A retrospective study of the correlation of in vitro chemosensitivity using ATP‑TCA with patient clinical outcomes in acute myeloid leukemia
Abstract
Purpose To evaluate the predictive value of the in vitro chemosensitivity using ATP-TCA method to compare the clinical efficacy of patients with AML.Methods Bone marrow or peripheral blood samples were collected from 65 patients with AML, and the in vitro chemosensi- tivity of four drugs (cytarabine/idarubicin/decitabine/aclacinomycin) was measured by an ATP-tumor chemosensitivity assay. Results Aclacinomycin and cytarabine had the highest chemosensitivity rates (66.7%, 8/12 and 58.5%, 38/65, respectively), while idarubicin and decitabine had rates of 6.5% (3/46) and 0% (0/35), respectively. Complete remission (CR) was achieved in 66.2% (43/65) of patients, and there was a statistically significant correlation between CR and in vitro chemosensitivity for cytarabine (47.7% vs 18.5%, p = 0.002), but not for the anthracyclines (p = 0.950). In addition, three other factors significantly correlated with CR: disease status (p = 0.005), FLT3-ITD/TKD mutation (p = 0.003) and chemotherapy regimens (p = 0.004). Furthermore, multiple logistic regression analysis revealed that the sensitivity of cytarabine was one of the significant risk factors for CR [hazard ratio (HR) = 5.52; 95% confidence interval (CI) = 1.47–20.70; p = 0.011].Conclusions The in vitro chemosensitivity as tested by ATP-TCA demonstrated a significant correlation with CR for chemo- therapy and can be a useful tool to optimize personalized treatments for patients with AML.
Introduction
Acute myeloid leukemia (AML) is a heterogeneous hemato- logic malignancy characterized by the clonal expansion of myeloid blasts in the peripheral blood, bone marrow, and/or other tissues [1]. Chemotherapy is widely used as an effec- tive method for the treatment of AML and has been divided into induction and postremission therapy. Chemotherapy drugs can kill unrelated cells and damage the patient’s body, further inducing serious drug resistance. Therefore, predicting the chemoresistance of tumor cells and developing indi- therapy reduce the damage to the patient’s body and reduce the cost of treatment, which are particularly important. At present, chemosensitivity studies in vitro are considered one of the approaches to solve this problem.There are many methods of chemosensitivity studies in vitro that have been reported, such as the 3-(4,5-dimethyl- thiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, microculture-kinetic drug-induced apoptosis (MICK) assay, and ATP-chemosensitivity (ATP-TCA) assay [2–4]. Among them, the ATP-TCA is a common method to detect the sen- sitivity of antitumor drugs in vitro with the advantages of high sensitivity, simple operation, satisfactory reproducibil- ity, and high accuracy. Previously, researchers have found that ATP-TCA can be used in lung cancer, ovarian cancer, cervical cancer, esophageal cancer, breast cancer, and other cancers for chemosensitivity studies, indicating a beneficial relationship with clinical outcomes [4–8]. However, the cor- relation studies were limited to in vitro chemosensitivity, and the results of chemotherapy for AML patients and methods of in vitro chemosensitivity have been studied and reported for MTT and MICK assays, without ATP-TCA [9–11]. In this study, we measured in vitro chemosensitivity with ATP- TCA from the bone marrow or peripheral blood samples of 65 patients with AML from June 2016 to December 2018 at the First Affiliated Hospital of Suzhou University.
Further- more, we compared the results of in vitro chemosensitivity measured by ATP-TCA and the clinical outcomes of induc- tion chemotherapy for newly diagnosed patients and treat- ment for patients with relapsed or refractory disease.The bone marrow or peripheral blood samples were collected from 65 patients with AML from June 2016 to December 2018 at the First Affiliated Hospital of Soochow University. Written informed consent was obtained from each patient, and this study was approved by the Research Ethics Board of the First Affiliated Hospital of Soochow University. The diagnosis of AML was due to cytomorphology, immunophe- notype, and cytogenetics by the French–American–British (FAB) classification. All patients were treated with chemo- therapy regimens based on cytarabine (AraC), classified as follows: (1) Regimen 1: AraC + idarubicin (IDA); (2) Regi- men 2: AraC + IDA + decitabine (DAC); and (3) Regimen 3: AraC ± DAC/aclacinomycin (ACLA)/IDA/etoposide/ vindesine/cladribine (one or more of the above).Separation of mononuclear cells from bone marrowTwo milliliters of fresh bone marrow/7 mL of peripheral blood was diluted with saline (v/v, 1:1), and mononuclear cells were separated with a Ficoll–Hypaque gradient (v/v, 2:1) at 2500 rpm for 20 min at room temperature. Then, the intermediate layer was transferred and centrifuged at 1000 rpm for 6 min, and the supernatant was discarded. Cells were stained with trypan blue, and cell numbers were counted with a blood cell counting plate.Chemosensitivity measurements with ATP‑TCAThe chemotherapeutic drugs were diluted and placed into a 96-well plate with five test drug concentrations (TDC) (n = 2) of 200%, 100%, 50%, 25%, and 12.5% of peak plasma concentrations (PPC) in the conventional clinical dose according to the manufacturer’s protocol or pharmacopoeia.
The rest of the wells were left as cells without drug treat- ment (M0) and blank controls without cells. Then, 100 µL of the cell suspensions was placed into each well and cultured in an incubator (37 °C, 5% CO2 and more than 95% humid- ity) for 5–7 days. The medium was a selective culture of nutrients that could accelerate normal cell death and tumor cell survival and was not changed or treated with new drugs during this period.After incubation, 40 μL of ATP extraction agent (TCE) was added to each well, and ATP extraction was completed after 40 s at room temperature. A total of 60 μL of the cul- ture suspension of each well was transferred to a fluorescent determination plate, and then each well was tested immedi- ately after 40 µL of LU–LU reagent was added. The fluores- cence intensity of each well was measured at 560 nm with a fluorescence reader, and the average fluorescence intensity value was calculated.The inhibition rate was calculated by the following equa- tion: inhibition rate (%) = [1 − (X − MI)/(M0 − MI)] × 100%. The dose–response curves could be obtained by the inhibi- tion rates. TDCs and the IC90 and IC50 values were calcu- lated by the dose–response curves.Criteria of in vitro chemosensitivityIn vitro chemosensitivity was classified into the follow- ing four categories [12]: ① strong sensitivity: IC90 ≤ 100% TDC and IC50 ≤ 25% TDC; ②moderate sensitivity: IC90 > 100% TDC and IC50 ≤ 25% TDC; ③ mild sensitiv- ity: IC90 ≤ 100% TDC and IC50 > 25% TDC; and ④ resistance: IC90 > 100% TDC and IC50 > 25% TDC.
In the following specific analysis, regarding the results of in vitro chemosensitivity, those with “strong sensitivity” were set to the sensitive group and those with “moderate sensitivity”, “mild sensitivity”, and “resistance” were set to the resistant group.Evaluation of clinical outcomes and definitionsPatients who achieved complete remission (CR) or CR with incomplete platelet recovery (CRp) were considered responders and were classified into the CR group. Those who achieved any other response were considered nonre- sponders and were classified into the noncomplete remis- sion (NCR) group. The response criteria for AML con- formed to the International Working Croup Criteria [13]. CR required the amount of myeloblasts in bone narrow to be ˂5%, absolute neutrophil count ≥ 1.0 × 109/L, platelet count ≥ 100 × 109/L, and absence of circulating blasts or extramedullary disease. CRp was defined as CR without platelet transfusion for 1 week. Data were analyzed by SPSS 19.0 software. Continuous variables were analyzed using descriptive statistical methods (median, range, minimum, and maximum). T test, χ2 test, and Fisher’s exact probability test were used to compare the categorical variables. Multiple logistic regression analysis was performed to identify possible factors for CR. P ≤ 0.05 was considered statistically significant.
Results
A total of 65 patients with AML were analyzed in our study, and all patients had the chemosensitivity test per- formed before chemotherapy. Most of the samples for the chemosensitivity test were from bone marrow (47/65; 72.3%). The median age of the patients was 37 years (range 24–49 years), and 34 (52.3%) patients were male. There were 49 (75.4%) patients with AML who were newly diag- nosed and 16 (24.6%) with relapsed or refractory disease. The median white blood cell (WBC) count at diagnosis was 14.14 × 109/L (range 5.19–47.23 × 109/L), which was higher than the normal upper limit (10.0 × 109/L). About thirty percent of the patients (20/65) had an FMS-like tyros- ine kinase 3 (FLT3) mutation (including FLT3-internal tan- dem duplications (ITD) and FLT3-tyrosine kinase domain (TKD) mutations), 24.6% (16/65) had a CCAAT/enhancer- binding protein alpha (CEBPA) mutation, and 12.3% (8/65) had a nucleophosmin-1 (NPM-1) mutation. All patients were treated with chemotherapy regimens based on AraC; 43 patients achieved CR, and the CR rate was 66.2% (Table 1).The in vitro responses of AML cells to the four drugs were tested using the ATP-TCA method. The median IC90 val- ues of the drugs from low to high were as follows: ACLA˂ AraC ˂ IDA ˂ DAC, which was the same order for IC50 values (Figs. 1, 2).
Among the four drugs, AraC had a higher chemosensitivity rate of 58.5% (38/65). In the 12 samples treated with ACLA, 8 showed strong sensitivity (66.7%). IDA demonstrated a low chemosensitivity to AML cells in vitro (3/46; 6.5%), and the sensitivity rate for DAC was 0.0% (0/35; Table 2).Correlation between in vitro chemosensitivity and clinical outcome of chemotherapyCR was achieved in 66.2% (43/65) of patients within the test cohort. Statistically, a correlation existed between the AML acute myeloid leukemia, WBC white blood cell, FLT3 FMS-like tyrosine kinase 3, CEBPA CCAAT/enhancer-binding protein alpha, NPM-1 nucleophosmin-1, CR complete remission (including CR and CRp), NCR noncomplete remissionthere was no statistically significant difference (p = 0.950; Table 3).Correlations with complete remission and multivariate analysisThe patient characteristics, chemotherapy regimens, and chemosensitivity of AraC were compared between the CR and NCR groups. Sex, age, WBC count at diagnosis, CEBPA mutation, and NPM-1 mutation did not differ between the two groups. Patients with relapsed or refractory disease and FLT3-ITD/TKD mutations were significantly more frequent in the NCR group than in the CR group (45.5% vs. 14.0%, p = 0.005 and 54.5% vs. 18.6%, p = 0.003, respectively). The chemotherapy regimens were significantly different between the two groups (p = 0.004). Finally, AraC with strong sensi- tivity was more frequent in the CR group (72.1% vs. 31.8%, p = 0.002) (Table 4).Multiple logistic regression analysis was performed to identify the significant factors for CR. After an adjustment of four variables, disease status (OR 8.71, 95% CI 1.84–41.15,p = 0.006), FLT3-ITD/TKD mutation (OR 0.11, 95% CI0.02–0.46, p = 0.003), and sensitivity of AraC (OR 5.52, 95% CI 1.47–20.70, p = 0.011) were found to be independent factors for CR. In addition, the chemotherapy regimens did not affect CR in the multivariable logistic regression model (Table 5).
Discussion
AML is a potentially fatal disease that shows only 30–40% long-term disease-free survival for young patients, and chemotherapy is the main treatment [14–16]. The stand- ard chemotherapy regimens used are based on a backbone of AraC plus an anthracycline (named the 3 + 7 regimen), which has marginally changed in the past 40 years [1]. Almost all newly diagnosed patients are treated with this regimen, but the responses to chemotherapy are different. These individual differences depend on patient character- istics, karyotypic or molecular abnormalities, antecedent myelodysplasia, chemotherapy regimens, and pharmacologi- cal resistance [11, 17, 18]. Among them, cell resistance is a major cause of treatment failure, which affects the survival outcomes of patients with AML [19, 20]. Thus, a test of cell resistance before chemotherapy may aid in the develop- ment of a more effective individualized regimen for patients. Therefore, we carried out this study to evaluate the predic- tive value of the in vitro sensitivity of chemotherapeutic drugs to AML cells and to compare the clinical efficacy of 65 patients with AML.Compared with conventional tumor susceptibility assays,such as the MTT [21, 22], MICK [3], human tumor clono- genic (HTCA) [23] and other assays [24, 25], ATP-TCA is a technology with higher sensitivity, higher specificity, and a lower requirement of the number of tumor cells [26]. In our study, we used the ATP-TCA assay to test the chemosensitiv- ity in vitro for AraC, IDA, ACLA, and DAC. The AML cells showed a higher in vitro sensitivity to AraC (38/65, 58.5%) and a lower sensitivity to IDA and DAC (3/46, 6.5% and 0/0, 0.0%, respectively). For ACLA, eight samples showed strong sensitivity (8/12, 66.7%).
The low sensitivity in vitro of DAC may be due to its mechanism of action. In vitro, high- dose DAC causes DNA damage and DNA synthesis arrest and leads to cytotoxicity, but low-dose DAC induces only DNA methyltransferase inhibition with minimal cytotoxicity [27]. In our study, the median IC90 and IC50 values of DAC were very high. In addition, we found another interesting phenomenon in which the eight AML patients with AraC resistance showed strong or moderate sensitivity to ACLA, but further study is needed to analyze this correlation.We analyzed the relationship between the in vitro and in vivo efficacy of susceptibility for AraC and 2 anthracy- clines (not including DAC because of the resistance in vitro). We found that the CR rate was lower in patients with AraC resistance than in patients with AraC sensitivity (18.5% vs. 47.7%, p = 0.002). However, there was no significant dif- ference between the chemosensitivity of anthracyclines and CR, which might be due to the small number of samples. These results suggested that the in vitro susceptibility of AraC had a good correlation with clinical efficacy and could be used to optimize individualized treatment for AML.In our study, 43 patients obtained CR (43/65, 66.2%), and a multivariate analysis of the variables associated with CR indicated statistical significance for disease status, FLT3- ITD/TKD mutation, and AraC sensitivity in vitro.
The disease status, relapsed or refractory, is associated with a poor prognosis, and only a small fraction of these patients can achieve successful second remission [28]. Our study showed the same result that patients with relapsed or refractory dis- ease were more frequent in the NCR group. Another risk factor for CR was the presence of an FLT3 mutation, which was present in 30% of AML patients, with the majority being FLT3 ITD and less frequently FLT3 TKD [29, 30]. Until now, numerous studies have shown poor prognosis in patients with FLT3-ITD mutations, and the median overall survival (OS) from the time of diagnosis for these patients was only 6–10 months [31]. The presence of the FLT3-TKD mutation has also been shown to be associated with shorter remission durations and decreased OS [32]. In our study, patients with FLT3-ITD/TKD mutations were likely to have poor clinical outcomes, which was consistent with the find- ings of most studies. In our study, we found that AraC sen- sitivity in vitro was significantly associated with CR, which showed that it could be a predictive factor for the clinical outcomes of AML patients. It was reported that sensitivity to both AraC and daunomycin (DNR) in AML patients was related to continuous CR or relapse, and the MTT assay may be a useful tool for individually optimizing chemotherapy [11], which was partially consistent with our results.Because this study is retrospective, patients were notclassified according to the risk level at the time of initial diagnosis, and the effects of other genetic abnormalities that are correlated with the treatment outcomes of AML were not analyzed. Therefore, the results of the study need fur- ther confirmation. In addition, in our study, all patients were diagnosed with AML and started treatment after 2016, and most patients had accepted hematopoietic stem cell trans- plantation, which was a main factor affecting OS. Therefore, we did not analyze the correlation between chemosensitivity in vitro and long-term survival.
In conclusion, the chemosensitivity in vitro measured by the APT-TCA technique in this study is in good agreement with the in vivo efficacy. The sensitivity in vitro of AraC can predict the treatment Idarubicin outcomes of chemotherapy, and the utilization of this technique may provide a theoretical basis for choosing an individualized regimen for AML.