Clinical and genetic features of therapy-related myeloid neoplasms after chemotherapy for acute promyelocytic leukemia

Acute promyelocytic leukemia (APL) is a distinct subtype of acute myeloid leukemia (AML) characterized by a t(15;17) translocation leading to a PML-RARA fusion gene. APL is a highly curable disease with excellent complete remission (CR) and long-term survival rates. All-trans retinoic acid (ATRA) combined with anthracycline-based chemotherapy yields a CR rate of approximately 90% for newly diagnosed APLs. The relapse rate is approximately 20%, and with the development of new molecular target therapies such as arsenic trioxide, a cure can now be expected even for relapsed patients.

However, the development of therapy-related myeloid neoplasms (t-MN) is being reported with an increasing frequency of 0.97% to 6.5% in patients successfully treated for APL.^1-3  Patients are typically in hematologic and cytogenetic remission for APL, and at follow-up after treatment, hematologic abnormalities may be detected without evidence of the original PML-RARA fusion gene seen at APL diagnosis. The t-MN secondary to APL is usually difficult to treat, and it is one of the prognosis-limiting factors for the curable APL disease. Therefore, we attempted to clarify the different clinical features and hematologic findings between t-MN and relapsed APL cases, and to identify gene alterations associated with t-MN.

Patients with APL were diagnosed at Hiroshima University Hospital and its affiliated hospitals between 1996 and 2008. Diagnoses of APL were initially made morphologically according to the French-American-British classification and were confirmed by the presence of t(15;17) and/or presence of the PML/RARA. Patients were treated with ATRA combined with chemotherapy and with intensified maintenance chemotherapy.^4,5  Patients who were resistant to ATRA therapy were treated with arsenic trioxide. Drugs used during APL treatment were idarubicin, cytarabine, mitoxantrone, daunorubicin, etoposide (VP-16), behenoyl cytarabine, mercaptopurine, vindesine, and aclarubicin. All patients received an intrathecal administration of methotrexate, cytarabine, and prednisolone during consolidation therapy. Patients with APL were examined as approved by the Institutional Review Board at Hiroshima University. Patients gave written informed consent in accordance with the Declaration of Helsinki.

Mononuclear cells were isolated from bone marrow samples, and genomic DNA and total RNA were prepared as described previously.⁶  Identification of RUNX1, CEBPA, FLT3, and NRAS mutations was performed as described previously.^6-8 

Probabilities of survival were estimated using the Kaplan-Meier method and compared by the log-rank test.

A total of 124 patients with APL were newly diagnosed and were consecutively enrolled in the study. The median age at diagnosis was 52 years (range, 18-86 years). An 86-year-old patient died of lung hemorrhage before starting induction therapy. Other patients received induction therapy, including ATRA with chemotherapy. Two patients were resistant to ATRA therapy and were treated with arsenic trioxide. Six patients who were 70 years of age or older died during induction therapy, whereas 117 patients (94.4%) achieved CR. After achieving CR, 2 patients dropped out of the study because they moved to a distant place, and 2 patients could not continue the therapy because of neuropsychologic diseases and died within 1 year. Others received intensive consolidation chemotherapy, but 5 patients died of therapy-related complications. All of the remaining 108 patients completed consolidation therapy. The predicted 10-year overall survival was 74.6% for the whole population, and 81.4% and 36.9% in patients younger than 70 years and 70 years or older, respectively. Thus, the efficacy of the treatment for APL patients in this study is comparable with that reported in clinical trials.^4,5 

After a median follow-up of 8.6 years (range, 1.7-16.3 years), 5 patients died of nonhematologic causes, 10 patients (9.3%) relapsed, and 11 patients (10.2%) developed t-MN among the 108 patients during their first CR from APL. We also analyzed 61 patients with RUNX1-ETO or CBFB-MYH11 (core-binding factor leukemia) during the same research period and found 13 relapses (21.3%) but no t-MN. Thus, there is a possibility that t-MN after successful treatment of APL may be more popular than AML other than APL. We noted more t-MN patients instead of fewer relapse patients compared with other reports. The fact that all patients in this study had received VP-16 for consolidation therapy whereas no VP-16 was administered in other trials could explain, in part, the higher incidence of t-MN.³  Furthermore, the accumulation of chemotherapeutic agents in the maintenance phase may increase the risk of t-MN.⁵  Meanwhile, there is a possibility that t-MN might have been diagnosed as relapsing APL because of positive PML-RARA from coexisting APL cells. Therefore, we attempted to clarify the differences in clinical features and hematologic findings between t-MN and relapsed APL patients after successful treatment of APL.

The patients who developed t-MN with progression to AML, therapy-related myelodysplastic syndrome, refractory cytopenia with multilineage dysplasia (t-MDS-RCMD), or relapse are shown in Tables 1 and 2. The median white blood cell (WBC) count at APL diagnosis was 6940/μL (range, 1.1-42.6/μL) in the t-MN patients and 33 630/μL (range, 1.1-148.6/μL) in the relapse patients, whereas it was 13 130/μL (range, 0.3-152.8/μL) in the disease-free patients who remained continuously in their first CR for 1.5 to 16.1 years (median, 8.3 years). The WBC count was significantly lower in the t-MN patients than in the relapse patients (P = .048). Relapse risk based on pretreatment WBC count and platelet count indicated that most of the t-MN patients were distributed in low- and intermediate-risk groups (Figure 1A). In particular, the WBC count at APL diagnosis was less than or equal to 4500/μL in most of the t-MN patients, whereas it was more than 4500/μL in the relapse patients. Previous reports also showed that most of the t-MN patients had a low WBC count.^3,9  After the first APL diagnosis, the median intervals were 2.6 years (range, 0.6-10.1 years) to relapse, 2.3 years (range, 1.6-3.0 years) to t-MDS-RCMD, and 3.3 years (range, 1.0-9.7 years) to t-MN with progression to AML (P = .83). However, overall survival from the start of chemotherapy was significantly worse for the t-MN with progression to AML patients than for the relapse patients (P = .022, Figure 1B). VP-16 or intensive maintenance chemotherapy may reduce the risk of relapse; however, the frequency of t-MN may increase. Recent therapeutic approaches in the treatment of refractory APL may improve the second CR rate. Our results suggest that preventing t-MN may be more important than preventing relapse, especially in patients with low WBCs.

We examined surface markers of the leukemic cells (Figure 1C). At the primary APL diagnosis, most of the leukemic cells from the APL patients showed CD34⁻HLA-DR⁻CD33⁺ phenotypes. At relapse, the phenotypes showed the same pattern, whereas at the t-MN CD34 and HLA-DR phenotypes changed to positive and the positivity of CD33 became lower. The relapse patients and the t-MN patients were easily distinguished by their CD34/HLA-DR distribution pattern; APL cells were CD34⁻HLA-DR⁻, whereas t-MN cells were CD34⁺ and/or HLA-DR⁺ (Figure 1D). These results indicate that the leukemic cells changed to primitive myeloid cells when t-MN developed after successful treatment of APL, and they were considered to have an independent clonal origin.¹⁰ 

To confirm their clonality, gene abnormalities that may play an important role in leukemogenesis were analyzed.¹¹  All of the relapse patients had the PML-RARA gene, whereas none of the patients with t-MN had PML-RARA. Instead, we found translocations involving 21q22 of RUNX1 (RUNX1-MTG16) or 11q23 of MLL (MLL-FOXO3 and MLL-CBP; Table 2), as previously reported.^9,12,13  Both of the t-MDS-RCMD patients had an isolated del(20q), which may be associated with a good prognosis. Furthermore, 4 RUNX1 mutations and one CEBPA mutation were detected. These abnormalities were not detected at the primary APL diagnosis or in the relapsed patients with APL or core-binding factor leukemia. RUNX1 mutations are known to be frequently involved in t-MN; however, no such reports included t-MN from APL. One patient with a RUNX1 mutation also acquired an NRAS mutation at t-MN. Three of 4 RUNX1 mutations were associated with monosomy 7, which is frequently reported in t-MN after successful treatment of APL.^3,9  It is possible that previously reported monosomy 7 cases were also associated with RUNX1 mutations, but this was not investigated in the earlier studies.

In addition, a recent case of t-MN at 8 months after APL diagnosis showed a phenotype of increasing CD34⁺HLA-DR⁺ cells, gating on the CD45RO (Figure 1E). Karyotypic analysis showed a del(7q) abnormality, although the percentage of blasts was less than 5% in the bone marrow, and the RUNX1 mutation was detected at 12 months. Monitoring of surface markers, karyotype, and the RUNX1 mutation, as well as PML-RARA monitoring, may be useful for the early diagnosis of t-MN in patients successfully treated for APL.

A previous report showed that RUNX1 mutations may have a leukemogenic potential in CD34⁺ cells from patients in the chronic phase of myeloproliferative neoplasms.⁷  In addition, in APL, it is assumed that RUNX1 or other abnormalities may be induced in CD34⁺ cells during chemotherapy resulting in t-MN after successful treatment of APL. Some patients with JAK2V617F⁺ myeloproliferative neoplasms were reported to transform to JAK2V617F⁻ AML, suggesting that leukemia developed from a pre-JAK2 hematopoietic stem cell¹⁴  or from a normal hematopoietic stem cell.¹⁵ PML-RARA⁻ t-MN may also develop from a “receptive” hematopoietic stem cell or from a normal hematopoietic stem cell, which is a myeloid committed progenitor, by the accumulation of chemotherapy-induced gene abnormalities, including RUNX1 mutations.

The authors thank Ryoko Matsumoto-Yamaguchi, who provided excellent technical support in carrying out these experiments.

This work was supported in part by the Ministry of Education, Culture, Sports, Science and Technology of Japan (Grants-in-Aid for Scientific Research: grants 21591206 and 22591038, H. Harada, Y.H.). This work was carried out in part at the Analysis Center of Life Science, Hiroshima University.

Contribution: J.I. collected data and wrote the paper; Y.H. assembled and analyzed the data and revised the manuscript; T.S., H.T., Y.O., H. Hyodo, and A.K. provided patient samples and clinical information; and H. Harada designed the research and revised the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Hironori Harada, Department of Hematology and Oncology, Research Institute for Radiation Biology and Medicine, Hiroshima University, 1-2-3 Kasumi, Minami-ku, Hiroshima 734-8553, Japan; e-mail: herf1@hiroshima-u.ac.jp.