia: results of the GRAAPH-2003 study. Blood 2007; 109: 1408–1413.

    CAS  Article  Google Scholar 

  21. 21

    Heerema NA, Raimondi SC, Anderson JR, Biegel J, Camitta BM, Cooley LD et al. Specific extra chromosomes occur in a modal number dependent pattern in pediatric acute lymphoblastic leukemia. Genes Chromosomes Cancer 2007; 46: 684–693.

    CAS  Article  Google Scholar 

  22. 22

    van Dongen JJ, Langerak AW, Bruggemann M, Evans PA, Hummel M, Lavender FL et al. Design and standardization of PCR primers and protocols for detection of clonal immunoglobulin and T-cell receptor gene recombinations in suspect lymphoproliferations: report of the BIOMED-2 Concerted Action BMH4-CT98-3936. Leukemia 2003; 17: 2257–2317.

    CAS  Article  Google Scholar 

  23. 23

    Kaplan E, Meier P . Nonparametric estimation from incomplete observations. J Am stat Assoc 1958; 53: 457–481.

    Article  Google Scholar 

  24. 24

    Peto R, Peto J . Asymptotically efficient rank invariant test procedures. J R Stat Soc 1972; 135: 185–206.

    Google Scholar 

  25. 25

    Gray R . A class of k-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat 1998; 16: 1141–1154.

    Article  Google Scholar 

  26. 26

    Kozmik Z, Wang S, Dorfler P, Adams B, Busslinger M . The promoter of the CD19 gene is a target for the B-cell-specific transcription factor BSAP. Mol Cell Biol 1992; 12: 2662–2672.

    CAS  Article  Google Scholar 

  27. 27

    Pui CH, Behm FG, Downing JR, Hancock ML, Shurtleff SA, Ribeiro RC et al 11q23/MLL rearrangement confers a poor prognosis in infants with acute lymphoblastic leukemia. J Clin Oncol 1994; 12: 909–915.

    CAS  Article  Google Scholar 

  28. 28

    Bene MC, Castoldi G, Knapp W, Ludwig WD, Matutes E, Orfao A et al. Proposals for the immunological classification of acute leukemias. European Group for the Immunological Characterization of Leukemias (EGIL). Leukemia 1995; 9: 1783–1786.

    CAS  PubMed  Google Scholar 

  29. 29

    Bertolino E, Reddy K, Medina KL, Parganas E, Ihle J, Singh H . Regulation of interleukin 7-dependent immunoglobulin heavy-chain variable gene rearrangements by transcription factor STAT5. Nat Immunol 2005; 6: 836–843.

    CAS  Article  Google Scholar 

  30. 30

    Paulsson K, Cazier JB, Macdougall F, Stevens J, Stasevich I, Vrcelj N et al. Microdeletions are a general feature of adult and adolescent acute lymphoblastic leukemia: unexpected similarities with pediatric disease. Proc Natl Acad Sci USA 2008; 105: 6708–6713.

    CAS  Article  Google Scholar 

  31. 31

    Mullighan CG, Miller CB, Radtke I, Phillips LA, Dalton J, Ma J et al. BCR-ABL1 lymphoblastic leukaemia is characterized by the deletion of Ikaros. Nature 2008; 453: 110–114.

    CAS  Article  Google Scholar 

  32. 32

    Miller CS, Mullighan CG, Su X, Ma J, Wang M, Zhang J et al. Pax5 haploinsufficiency cooperates with BCR-ABL1 to induce acute lymphoblastic leukemia. Blood 2008; 112: 114.

    Article  Google Scholar 

  33. 33

    Mortuza FY, Moreira IM, Papaioannou M, Gameiro P, Coyle LA, Gricks CS et al. Immunoglobulin heavy-chain gene rearrangement in adult acute lymphoblastic leukemia reveals preferential usage of J(H)-proximal variable gene segments. Blood 2001; 97: 2716–2726.

    CAS  Article  Google Scholar 

  34. 34

    Nutt SL, Urbanek P, Rolink A, Busslinger M . Essential functions of Pax5 (BSAP) in pro-B cell development: difference between fetal and adult B lymphopoiesis and reduced V-to-DJ recombination at the IgH locus. Genes Dev 1997; 11: 476–491.

    CAS  Article  Google Scholar 

Download references


We thank the financial support of the association Laurette-Fugain. JF was supported by an ARC fellowship. CB and NPH were supported by a grant of Institut National du Cancer. ED, CB and PB were supported by the CITTIL (Cooperacion de investigacion transpirenaica en la terapia innovadora de la leucemia).

This work would not have been possible without the help of all the people who take care of the patients involved in the GRAALL studies, especially the centers that contributed directly to this study. Amiens: Damaj, Royer, Dubus, Capiod, Marolleau; Angers: Francois, Hunault, Ifrah, Marie, Genevieve, Baranger, Chassevent, Blanchet; Avignon: Boulat, Derre; Bayonne: Banos, Bauduer, Burtin; Bobigny: Gardin, Fenaux, Beve, Boulalam, Eclache, Fenaux; Bordeaux: Boiron, Leguay, Pigneux, Bilhou-Nabera, Perry, Lacombe, Tabrizi, Lippert, Marit; Brest: Guillerme, Berthou, Lecalvez, De Braekeleer, Ugo; Caen: Reman, Lepesant, Salaun, Plessis, Naguib, Leporrier; Clamart: De Revel, Samson, Desangles; Clermont-Ferrand: Chaleteix, Villemagne, Latiere, Berger, Giollant, Tchirkov, Tournilhac; Dijon: Caillot, Casanovas, Grandjean, Menadier, Favre-Audry, Mugneret, Teyssier; Lens: Stalnikiewicz, Poulain; Lille: Darre, De Botton, Lepelley, Lai, Soenen, Preudhomme, Grardel, Bauters; Limoges: Turlure, Bordessoule, Chaury, Trimoreau, Gachard; Lyon: Le, Nicolini, Tavernier, Thiebaut, Thomas, Lheritier, Girard, Wattel, Tigaud, Hayette, Michallet; Marseille: Vey, Charbonnier, Stoppa, Mouton, Sainty, Moziconacci, Arnoulet, Lafage-Pochitaloff, Blaise; Meaux: Frayfer, Mossafa; Mulhouse: Arkam, Ojeda Uribe, Iglarz, Drenou, Jeandidier, Isaac; Nancy: Witz, Bene, Witz, Gregoire, Monhoven; Necker: Buzyn, Couderc, Asnafi, Valensi, Radford-Weiss, Delabesse, Macintyre, Varet; Paris Pitié-Salpétrière: Dhedin, Aliammar, Merle-Beral, Nguyen-Khac, Davi, Leblond, Vernant; Paris Saint-Louis: Raffoux, Treilhou, Maarek, Daniel, Soulier, Cayuela, Miclea, de Labarthe, Dombret; Reims: Himberlin, Baury, Daliphard, Luquet, Cornillet-Lefebvre, Delmer; Rennes: Escoffre-Barbe, Lamy, Picouleau, Roussel, Henry, Ly Sunnaram, Fest; Rouen: Lepretre, Contentin, Jardin, Lenain, Tilly, Tallon, Lenormand, Stamatoullas-Bastard, Penther, Bastard; Saint-Etienne: Cornillon, Jaubert, Guyotat, Marchand, Campos, Nadal, Flandrin; Toulon: De Jaurreguiberry; Toulouse: Huguet, Recher, Daniel, Kuhlein, Dastugue, Demas, Attal; Tours: Delain, Delepine, Degene, Barin, Colombat; Valenciennes: Fernandes, Poulain, Daudignon; Versailles: Choquet, Rousselot, Taksin, Pousset, Terre, Castaigne; Villejuif: Arnaud, Bayle, Bourhis, Auger, Bernheim.

Among these participants, we specially acknowledge the following cytogeneticists that provided cytogenetic pellets and data to perform PAX5 and TCF3 FISH analyses, Eric Lippert (Bordeaux), Odile Maarek (Paris Saint-Louis), Christian Bastard and Dominique Penther (Rouen), Isabelle Tigaud (Lyon), Florence Nguyen-Khac (Paris Pitié-Salpétrière), Christine Terré (Versailles) and Ghislaine Plessis (Caen). We thank the extended FISH analysis of Francesca Correia (Toulouse).

The design and analysis of the experiments were performed by ED and CB. JF performed the PAX5 quantitative PCR. MB cloned and analyzed the PAX5-ELN case. They cloned the PAX5 mutants helped by ND, SS, EC, CQ, NPH and SD. ED, CB and PB overviewed the results. The cytogenetic results were collected and analyzed by MLP and ND. FISH analyses were performed by MLP and CB. The immunophenotype results were collected by MCB. The molecular data were collected by EAM and ED. KB performed VHDHJH sequence analysis. JDV performed the microarray analysis. JMC, NG, CP, HC, OB, KB and EAM provided DNA samples. VL collected clinical data, reviewed by YC, NI, AD, AP, FH and HD. Paper was written by ED, CB, JF and HD.

Author information



Corresponding author

Correspondence to É Delabesse.

Additional information

Supplementary Information accompanies the paper on the Leukemia website (

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Familiades, J., Bousquet, M., Lafage-Pochitaloff, M. et al. PAX5 mutations occur frequently in adult B-cell progenitor acute lymphoblastic leukemia and PAX5 haploinsufficiency is associated with BCR-ABL1 and TCF3-PBX1 fusion genes: a GRAALL study. Leukemia 23, 1989–1998 (2009).

Download citation


  • oncogenesis
  • BCR-ABL1
  • PAX5
  • TCF3-PBX1

Further reading


Quick links