Chromosomal rearrangement involving 11q23 locus in chronic myelogenous leukemia: a rare phenomenon frequently associated with disease progression and poor prognosis

Although frequently encountered in AML and acute lymphoid leukemia (ALL), 11q23 rearrangement occurs very rarely in CML, in less than 1% of cases in this study by conventional karyotype. The translocation partners for 11q23 are diverse with t(9;11) (p22;q23) being the most common. Most cases developed 11q23 translocation in BP, and were often associated with other chromosomal changes. Patients had a low response rate to TKI treatment, and the overall survival was poor.

Clonal evolution with additional chromosomal abnormalities besides t(9;22) is considered as a feature of disease progression in CML. The predominant emergence of 11q23 in BP (8/11) suggests that 11q23/MLL may play a direct role in CML blast transformation. In AML and ALL, the role of MLL in leukemic transformation has been extensively studied. Although the fusion partners for MLL are diverse with more than 70 fusion partners described, most MLL rearrangements join the N-terminal region of MLL with C-terminal portion of its partners, forming chimeric oncoproteins [21]. Although the mechanisms of MLL-induced leukemia are complex and not fully understood, the chimeric proteins have been shown to induce HOX expression, which can block hematopoietic differentiation and induce leukemia [20]. Whether MLL plays a similar role in CML blastic transformation needs further confirmation.

The prognostic significance of clonal evolution in CML is variable and depends on many factors, including the particular type of cytogenetic changes, the time and phase of emergence of clonal evolution, other concurrent chromosomal abnormalities, the presence or absence of other AP features, and treatment regimens. In the era of pre-TKI therapy, INFa induced a complete suppression of clonal evolution in 46% of patients with only Ph chromosome and/or diploid karyotype present after resolution of clonal evolution. About 7% of patients with clonal evolution achieved a complete cytogenetic remission on INFa therapy, and the absence of other AP features was associated with a better response [23]. With the first generation of TKI, imatinib, one study showed that clonal evolution had no significant effect on achieving major or complete cytogenetic response, but it was a poor prognostic factor for survival in both CP and AP [24]. Another study showed that clonal evolution was associated with lower cytgenetic response rate in patients in AP with other signs of acceleration [25]. For patients on second generation of TKI therapy, cases with clonal evolution but without other features of AP had similar hematologic/cytogenetic response and overall survival to cases in CP without clonal evolution. However, when other features of AP were present, the presence of clonal evolution adversely affected overall survival [26].

As mentioned above, the prognostic significance of clonal evolution also depends on the specific type of chromosomal changes. Fabarius et al. reported that the major route, not minor route cytogenetic aberrations at initial diagnosis, had a negative prognostic impact on CML patients [27]. Consistently, a recent study indicated that trisomy 8, chromosome 17 abnormalities, and complex cytogenetic changes carried worse prognosis than others [26]. Accordingly, in European LeukemiaNet recommendations, the major routes, not the minor routes of chromosomal abnormalities, are defined as one of criteria for AP [28]. Although 11q23 abnormalities are considered as “minor route” abnormalities, they appear to carry a poor prognosis. In our cohort, nine patients (9/11) died at the last follow-up. The majority of cases had 11q23 abnormalities in BP(8/11). Only one patient (case 1) had 11q23 in CP, in the absence of other AP features. This patient achieved complete cytogenetic response and major molecular response after nilotinib treatment. Two patients (cases 2 and 3) had 11q23 in AP, accompanied with other AP features including increased blast count and thrombocytopenia. These two cases showed poor response to TKI therapy and transformed to blast phase after a period of 1.6 and 12 months.

Interestingly, not all cases with 11q23 rearrangements showed MLL involvement by FISH analysis. In our series, 30% (3/10) cases showed negative MLL rearrangement by FISH study using a commercial MLL breakapart probe. Although not frequent, similar findings had been described in CML and other myeloid neoplasms. Giurliano et al. reported two cases of myeloid neoplasms (AML and atypical CML) with cytogenetic changes involving 11q23 but with intact MLL gene by FISH: one with t(3;11)(q21;q23) and another with t(6;11)(q15;q23) [29]. Cox et al. studied 378 adult acute leukemia cases and found that in 18 cases with rearrangements involving 11q22-25 bands by conventional karyotyping, only 11 had MLL gene involvements [30]. In CML, two cases with 11q23 rearrangements but intact MLL were previously reported. One case had t(2;11)(p21;q23) [12] and the other had t(4;11)(q21;q23) [5]. In our study, three patients showed 11q23 translocations by conventional cytogenetics but were negative for MLL involvement by FISH, including t(2;11)(q32;q23) (case 4), t(4;11)(q21;q23) (case 1), and t(6;11)(p11.2;q23) (case 2). The exact mechanisms and roles of 11q23+/MLL− in leukemia are unknown. We proposed two potential mechanisms here. First, although MLL gene appeared intact in these cases, MLL gene may be controlled and dysregulated by other promoters or enhancers as a result of translocations. Secondly, other genes at the 11q23 locus may play roles in leukemogenesis, such as FLI1 [31,32]. Future molecular study to identify the breakpoints of these 11q23+/MLL− translocations and their fusion partner is needed.

In case 11, although the patient initially presented with chronic phase CML and responded to TKI, he later developed blast transformation with 11q23 translocation. FISH for BCR-ABL1 was negative. The development of acute leukemia in Ph-negative clones is intriguing. The occurrence of chromosomal changes in Ph-negative cells is less common than in Ph-positive cells, and the significance of this phenomenon is not fully understood [33]. Jabbour et al. analyzed 258 CML patients in CP with imatinib therapy [34]. After a median follow-up of 37 months, 9% patients developed chromosomal abnormalities in Ph-negative cells, with − Y and trisomy 8 being the most common. Others included del (7), del (20), t(11;17), etc. Although similar cytogenetic changes are associated with myelodysplastic syndromes, these cytogenetic abnormalities in CML seem transient and most disappeared after a median of 5 months. Development of MDS or acute leukemia is extremely rare in Ph-negative cells. In a study of 1,701 CML patients, Kovitz found that 3 patients developed AML and MDS in Ph-negative cells [35]. Monosomy 7 and complex cytogenetic changes were the most common chromosomal abnormalities in reported cases of MDS/AML developed in Ph-negative cells [36-40]. In our case, the patient developed AML in Ph-negative cells carrying 11q23/MLL abnormality. To our best knowledge, this is the first case of acute leukemia developing in Ph-negative cells with 11q23/MLL involvement. The mechanisms behind the development of chromosomal abnormalities in Ph-negative cells are not clear, but it is less likely due to TKI therapy [34]. A plausible explanation is that t(9;22) is not a random event, but rather develops from a preceding genetic abnormality, which occurs at the level of hematopoietic stem cells. This genetic abnormality not only induces t(9;22) but also other chromosomal abnormalities. During CML treatment, cells with t(9;22) are eradicated and cells with other chromosomal abnormalities emerge [41].

As mentioned above, occasional case reports of CML with 11q23 translocations are available in the literature. The fusion partners previously reported for 11q23 in CML include 1q21 [10], 2p21 [12], 4q21 [5], 6q25 [7], 9p22 [11,8,6], 17q21 [9], 19p13.1 [14], and 19p13.3 [13]. In our study, 9p21-22/AF9 was the most common partner for 11q23. In addition, we identified several new fusion partners for 11q23 in CML: 2q32, 6p11.2, 17q25, and 18q11.2. Of these, t(11;17)(q23;q25) potentially involves the SEPT9 gene on 17q25 and had been reported in AML and myelodysplastic syndrome with a poor prognosis [42], whereas t(2;11)(q32;q23), t(6;11)(p11.2;q23), and t(11;18)(q23;q11.2) translocations have never been previously described in the literature. It will be of interest to explore and characterize the potential genes involved in these reciprocal translocations and their roles in leukemia.

In conclusion, 11q23 translocations are rare events in CML, frequently occurring in blast phase and associated with a poor prognosis. Not all 11q23 translocations involve MLL gene. We described a first case of acute leukemia developed in Ph-negative cells with 11q23/MLL involvement. Several novel genetic loci are identified as fusion partners for 11q23.