Cancer Genetics 207 (2014) 503e510
Transient presence of clonal chromosomal aberrations in Ph-negative cells in patients with chronic myeloid leukemia remaining in deep molecular response on tyrosine kinase inhibitor treatment _ Ratajczak a, Michał Gniot a,*, Krzysztof Lewandowski a, Błazej Maria Lewandowska a, Agata Lehmann-Kopydłowska a, a,b _ , Mieczysław Komarnicki a Małgorzata Jarmuz-Szymczak a
Department of Hematology, University of Medical Sciences, Szamarzewskiego, Poznan, Poland; b Institute of Human Genetics, Polish Academy of Sciences, Strzeszynska, Poznan, Poland Advancements in treatment of chronic myeloid leukemia (CML) turned this formerly fatal neoplasm into a manageable chronic condition. Therapy with tyrosine kinase inhibitors (TKIs) often leads to significant reduction of disease burden, known as the deep molecular response (DMR). Herein, we decided to analyze the cohort of CML patients treated in our center with TKIs, who obtain and retain DMR for a period longer than 24 months. The aim of the study was to evaluate the frequency of clonal cytogenetic aberrations in Philadelphia-negative (Ph) cells in patients with DMR during TKI treatment. The analyzed data was obtained during routine molecular and cytogenetic treatment monitoring, using G-banded trypsin and Giemsa stain (GTG) karyotyping and reverse transcription quantitative PCR. We noticed that approximately 50% of patients (28 of 55) in DMR had, at some follow-up point, transient changes in the karyotype of their Ph bone marrow cells. In 9.1% of cases (5 of 55), the presence of the same aberrations was observed at different time points. The most frequently appearing aberrations were monosomies of chromosomes 19, 20, 21, and Y. Statistical analysis suggests that the occurrence of such abnormalities in CML patients correlates with the TKI treatment time. Keywords CML, cytogenetic abnormalities, TKI treatment, deep molecular response ª 2014 Elsevier Inc. All rights reserved.
Chronic myeloid leukemia (CML) is a hematological neoplasm characterized by the presence of the reciprocal translocation t(9;22), resulting in two abnormal fusion genes: ABL1-BCR and BCR-ABL1 (1,2). The first gene seems to have no biological activity, and the second gene, which contains the entire catalytic domain of ABL1, was proven to be responsible for the development of the disease. A constant, high catalytic activity of BCR-ABL1 tyrosine kinase causes increased cell proliferation, reduced adhesion of
Received June 12, 2014; received in revised form October 6, 2014; accepted October 7, 2014. * Corresponding author. E-mail address: [email protected] 2210-7762/$ - see front matter ª 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cancergen.2014.10.003
leukemic cells to bone marrow stroma, and is responsible, at least in part, for the genomic instability of malignant cells. During the natural outcome of the disease, if not treated specifically, it evolves from chronic to advanced phases: acceleration and blast crisis. It results in relatively short overall survival time, not exceedingdaccording to the historical datad3e6 years. When tyrosine kinase inhibitors (TKIs) were introduced in 1999, this formerly lethal disease became, in the vast majority of patients, a manageable chronic condition (3,4). Introduction of imatinib (IM), the first available BCR-ABL1 inhibitor, into CML treatment resulted in significant prognosis improvement. In the IRIS study, the overall survival rate, excluding non-CML-related deaths after 8 years of follow-up reached 93% (5). Further improvement was noted when
504 second-generation TKIs became available (6,7). Administration of nilotinib or dasatinib as the first-line treatment resulted in a higher response rate, as well as faster and deeper cytogenetic and molecular responses. Also, the toxicity profiles of these drugs were found to be acceptable and manageable (8,9). Although most of the patients respond very well to the therapy, TKI resistance was observed in a small proportion of them. This resistance may be related to the presence of BCR-ABL1 tyrosine kinase domain (KD) mutations (10) and/ or proliferation signal transduction through BCR-ABL1 independent cell signaling pathways (11). TKI resistance can be associated with the appearance of other clonal chromosomal aberrations (CCA) in Philadelphia chromosomeepositive (Phþ) leukemic cells. Their presence has been correlated with poor disease outcome (12). Currently, according to European LeukemiaNet (ELN), the clonal evolution of Phþ cells during the TKI treatment is one of the criteria of treatment failure (13). The clinical significance of CCA acquisition by Phnegative cells (CCA/Ph) is unknown. In the ELN 2013 guidelines, their presence is considered a warning sign, whereas the NCCN guidelines do not mention them at all (14). Since stopping the therapy is currently widely discussed, this issue is becoming very important, especially in patients with deep molecular response (DMR), who are candidates for clinical trials with TKI discontinuation. Therefore, in this study, a decision was made to analyze the presence of cytogenetic abnormalities in the Ph hematopoietic cells in our cohort of patients who reached at least MR4 during TKI treatment.
Materials and methods Patients Of the 146 CML patients treated with TKIs in our center, 65 met the criteria of DMR, which was defined as at least a 4 log reduction of the BCR-ABL1 transcript level in two consecutive measurements (MR4). Seven patients had to be excluded from final evaluation due to incomplete laboratory and/or clinical data. In another three cases, only fluorescence in situ hybridization (FISH) data were available due to a lack of metaphases in the myeloid cell culture. Therefore, those patients were also excluded from analysis. The group included in the analysis consisted of 55 patients (median age of 52 years: 57.5 for women and 50 for men). All of them remained in DMR for >12 months (12e76, median of 41 months). In 45 of these patients, DMR was obtained during treatment with imatinib. In 10 patients, responses of at least MR4 were obtained on second-line treatment with dasatinib or nilotinib. In six cases, the reason for a change in the IM therapy was acquired drug resistance (in four patients, due to the presence of BCR-ABL1 KD mutations; in two additional patients, due to non-mutational resistance). Imatinib intolerance was diagnosed in four additional patients. None of the analyzed patients had any cytogenetic abnormalities other than the t(9;22) at the time of diagnosis, except one patient with a constitutive t(1;10). Detailed characteristics of the study group are presented in Table 1. During the observation period, none of the patients from the entire study
M. Gniot et al. Table 1 The characteristics of CML patients remaining in DMR during TKI treatment
Imatinib Nilotinib Dasatinib Total TKI median treatment time (mo) Age (y), median Male/female ratio
Patients in DMR
Patients with CCAs
Patients without CCAs
45 5 5 55 62
21 4 3 28 93
24 1 2 27 43
52 21/34
52 14/14
59 7/20
group progressed to acceleration phase/blast crisis or lost complete cytogenetic response.
Methods Reverse transcription quantitative PCR (RQ-PCR) was performed every 3 months until major molecular response (MMR) was achieved, and then every 3 months subsequently. Chromosome banding analysis of marrow cell metaphases was done at 3, 6, and 12 months until complete cytogenetic remission (CCyR) was reached, and then every 12 months subsequently. Molecular analysis was performed using RQ-PCR according to the ELN protocol (15e17), and was validated in cooperation with ELN (18). The RQ-PCR method was also independently validated at the Institute of Medical and Veterinary Sciences in Adelaide, Australiadone of three main international reference laboratories for CML monitoring. For the purpose of the study, the terms of MR4, MR4.5, and MR5 were used to define the level of DMR. Complete molecular response (CMR) was diagnosed when RQ-PCR and nested-PCR for the BCR-ABL1 transcript (with sensitivity of quantitative PCR from the same sample of at least 105 below baseline) were negative. A classic cytogenetic analysis was performed on bone marrow cells that were stimulated with granulocyte colonystimulating factor (GCSF) for 48 hours. The chromosomes were G-banded with trypsin and Giemsa stain (GTG) using standard methods. The karyotype was described according to the ISCN 2009 guidelines (19). In select cases, the FISH technique with commercially available probes (Vysis BCL 6 (3q27), AML/ETO (21q22, 8q22), CBFB (16q22), TCF3/PBX1 (19p13.3, 1q23); Abbott Molecular, Des Plaines, IL) was used to confirm the presence of numerical aberrations in interphase cells and/or metaphases (Figure 1). FISH studies were performed according to standard protocols, and at least 400 nuclei were counted. Statistical analysis was performed using Statistica 8.0 (StatSoft, Tulsa, OK). Appropriate parametric and nonparametric tests were used, according to the distribution of the variables. The frequency comparison between groups with and without CCA was performed using the Fisher exact test. The non-parametric Mann-Whitney U test was used to compare TKI treatment time in subgroups of patients with and without CCA in Ph cells.
Ph-negative aberrations in CML
505
Figure 1 Case of imatinib-treated patient with mononomy of chromosome 3, 8, 19, and 21 in Ph-negative cells (composite karyotype) in consecutive evaluations, which were performed 90 months after diagnosis and treatment initiation (Table 3, patient 21). (A) G-banded metaphase from bone marrow cells with monosomy of chromosome 3. (B) Interphase FISH with AML1/ETO dual-color, dual-fusion translocation probe: a cell with deletion of the AML1 and ETO genes (one red and one green signal) and a normal cell (two red and two green signals). (C) Interphase FISH with BCL6 probe: a cell with deletion of BCL6 (one yellow signal) and a normal cell (two yellow signals). (D) Representative G-banded karyogram showing the monosomy of chromosome 3.
Results In 28 (50.9%) of 55 patients with DMR, the CCAs in Ph clones were detected during the observation time. In 5 (9.1%) of 55 patients, reappearance of the same aberrations was observed at different time points; however, in most cases, the chromosomal abnormalities were detected only
once during the observation period. CCAs seem to occur independently of the depth of molecular response (Table 2). The most frequently appearing aberrations were losses of chromosome 19, 20, 21, and Y (Table 3). None of the patients analyzed had myelodysplastic changes in the cytomorphological evaluation of bone marrow cells. However, in five patients, despite the continuous DMR, persistent
506 Table 2
M. Gniot et al. The frequency of CCAs in Ph myeloid cells in CML patients during TKI treatment with different levels of MRa
Current level of MR
Total patients
Patients with CCA
Patients without CCA
Patients with CCA, individual
Patients with CCA, cumulative
CMR MR5 MR4.5 MR4 Total
13 17 11 14 55
5 8 8 7 28
8 9 3 7 27
38.46% 47.06% 72.73% 50.00% 50.91%
5/13 13/30 21/41 28/55
a
(38.46%) (43.33%) (51.21%) (50.91%)
Classical cytogenetic analysis of bone marrow cells stimulated with GCSF for 48 hours.
granulocytopenia (grade 1 Common Terminology Criteria for Adverse Effects) during TKI treatment was observed. Cytogenetic evaluation of those five patients confirmed the presence of a large spectrum of different aberrations, including the monosomies of chromosomes 16, 20, 21, and Y, as well as the monosomies of chromosomes 3, 8, 19, and 21, which were described as a composite karyotype for one patient (Figure 1). Moreover, the aberrant metaphases never exceeded 25%. The FISH study confirmed the monosomies of chromosomes 3 and 8 (patient 21), chromosome 16 (patient 12), chromosome 19 (patient 23) and chromosome 21 (patients 7 and 20). The potential factors that could have been associated with the presence of CCA/Ph were analyzed. The t-test showed no differences in patients’ age distribution in groups with and without CCA (P Z 0.4036). Among the patients with the presence or absence of CCA in Ph cells, a significant variation in TKI treatment time was noticed. While a median treatment time in the entire study group was 62 months, in subgroups of patients with and without CCA, it was 93 and 43 months, respectively. The difference turned out to be statistically significant (P Z 0.0011, Mann-Whitney U test, Figure 2). However, it must be mentioned that when the same analysis was performed only in patients who were treated with second-generation TKIs, the difference between the occurrence or non-occurrence of CCA was found to be insignificant (P Z 0.1625, Fisher exact test). Moreover, no correlation between interferon-alpha (INFa) or hydroxyurea pre-treatment and the occurrence of CCA/Ph was found.
Discussion At the time of diagnosis, in 90e95% of patients with CML, the presence of the typical t(9;22)(q34;q11) is found in erythroid, granulocytic, monocytic, megakaryocytic, and precursor Tand B-lymphoid cells. Detailed genomic studies confirmed that in approximately 5% of CML patients in the chronic phase, a variant translocation affecting another chromosome(s) is present. Moreover, in 20% of CML patients, additional CCA in Phþ cells, most frequently including Y, þ8, or an additional copy of chromosome 22 are present. The frequency of CCA/Phþ is higher in advanced stages of the disease (approximately 80% of patients in the blastic transformation phase). The most typical cytogenetic changes in advanced disease phases are þ8 (33%), an additional copy of the Ph chromosome (30%), and abnormalities of chromosome 17 (20%). In our study group, none of the
patients had either CCA/Phþ or CCA/Ph at the time of diagnosis. The frequency of CCA/Ph before initiation of the treatment is not fully recognized. Bozkurt et al. confirmed the presence of CCA/Ph in 4.8% of untreated patients with CML (20). Despite excellent results of the TKI treatment, some CML patients with prolonged cytogenetic remissions acquired CCAs in the Ph bone marrow cells. The most commonly reported CCA/Ph abnormalities include þ8, 7, 20q, 5, and Y (21e29). Unfortunately, until now, no precise data has been available regarding the frequency of CCA/Ph in patients who obtained pronounced reduction of the BCR-ABL1 transcript level during TKI treatment. The origin of molecular and cytogenetic abnormalities in CML is not fully understood. It has been suggested that cancer development depends on continuous acquisition of genetic variations by random mutations, with environmental selection pressure acting on the phenotype diversity (30). Therefore, it cannot be excluded that the genomic instability of CML stem cells leads not only to the appearance of the t(9;22), but also other aberrations in Phþ and Ph stem cells. Herens et al. were among the first to report, in 2003, the Ph aberrations in CML (31). The authors suggested that the aberrant clones were present before the treatment and that eradication of the Phþ population allowed them to emerge. However, they were unable to detect any additional aberrations in the pretreatment material. The polyclonal nature of CML was also postulated by Abruzzese et al. (32). Cytogenetic abnormalities in Ph clones may have a similar nature, as reported in the case of BCR-ABL1 mutations in Ph cells in TKI-resistant CML patients. Next generation sequencing and cloning studies have confirmed that during TKI treatment, sequential, branching, and parallel routes to compound BCR-ABL1 mutations are common, including low-level defects in independent clones. Therefore, it is possible that the same phenomenon that reflects the genomic instability of CML stem cells plays a role in the case of Ph stem cells (33). The clinical significance of the CCA/Ph appearance in patients in CCyR and DMR during TKI treatment is currently not clear. The only sure conclusion is that chromosome 7 monosomy (or del 7q) is associated with a high risk of CML transformation to secondary acute myeloid leukemia (AML) or, less frequently, to myelodysplastic syndrome (MDS) (27,34). Moreover, presence of CCA/Ph does not seem to have an obvious effect on the prognosis (27,28,35). In our study group of five patients who harbored CCA/Ph during TKI treatment, prolonged moderate granulocytopenia was observed. It was, however, not associated with the presence of MDS-like morphological changes in the bone marrow cells on cytomorphological evaluation. Moreover, during the time
Ph-negative aberrations in CML Table 3
Patient no. b
1 2 3 4b 5
6 7b
8 9b 10 11 12b
13 14b 15b 16b 17 18 19 20 21 22 23
24 25 26
507
The CCA/Ph cells in CML patients who obtained DMR during TKI treatmenta Treatment time to aberration (mo)
Overall treatment time (mo)
Molecular response level
78 6 42 78
112 34 54 93
CMR CMR CMR CMR
IM, DASA IM IM IM
60 72
100
CMR CMR
IM, NILO
32 40
63 107
MR5 MMR
45,XX,21[3]/46,XX[28] 45,XX,19[4]/46,XX[24] 45,X,Y[3]/46,XY[27] 45,XX,20[3]/46,XX[27] 47,XX,þ4[2]/46,XX[22] 45,XX,16[3]/46,XX[24]
70 36 80 84 54 6
63 101 107 149 38
MR5 MR5 MR5 MR5 MR5 MMR
42w45,X,X[3],3[4],11[5],16[4], 20[5],21[5], 22[6][cp14]/46, XX[15] 45,XY,7[3]/45,XY,19[3]/46,XY[21] 45,XY,17[3]/46,XY[25] 45,XY,19[4]/46,XY[21] 45,XY,12[3]/46,XY[25] //45,XY,21[3]/46,XY[22] //45,XY,19[3]/46,XY[23] 45,XX,22[4]/46,XX[22] 45,XY,21[3]/46,XY[24] 45,XY,22[3]/46,XY[21] 44w46,XY,12[4],19[3], 21[5] [cp9]/46,XY[19] 44w45,XX,3[3],8[3],19[5],21[3] [cp9]/46,XX[15] 45,X,X[3]/46,XX[20]
24
Karyotype (no. of analyzed metaphases) 45,XY,20[4]/45,XY,22[3]/46,XY[19] 45,XX,8[3]/46,XX[25] 45,XY,21[3]/46,XY[25] 44w46,XY,20[5],22[5][cp8]/46, XY[22] 45,XX,20[3]/46,XX[21] 43w45,X,X[3],21[5],22[5][cp9]/ 46,XX[16] 45,XY,20[4]/46,XY[21] 45,XX,21 [2]/45,XX,22 [1]/46, XX [22]
Coexisting abnormalities
Granulocytopenia
Granulocytopenia Granulocytopenia
Treatment
IM, DASA IM
IM IM, NILO IM IM IM
MR5
95 80 116 60 33 46 122 74 86 18
111 106 147 101 62
42
MR5 MR4,5 MR4,5 MR4,5 MR4.5 CMR MR4.5 MR4.5 MR5 MR4.5
90
115
MR4.5
Granulocytopenia
IM
6
34
MR4
IM, NILO
46,XX,t(1;10)(q21;q22),19,21[1]/ 46,XX,t(1;10)(q21;q22),19[2]/ 46,XX,t(1;10),22[3]/ 46,XX,t(1;10)(q21;q22)[24] 46,XX,t(1;10)(q21;q22),16[3]/ 45,XX,t(1;10) (q21;q22),19[3]/ 46,XX,t(1;10)(q21;q22) [24]
42
76
MMR
Other abnormalities in single metaphases Constitutive t(1;10)
48
MMR
45,XX,t(1;10)(q21;q22),19[3]/ 46,XX,t(1;10)(q21;q22) [19] 45,XY,16[3]/46,XY[22] 47,XX,þ8[1]/45,XX,20[3]/46,XX[23]
54
MR4
24 30
61 66
MR4 MMR
45,XX,5[4]/46,XX[35] 45,X,Y[4]/46,XY[26] 45,X,Y[3]/46,XY[37]
54 60 76
93
MR4 MR4 MR4
149 117
IM IM IM IM, NILO BMT, IM
CLL
IM IM IM
Granulocytopenia
IM, DASA
IM IM IM
(Continued on the next page)
508 Table 3
M. Gniot et al. (Continued )
Patient no.
Karyotype (no. of analyzed metaphases)
27
45,X,Y[4]/46,XY[21]
28
46,XY,21[3]/46,XY[23] 45,X,Y[3]/46,XY[17] 45,X,X[3]/45,XX,19[3]/46,XX[38]
Treatment time to aberration (mo) 6 24 48 108
Overall treatment time (mo)
Molecular response level
Coexisting abnormalities
Treatment
62
MMR
IM
54
MR4 CMR MR4
IM
Abbreviations: MMR, major molecular response; CMR, complete molecular response; CLL, chronic lymphocytic leukemia; IM, imatinib; DASA, dasatinib; NILO, nilotinib; BMT, bone marrow transplantation; MR4, MR4.5, MR5, reduction by 4, 4.5, and 5 logs according to the International Scale. a The shaded fields represent the supplementary data, which were excluded from statistical analysis due to the fact that the observations were made before reaching DMR and/or did not meet the ISCN criteria. These fields stress the recurrent/stable character of CCA/Ph found in further analyses. b Patient treated with interferon alpha before TKI therapy initiation.
of follow-up (median 62 months), not a single TKI-treated CML patient had developed a secondary Ph myeloid neoplasm, nor did they progress to a more advanced phase of the disease. However, in one patient from the CCA/Ph group, despite continuous MR4, the symptoms of chronic lymphocytic leukemia (CLL) occurred after 9 years of treatment with imatinib. The FISH analysis revealed that the patient had a population of cells with a split ATM gene signal (Table 3, patient 16). The study of the retrospective samples showed that this aberration was present for at least 4 years before CLL manifestation. Numerous precautions were introduced to minimize the potential bias in this study. In our study group, the most common CCA in Ph clones were chromosome 19, 20, 21 and Y monosomies. It is known that small chromosomes are more prone to loss during processing of the material. To avoid reporting of artificial aberrations related to laboratory
Figure 2 Median tyrosine kinase inhibitor treatment time in a group of CML patients with transient appearance of Ph-negative CCA in cytogenetic evaluation of cultured and GCSF stimulated bone marrow cells, using GTG techniques.
procedures, the ISCN guidelines were introduced (19). According to these guidelines, the monosomies found in more than three metaphases should be reported in the karyotype. In light of these guidelines, the over-reporting of the small chromosomes’ monosomies should not be an issue. In the three karyotypes presented in Table 3, changes were detected in lower than the normally accepted number of metaphases; however, those results were marked as supplementary data because the patients were not in DMR by that time. We decided to include them in the table only because these were the same aberrations that occurred in further analyses of the same patients, and they suggest the recurrent character of the observed changes. Some controversies may surround the Y monosomies. Until today, they were reported as age-related and affecting mainly elderly males. The loss of chromosome Y has been reported to occur in approximately 10% of patients with various hematological disorders (36). Recently, data that showed the strong correlation between the presence of Y monosomy in a fraction of bone marrow cells and cancer risk was published (37). Moreover, in CML patients, the loss of the Y chromosome in Phþ metaphases has been proven to have a similar prognostic impact as any other CCA (38). Therefore, in our opinion, relatively high-frequency Y monosomies (3 of 21 male patients; 14.1%) should be taken into consideration. In the majority of our investigated CML cases, CCA/Ph was detected incidentally during routine cytogenetic evaluation. CCA/Ph reoccurred in only five cases (Table 3). A similar phenomenon was reported by Kim et al. (35). They confirmed the transient presence of trisomy of chromosome 8 in Ph cells, and, in this study, no association between the occurrence of CCA in Ph cells and transformation to AML or MDS was found. It should be noted, however, that trisomy of chromosome 8 is one of the most frequent Ph aberrations that emerges during imatinib therapy in CML patients (27,32). The question remains: Was the sensitivity of the cytogenetic test used in diagnostic work-up sufficient? Recently, it was documented that the genome of CML patients harbors many cytogenetic changes. These might be hidden in subpopulations due to clonal events or due to their extremely complex characteristics. The use of several cytogenetic and molecular techniques (multicolor-FISH, comparative genomic
Ph-negative aberrations in CML hybridization (CGH) array) has been postulated to identify “cryptic” cytogenetic aberrations. However, this strategy is often questioned because it is time-consuming and costly, and cannot be converted to benefit the patients (39). Another problem is that the cytogenetic aberrations we observed are present in
Transient presence of clonal chromosomal aberrations in Ph-negative cells in patients with chronic myeloid leukemia remaining in deep molecular response on tyrosine kinase inhibitor treatment _ Ratajczak a, Michał Gniot a,*, Krzysztof Lewandowski a, Błazej Maria Lewandowska a, Agata Lehmann-Kopydłowska a, a,b _ , Mieczysław Komarnicki a Małgorzata Jarmuz-Szymczak a
Department of Hematology, University of Medical Sciences, Szamarzewskiego, Poznan, Poland; b Institute of Human Genetics, Polish Academy of Sciences, Strzeszynska, Poznan, Poland Advancements in treatment of chronic myeloid leukemia (CML) turned this formerly fatal neoplasm into a manageable chronic condition. Therapy with tyrosine kinase inhibitors (TKIs) often leads to significant reduction of disease burden, known as the deep molecular response (DMR). Herein, we decided to analyze the cohort of CML patients treated in our center with TKIs, who obtain and retain DMR for a period longer than 24 months. The aim of the study was to evaluate the frequency of clonal cytogenetic aberrations in Philadelphia-negative (Ph) cells in patients with DMR during TKI treatment. The analyzed data was obtained during routine molecular and cytogenetic treatment monitoring, using G-banded trypsin and Giemsa stain (GTG) karyotyping and reverse transcription quantitative PCR. We noticed that approximately 50% of patients (28 of 55) in DMR had, at some follow-up point, transient changes in the karyotype of their Ph bone marrow cells. In 9.1% of cases (5 of 55), the presence of the same aberrations was observed at different time points. The most frequently appearing aberrations were monosomies of chromosomes 19, 20, 21, and Y. Statistical analysis suggests that the occurrence of such abnormalities in CML patients correlates with the TKI treatment time. Keywords CML, cytogenetic abnormalities, TKI treatment, deep molecular response ª 2014 Elsevier Inc. All rights reserved.
Chronic myeloid leukemia (CML) is a hematological neoplasm characterized by the presence of the reciprocal translocation t(9;22), resulting in two abnormal fusion genes: ABL1-BCR and BCR-ABL1 (1,2). The first gene seems to have no biological activity, and the second gene, which contains the entire catalytic domain of ABL1, was proven to be responsible for the development of the disease. A constant, high catalytic activity of BCR-ABL1 tyrosine kinase causes increased cell proliferation, reduced adhesion of
Received June 12, 2014; received in revised form October 6, 2014; accepted October 7, 2014. * Corresponding author. E-mail address: [email protected] 2210-7762/$ - see front matter ª 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cancergen.2014.10.003
leukemic cells to bone marrow stroma, and is responsible, at least in part, for the genomic instability of malignant cells. During the natural outcome of the disease, if not treated specifically, it evolves from chronic to advanced phases: acceleration and blast crisis. It results in relatively short overall survival time, not exceedingdaccording to the historical datad3e6 years. When tyrosine kinase inhibitors (TKIs) were introduced in 1999, this formerly lethal disease became, in the vast majority of patients, a manageable chronic condition (3,4). Introduction of imatinib (IM), the first available BCR-ABL1 inhibitor, into CML treatment resulted in significant prognosis improvement. In the IRIS study, the overall survival rate, excluding non-CML-related deaths after 8 years of follow-up reached 93% (5). Further improvement was noted when
504 second-generation TKIs became available (6,7). Administration of nilotinib or dasatinib as the first-line treatment resulted in a higher response rate, as well as faster and deeper cytogenetic and molecular responses. Also, the toxicity profiles of these drugs were found to be acceptable and manageable (8,9). Although most of the patients respond very well to the therapy, TKI resistance was observed in a small proportion of them. This resistance may be related to the presence of BCR-ABL1 tyrosine kinase domain (KD) mutations (10) and/ or proliferation signal transduction through BCR-ABL1 independent cell signaling pathways (11). TKI resistance can be associated with the appearance of other clonal chromosomal aberrations (CCA) in Philadelphia chromosomeepositive (Phþ) leukemic cells. Their presence has been correlated with poor disease outcome (12). Currently, according to European LeukemiaNet (ELN), the clonal evolution of Phþ cells during the TKI treatment is one of the criteria of treatment failure (13). The clinical significance of CCA acquisition by Phnegative cells (CCA/Ph) is unknown. In the ELN 2013 guidelines, their presence is considered a warning sign, whereas the NCCN guidelines do not mention them at all (14). Since stopping the therapy is currently widely discussed, this issue is becoming very important, especially in patients with deep molecular response (DMR), who are candidates for clinical trials with TKI discontinuation. Therefore, in this study, a decision was made to analyze the presence of cytogenetic abnormalities in the Ph hematopoietic cells in our cohort of patients who reached at least MR4 during TKI treatment.
Materials and methods Patients Of the 146 CML patients treated with TKIs in our center, 65 met the criteria of DMR, which was defined as at least a 4 log reduction of the BCR-ABL1 transcript level in two consecutive measurements (MR4). Seven patients had to be excluded from final evaluation due to incomplete laboratory and/or clinical data. In another three cases, only fluorescence in situ hybridization (FISH) data were available due to a lack of metaphases in the myeloid cell culture. Therefore, those patients were also excluded from analysis. The group included in the analysis consisted of 55 patients (median age of 52 years: 57.5 for women and 50 for men). All of them remained in DMR for >12 months (12e76, median of 41 months). In 45 of these patients, DMR was obtained during treatment with imatinib. In 10 patients, responses of at least MR4 were obtained on second-line treatment with dasatinib or nilotinib. In six cases, the reason for a change in the IM therapy was acquired drug resistance (in four patients, due to the presence of BCR-ABL1 KD mutations; in two additional patients, due to non-mutational resistance). Imatinib intolerance was diagnosed in four additional patients. None of the analyzed patients had any cytogenetic abnormalities other than the t(9;22) at the time of diagnosis, except one patient with a constitutive t(1;10). Detailed characteristics of the study group are presented in Table 1. During the observation period, none of the patients from the entire study
M. Gniot et al. Table 1 The characteristics of CML patients remaining in DMR during TKI treatment
Imatinib Nilotinib Dasatinib Total TKI median treatment time (mo) Age (y), median Male/female ratio
Patients in DMR
Patients with CCAs
Patients without CCAs
45 5 5 55 62
21 4 3 28 93
24 1 2 27 43
52 21/34
52 14/14
59 7/20
group progressed to acceleration phase/blast crisis or lost complete cytogenetic response.
Methods Reverse transcription quantitative PCR (RQ-PCR) was performed every 3 months until major molecular response (MMR) was achieved, and then every 3 months subsequently. Chromosome banding analysis of marrow cell metaphases was done at 3, 6, and 12 months until complete cytogenetic remission (CCyR) was reached, and then every 12 months subsequently. Molecular analysis was performed using RQ-PCR according to the ELN protocol (15e17), and was validated in cooperation with ELN (18). The RQ-PCR method was also independently validated at the Institute of Medical and Veterinary Sciences in Adelaide, Australiadone of three main international reference laboratories for CML monitoring. For the purpose of the study, the terms of MR4, MR4.5, and MR5 were used to define the level of DMR. Complete molecular response (CMR) was diagnosed when RQ-PCR and nested-PCR for the BCR-ABL1 transcript (with sensitivity of quantitative PCR from the same sample of at least 105 below baseline) were negative. A classic cytogenetic analysis was performed on bone marrow cells that were stimulated with granulocyte colonystimulating factor (GCSF) for 48 hours. The chromosomes were G-banded with trypsin and Giemsa stain (GTG) using standard methods. The karyotype was described according to the ISCN 2009 guidelines (19). In select cases, the FISH technique with commercially available probes (Vysis BCL 6 (3q27), AML/ETO (21q22, 8q22), CBFB (16q22), TCF3/PBX1 (19p13.3, 1q23); Abbott Molecular, Des Plaines, IL) was used to confirm the presence of numerical aberrations in interphase cells and/or metaphases (Figure 1). FISH studies were performed according to standard protocols, and at least 400 nuclei were counted. Statistical analysis was performed using Statistica 8.0 (StatSoft, Tulsa, OK). Appropriate parametric and nonparametric tests were used, according to the distribution of the variables. The frequency comparison between groups with and without CCA was performed using the Fisher exact test. The non-parametric Mann-Whitney U test was used to compare TKI treatment time in subgroups of patients with and without CCA in Ph cells.
Ph-negative aberrations in CML
505
Figure 1 Case of imatinib-treated patient with mononomy of chromosome 3, 8, 19, and 21 in Ph-negative cells (composite karyotype) in consecutive evaluations, which were performed 90 months after diagnosis and treatment initiation (Table 3, patient 21). (A) G-banded metaphase from bone marrow cells with monosomy of chromosome 3. (B) Interphase FISH with AML1/ETO dual-color, dual-fusion translocation probe: a cell with deletion of the AML1 and ETO genes (one red and one green signal) and a normal cell (two red and two green signals). (C) Interphase FISH with BCL6 probe: a cell with deletion of BCL6 (one yellow signal) and a normal cell (two yellow signals). (D) Representative G-banded karyogram showing the monosomy of chromosome 3.
Results In 28 (50.9%) of 55 patients with DMR, the CCAs in Ph clones were detected during the observation time. In 5 (9.1%) of 55 patients, reappearance of the same aberrations was observed at different time points; however, in most cases, the chromosomal abnormalities were detected only
once during the observation period. CCAs seem to occur independently of the depth of molecular response (Table 2). The most frequently appearing aberrations were losses of chromosome 19, 20, 21, and Y (Table 3). None of the patients analyzed had myelodysplastic changes in the cytomorphological evaluation of bone marrow cells. However, in five patients, despite the continuous DMR, persistent
506 Table 2
M. Gniot et al. The frequency of CCAs in Ph myeloid cells in CML patients during TKI treatment with different levels of MRa
Current level of MR
Total patients
Patients with CCA
Patients without CCA
Patients with CCA, individual
Patients with CCA, cumulative
CMR MR5 MR4.5 MR4 Total
13 17 11 14 55
5 8 8 7 28
8 9 3 7 27
38.46% 47.06% 72.73% 50.00% 50.91%
5/13 13/30 21/41 28/55
a
(38.46%) (43.33%) (51.21%) (50.91%)
Classical cytogenetic analysis of bone marrow cells stimulated with GCSF for 48 hours.
granulocytopenia (grade 1 Common Terminology Criteria for Adverse Effects) during TKI treatment was observed. Cytogenetic evaluation of those five patients confirmed the presence of a large spectrum of different aberrations, including the monosomies of chromosomes 16, 20, 21, and Y, as well as the monosomies of chromosomes 3, 8, 19, and 21, which were described as a composite karyotype for one patient (Figure 1). Moreover, the aberrant metaphases never exceeded 25%. The FISH study confirmed the monosomies of chromosomes 3 and 8 (patient 21), chromosome 16 (patient 12), chromosome 19 (patient 23) and chromosome 21 (patients 7 and 20). The potential factors that could have been associated with the presence of CCA/Ph were analyzed. The t-test showed no differences in patients’ age distribution in groups with and without CCA (P Z 0.4036). Among the patients with the presence or absence of CCA in Ph cells, a significant variation in TKI treatment time was noticed. While a median treatment time in the entire study group was 62 months, in subgroups of patients with and without CCA, it was 93 and 43 months, respectively. The difference turned out to be statistically significant (P Z 0.0011, Mann-Whitney U test, Figure 2). However, it must be mentioned that when the same analysis was performed only in patients who were treated with second-generation TKIs, the difference between the occurrence or non-occurrence of CCA was found to be insignificant (P Z 0.1625, Fisher exact test). Moreover, no correlation between interferon-alpha (INFa) or hydroxyurea pre-treatment and the occurrence of CCA/Ph was found.
Discussion At the time of diagnosis, in 90e95% of patients with CML, the presence of the typical t(9;22)(q34;q11) is found in erythroid, granulocytic, monocytic, megakaryocytic, and precursor Tand B-lymphoid cells. Detailed genomic studies confirmed that in approximately 5% of CML patients in the chronic phase, a variant translocation affecting another chromosome(s) is present. Moreover, in 20% of CML patients, additional CCA in Phþ cells, most frequently including Y, þ8, or an additional copy of chromosome 22 are present. The frequency of CCA/Phþ is higher in advanced stages of the disease (approximately 80% of patients in the blastic transformation phase). The most typical cytogenetic changes in advanced disease phases are þ8 (33%), an additional copy of the Ph chromosome (30%), and abnormalities of chromosome 17 (20%). In our study group, none of the
patients had either CCA/Phþ or CCA/Ph at the time of diagnosis. The frequency of CCA/Ph before initiation of the treatment is not fully recognized. Bozkurt et al. confirmed the presence of CCA/Ph in 4.8% of untreated patients with CML (20). Despite excellent results of the TKI treatment, some CML patients with prolonged cytogenetic remissions acquired CCAs in the Ph bone marrow cells. The most commonly reported CCA/Ph abnormalities include þ8, 7, 20q, 5, and Y (21e29). Unfortunately, until now, no precise data has been available regarding the frequency of CCA/Ph in patients who obtained pronounced reduction of the BCR-ABL1 transcript level during TKI treatment. The origin of molecular and cytogenetic abnormalities in CML is not fully understood. It has been suggested that cancer development depends on continuous acquisition of genetic variations by random mutations, with environmental selection pressure acting on the phenotype diversity (30). Therefore, it cannot be excluded that the genomic instability of CML stem cells leads not only to the appearance of the t(9;22), but also other aberrations in Phþ and Ph stem cells. Herens et al. were among the first to report, in 2003, the Ph aberrations in CML (31). The authors suggested that the aberrant clones were present before the treatment and that eradication of the Phþ population allowed them to emerge. However, they were unable to detect any additional aberrations in the pretreatment material. The polyclonal nature of CML was also postulated by Abruzzese et al. (32). Cytogenetic abnormalities in Ph clones may have a similar nature, as reported in the case of BCR-ABL1 mutations in Ph cells in TKI-resistant CML patients. Next generation sequencing and cloning studies have confirmed that during TKI treatment, sequential, branching, and parallel routes to compound BCR-ABL1 mutations are common, including low-level defects in independent clones. Therefore, it is possible that the same phenomenon that reflects the genomic instability of CML stem cells plays a role in the case of Ph stem cells (33). The clinical significance of the CCA/Ph appearance in patients in CCyR and DMR during TKI treatment is currently not clear. The only sure conclusion is that chromosome 7 monosomy (or del 7q) is associated with a high risk of CML transformation to secondary acute myeloid leukemia (AML) or, less frequently, to myelodysplastic syndrome (MDS) (27,34). Moreover, presence of CCA/Ph does not seem to have an obvious effect on the prognosis (27,28,35). In our study group of five patients who harbored CCA/Ph during TKI treatment, prolonged moderate granulocytopenia was observed. It was, however, not associated with the presence of MDS-like morphological changes in the bone marrow cells on cytomorphological evaluation. Moreover, during the time
Ph-negative aberrations in CML Table 3
Patient no. b
1 2 3 4b 5
6 7b
8 9b 10 11 12b
13 14b 15b 16b 17 18 19 20 21 22 23
24 25 26
507
The CCA/Ph cells in CML patients who obtained DMR during TKI treatmenta Treatment time to aberration (mo)
Overall treatment time (mo)
Molecular response level
78 6 42 78
112 34 54 93
CMR CMR CMR CMR
IM, DASA IM IM IM
60 72
100
CMR CMR
IM, NILO
32 40
63 107
MR5 MMR
45,XX,21[3]/46,XX[28] 45,XX,19[4]/46,XX[24] 45,X,Y[3]/46,XY[27] 45,XX,20[3]/46,XX[27] 47,XX,þ4[2]/46,XX[22] 45,XX,16[3]/46,XX[24]
70 36 80 84 54 6
63 101 107 149 38
MR5 MR5 MR5 MR5 MR5 MMR
42w45,X,X[3],3[4],11[5],16[4], 20[5],21[5], 22[6][cp14]/46, XX[15] 45,XY,7[3]/45,XY,19[3]/46,XY[21] 45,XY,17[3]/46,XY[25] 45,XY,19[4]/46,XY[21] 45,XY,12[3]/46,XY[25] //45,XY,21[3]/46,XY[22] //45,XY,19[3]/46,XY[23] 45,XX,22[4]/46,XX[22] 45,XY,21[3]/46,XY[24] 45,XY,22[3]/46,XY[21] 44w46,XY,12[4],19[3], 21[5] [cp9]/46,XY[19] 44w45,XX,3[3],8[3],19[5],21[3] [cp9]/46,XX[15] 45,X,X[3]/46,XX[20]
24
Karyotype (no. of analyzed metaphases) 45,XY,20[4]/45,XY,22[3]/46,XY[19] 45,XX,8[3]/46,XX[25] 45,XY,21[3]/46,XY[25] 44w46,XY,20[5],22[5][cp8]/46, XY[22] 45,XX,20[3]/46,XX[21] 43w45,X,X[3],21[5],22[5][cp9]/ 46,XX[16] 45,XY,20[4]/46,XY[21] 45,XX,21 [2]/45,XX,22 [1]/46, XX [22]
Coexisting abnormalities
Granulocytopenia
Granulocytopenia Granulocytopenia
Treatment
IM, DASA IM
IM IM, NILO IM IM IM
MR5
95 80 116 60 33 46 122 74 86 18
111 106 147 101 62
42
MR5 MR4,5 MR4,5 MR4,5 MR4.5 CMR MR4.5 MR4.5 MR5 MR4.5
90
115
MR4.5
Granulocytopenia
IM
6
34
MR4
IM, NILO
46,XX,t(1;10)(q21;q22),19,21[1]/ 46,XX,t(1;10)(q21;q22),19[2]/ 46,XX,t(1;10),22[3]/ 46,XX,t(1;10)(q21;q22)[24] 46,XX,t(1;10)(q21;q22),16[3]/ 45,XX,t(1;10) (q21;q22),19[3]/ 46,XX,t(1;10)(q21;q22) [24]
42
76
MMR
Other abnormalities in single metaphases Constitutive t(1;10)
48
MMR
45,XX,t(1;10)(q21;q22),19[3]/ 46,XX,t(1;10)(q21;q22) [19] 45,XY,16[3]/46,XY[22] 47,XX,þ8[1]/45,XX,20[3]/46,XX[23]
54
MR4
24 30
61 66
MR4 MMR
45,XX,5[4]/46,XX[35] 45,X,Y[4]/46,XY[26] 45,X,Y[3]/46,XY[37]
54 60 76
93
MR4 MR4 MR4
149 117
IM IM IM IM, NILO BMT, IM
CLL
IM IM IM
Granulocytopenia
IM, DASA
IM IM IM
(Continued on the next page)
508 Table 3
M. Gniot et al. (Continued )
Patient no.
Karyotype (no. of analyzed metaphases)
27
45,X,Y[4]/46,XY[21]
28
46,XY,21[3]/46,XY[23] 45,X,Y[3]/46,XY[17] 45,X,X[3]/45,XX,19[3]/46,XX[38]
Treatment time to aberration (mo) 6 24 48 108
Overall treatment time (mo)
Molecular response level
Coexisting abnormalities
Treatment
62
MMR
IM
54
MR4 CMR MR4
IM
Abbreviations: MMR, major molecular response; CMR, complete molecular response; CLL, chronic lymphocytic leukemia; IM, imatinib; DASA, dasatinib; NILO, nilotinib; BMT, bone marrow transplantation; MR4, MR4.5, MR5, reduction by 4, 4.5, and 5 logs according to the International Scale. a The shaded fields represent the supplementary data, which were excluded from statistical analysis due to the fact that the observations were made before reaching DMR and/or did not meet the ISCN criteria. These fields stress the recurrent/stable character of CCA/Ph found in further analyses. b Patient treated with interferon alpha before TKI therapy initiation.
of follow-up (median 62 months), not a single TKI-treated CML patient had developed a secondary Ph myeloid neoplasm, nor did they progress to a more advanced phase of the disease. However, in one patient from the CCA/Ph group, despite continuous MR4, the symptoms of chronic lymphocytic leukemia (CLL) occurred after 9 years of treatment with imatinib. The FISH analysis revealed that the patient had a population of cells with a split ATM gene signal (Table 3, patient 16). The study of the retrospective samples showed that this aberration was present for at least 4 years before CLL manifestation. Numerous precautions were introduced to minimize the potential bias in this study. In our study group, the most common CCA in Ph clones were chromosome 19, 20, 21 and Y monosomies. It is known that small chromosomes are more prone to loss during processing of the material. To avoid reporting of artificial aberrations related to laboratory
Figure 2 Median tyrosine kinase inhibitor treatment time in a group of CML patients with transient appearance of Ph-negative CCA in cytogenetic evaluation of cultured and GCSF stimulated bone marrow cells, using GTG techniques.
procedures, the ISCN guidelines were introduced (19). According to these guidelines, the monosomies found in more than three metaphases should be reported in the karyotype. In light of these guidelines, the over-reporting of the small chromosomes’ monosomies should not be an issue. In the three karyotypes presented in Table 3, changes were detected in lower than the normally accepted number of metaphases; however, those results were marked as supplementary data because the patients were not in DMR by that time. We decided to include them in the table only because these were the same aberrations that occurred in further analyses of the same patients, and they suggest the recurrent character of the observed changes. Some controversies may surround the Y monosomies. Until today, they were reported as age-related and affecting mainly elderly males. The loss of chromosome Y has been reported to occur in approximately 10% of patients with various hematological disorders (36). Recently, data that showed the strong correlation between the presence of Y monosomy in a fraction of bone marrow cells and cancer risk was published (37). Moreover, in CML patients, the loss of the Y chromosome in Phþ metaphases has been proven to have a similar prognostic impact as any other CCA (38). Therefore, in our opinion, relatively high-frequency Y monosomies (3 of 21 male patients; 14.1%) should be taken into consideration. In the majority of our investigated CML cases, CCA/Ph was detected incidentally during routine cytogenetic evaluation. CCA/Ph reoccurred in only five cases (Table 3). A similar phenomenon was reported by Kim et al. (35). They confirmed the transient presence of trisomy of chromosome 8 in Ph cells, and, in this study, no association between the occurrence of CCA in Ph cells and transformation to AML or MDS was found. It should be noted, however, that trisomy of chromosome 8 is one of the most frequent Ph aberrations that emerges during imatinib therapy in CML patients (27,32). The question remains: Was the sensitivity of the cytogenetic test used in diagnostic work-up sufficient? Recently, it was documented that the genome of CML patients harbors many cytogenetic changes. These might be hidden in subpopulations due to clonal events or due to their extremely complex characteristics. The use of several cytogenetic and molecular techniques (multicolor-FISH, comparative genomic
Ph-negative aberrations in CML hybridization (CGH) array) has been postulated to identify “cryptic” cytogenetic aberrations. However, this strategy is often questioned because it is time-consuming and costly, and cannot be converted to benefit the patients (39). Another problem is that the cytogenetic aberrations we observed are present in
