The introduction of imatinib has been a major advance in the treatment of chronic myeloid leukemia (CML). Based on the results of the recently published IRIS study, imatinib has now been approved by the Food and Drug Administration as treatment for newly diagnosed CML patients.1 However, despite its remarkable short-term efficacy and toxicity profile, little is known about the potential for long-term toxicity. This may be an important issue, as increasingly patients are declining stem cell transplantation (SCT) in favor of imatinib.
A recent article by Bumm et al2 draws attention to the occurrence of clonal karyotypic abnormalities in Philadelphia chromosome (Ph)-negative cells following imatinib-induced cytogenetic response. To date more than 20 such cases have been described with the most common abnormalities trisomy 8 and chromosome 7 defects.3,4 All were late chronic phase and had received prior therapy, in some cases extensive. Various explanations for this phenomenon have been proposed. Regarding a potential direct role for imatinib, Bumm et al2 state “a strong argument against a causative role of imatinib is, however, the fact that the phenomenon has not yet been observed in patients treated with imatinib up-front.” This is no longer the case as we now report.
A 53-year-old man presented in October 2000. His hemoglobin level was 127 g/L (12.7 g/dL), white blood cell count 30.1 × 109/L, and platelet count 856 × 109/L. Cytogenetic analysis confirmed a complex Ph translocation t(1;9;22)(q32;q34;q11) and no other cytogenetic abnormalities. He was treated with hydroxyurea for 5 weeks before enrolling on the IRIS study, being randomized to receive 400 mg imatinib daily. After 5 months, treatment was interrupted for 5 weeks due to grade 3 neutropenia and thrombocytopenia. Imatinib was reintroduced at 300 mg and maintained at this level. Serial cytogenetic analysis was performed approximately every 12 weeks (Table 1). Within 12 months, a major cytogenetic response was evident with only 1 of 20 metaphases Ph positive. However the 19 Ph-negative metaphases revealed new abnormalities with trisomy 8 and monosomy 7 detected in 16 and 3 cells, respectively. Trisomy 8 persisted and at 14 months all 10 cells analyzed showed a 47, XY,+8 karyotype. Fluorescent in situ hybridization (FISH) analysis showed all 100 cells to be Bcr-Abl negative at this time. The diagnostic sample was retrospectively checked by FISH using an 8 centromeric probe, but none of the 100 nuclei showed trisomy 8. This confirmed that the trisomy 8 clone became evident only after imatinib treatment was initiated. Bone marrow did not show any dysplastic changes at any stage. Because of concerns about the abnormal karyotype he underwent allogeneic SCT in January 2002. He engrafted successfully with no complications and remains in complete cytogenetic remission.
This case is notable as this patient was in early chronic phase and had received little therapy other than imatinib. Could imatinib have influenced the emergence of these abnormal clones? Previous cases described have all been late chronic phase, suggesting that the emergence of clonal abnormalities could be as a result of a long disease process or prior therapy negatively inhibiting normal hematopoiesis, thus providing the selective pressure for the outgrowth of a resistant Ph-negative clone. Our case, we believe, supports speculation that imatinib may have a direct effect. The normal Abl gene product is involved in the cellular response to DNA damage, thus inhibition may lead to genetic instability.5 Imatinib also inhibits c-kit.6 Stem cells deficient in c-kit are unable to reconstitute hematopoiesis in myeloablated mice, and mutant c-kit causes a severe macrocytic anemia.7 Long-term administration could have potentially suppressive effects on normal stem cells leading to emergence of new clones. These effects may be more pronounced after long disease duration as the pool of Ph-negative stem cells declines.
Ultimately, however, we do not know the significance of these abnormalities in Ph-negative cells, and longer follow-up with larger patient numbers will be required. We are in agreement with Bumm et al2 on the need to establish a registry to collect data on these patients.
Chromosomal abnormalities in Ph-negative cells
Several case reports and small studies have appeared that describe clonal cytogenetic abnormalities in Ph-negative cells of chronic myeloid leukemia (CML) patients treated with imatinib. Trisomy 8 and monosomy 7 are most frequent, but other aberrations have also been reported.1
The cases published thus far were in late chronic or accelerated phase. Thus, the patient described by McMullin et al is unusual, since he had received only a short course of hydrea prior to imatinib. To our knowledge, there is only one similar published case. This patient exhibited t(2;11) both in Ph-negative and Ph-positive cells, implying that this aberration was present before the Philadelphia translocation was acquired,2 in contrast to the case of McMullin et al, where the abnormalities were confined to Ph-negative cells.
One of the main issues raised by these observations is whether imatinib has a direct mutagenic effect. Preclinical studies showed that imatinib and some of its intermediates are genotoxic in vitro.3 In vivo effects of intermediates can be excluded, since they are not present at relevant concentrations in the purified drug. Clastogenic effects of imatinib occurred in only one in vitro chromosome aberration assay at the highest concentration (125 mg/L), which is highly cytotoxic and exceeds drug levels achieved in patients by several orders of magnitude.4 Effects such as this are not unusual and known as high-toxicity clastogenicity (J. Ford and D. Roman, personal written communication, July 2003). In addition, in our study,5 we observed an association between chromosomal abnormalities in Ph-negative cells and prior exposure to cytarabine and idarubicin, arguing against a direct genotoxic effect of imatinib. Moreover, if imatinib was directly genotoxic, one would expect a similar incidence in newly diagnosed patients after comparable follow-up. The median time on imatinib in the patients with abnormalities in our study was 16 months.5 This is shorter than the 19 months median follow-up in the randomized comparison of imatinib versus interferon-alpha/cytarabine in newly diagnosed patients.6 Since abnormalities in Ph-negative cells were not specifically captured in this study, underreporting cannot be excluded. However, an incidence of 15% as in our study would hardly have gone undetected.
Thus, it seems likely that the proliferative stress placed on a small number of Ph-negative stem cells to restore hematopoiesis in the presence of imatinib may be more important. In this case, patients whose disease has evolved over a longer period of time before diagnosis will be at higher risk, due to their smaller pool of Ph-negative progenitor cells. The patient described by McMullin et al did not achieve a major cytogenetic response until 12 months, unlike most newly diagnosed patients.6 This may indicate that his disease, though in chronic phase by definition, was in fact in a later stage of evolution. Damage to Ph-negative stem cells by previous exposure to cytotoxic agents would add to the risk, consistent with our findings.5 Whether the stress to Ph-negative cells is aggravated by inhibition of Kit remains speculative; Kit mutant mice exhibit compromised hematopoiesis but are not prone to leukemia.7
In reality, the 2 possibilities cannot be distinguished with certainty. What might be more revealing is whether similar abnormalities appear in solid tumor patients treated with imatinib. No less important than the etiology of this phenomenon is its prognosis. Progression to myelodysplastic syndrome (MDS) has occurred in several patients, and a decision to proceed to allografting was made.8 Given the dismal prognosis of chromosome 7 abnormalities in MDS, this is certainly justified. However, the significance of other abnormalities, particularly isolated trisomy 8 in the absence of dysplastic changes, is less clear. For adequate counseling of patients, a systematic collection of information is required. A registry has been set up at Oregon Health & Science University, to which contributions are invited.
Correspondence: Michael W. N. Deininger, BMT/Leukemia Center, Oregon Health & Science University, 3181 SW Sam Jackson Park Road, Mail code L592, Portland, OR 97239; e-mail: deininge@ohsu.edu.
