Artigo Acesso aberto Revisado por pares

Intracellular cytarabine triphosphate in circulating blasts post-treatment predicts remission status in patients with acute myeloid leukemia

2019; Elsevier BV; Volume: 74; Linguagem: Inglês

10.1016/j.exphem.2019.04.005

ISSN

1873-2399

Autores

Elizabeth Anderson, B. Rees, Jonathon Hull, Jonathan Heywood, Andrea Preston, Rachel Protheroe, Emily J. Foulstone, Rosemary Greenwood, Vyv Salisbury, Priyanka Mehta,

Tópico(s)

Chronic Lymphocytic Leukemia Research

Resumo

•Intracellular ara-CTP is quantifiable in circulating AML blasts post-treatment.•Intracellular ara-CTP 4 hours post-treatment correlates with day 28 burden decrease.•Sampling for analysis can be integrated into a busy hemato-oncology ward. Cytarabine remains the backbone of therapy in acute myeloid leukemia (AML). The ability to assess intracellular cytarabine triphosphate (ara-CTP) levels in patients receiving cytarabine represents a major goal in the prediction of treatment response. This study, conducted within a clinical setting, aimed to assess ara-CTP levels in circulating peripheral blasts from non-M3 AML patients receiving cytarabine at one of three dosing levels, using a novel biosensor assay. Results from the initial 72 hours post-commencement were correlated with day 28 remission status, with feasibility parameters concurrently assessed. Intracellular ara-CTP was detectable in ex vivo blasts post-treatment for standard-dose (SD) and high-dose (HD) patients (p < 0.05), and quantification revealed a 27-fold increase in intracellular steady-state concentration between the two dosing levels. For low-dose cytarabine, high rates of patient discharge and low intracellular concentrations limited analysis; however, assessment of intracellular ara-CTP concentration was achievable in a dwindling population of blasts for SD and HD treatment cohorts, with 4 hours post-treatment commencement potentially being most predictive of clinical response (r = –0.912, p = 0.0113). Concurrent assessment of peripheral leukemia-associated immunophenotype (LAIP)-positive cells revealed a decline in burden (0–72 hours), which correlated with remission status (p < 0.05). Unexpectedly high rates of night sampling led to challenges associated with sampling rates, but did not have an impact on patient compliance. Additional training of night staff improved feasibility substantially. Multiple peripheral sampling during the initial 72 hours of treatment is feasible in newly diagnosed patients, and ara-CTP is detectable over the initial 24 hours, facilitating prediction of chemosensitivity of leukemic blasts to cytarabine. Cytarabine remains the backbone of therapy in acute myeloid leukemia (AML). The ability to assess intracellular cytarabine triphosphate (ara-CTP) levels in patients receiving cytarabine represents a major goal in the prediction of treatment response. This study, conducted within a clinical setting, aimed to assess ara-CTP levels in circulating peripheral blasts from non-M3 AML patients receiving cytarabine at one of three dosing levels, using a novel biosensor assay. Results from the initial 72 hours post-commencement were correlated with day 28 remission status, with feasibility parameters concurrently assessed. Intracellular ara-CTP was detectable in ex vivo blasts post-treatment for standard-dose (SD) and high-dose (HD) patients (p < 0.05), and quantification revealed a 27-fold increase in intracellular steady-state concentration between the two dosing levels. For low-dose cytarabine, high rates of patient discharge and low intracellular concentrations limited analysis; however, assessment of intracellular ara-CTP concentration was achievable in a dwindling population of blasts for SD and HD treatment cohorts, with 4 hours post-treatment commencement potentially being most predictive of clinical response (r = –0.912, p = 0.0113). Concurrent assessment of peripheral leukemia-associated immunophenotype (LAIP)-positive cells revealed a decline in burden (0–72 hours), which correlated with remission status (p < 0.05). Unexpectedly high rates of night sampling led to challenges associated with sampling rates, but did not have an impact on patient compliance. Additional training of night staff improved feasibility substantially. Multiple peripheral sampling during the initial 72 hours of treatment is feasible in newly diagnosed patients, and ara-CTP is detectable over the initial 24 hours, facilitating prediction of chemosensitivity of leukemic blasts to cytarabine. Prediction of response to therapy and risk stratification is a major goal in the early treatment of acute myeloid leukemia (AML). Whilst complete remission (CR) rates have greatly improved in younger patients with AML, high relapse rates and poor disease-free survival still pose a significant challenge [1Bose P Vachhani P Cortes JE Treatment of relapsed/refractory acute myeloid leukemia.Curr Treat Options Oncol. 2017; 18: 17Crossref PubMed Scopus (145) Google Scholar]. The backbone of AML therapy is cytarabine (ara-C), given at one of three dosing levels, depending on stage of disease (induction or relapse) and patient fitness (intensive or non-intensive therapy). Because of the rapid commencement of therapy after diagnosis, often prior to prognostic information being available, early assessment predicting response to cytarabine could be an important clinical decision-making tool. The relationship between intracellular ara-CTP accumulation and disease response has long been controversial, with some studies advocating a link [2Karp JE Donehower RC Dole GB Burke PJ Correlation of drug-perturbed marrow cell growth kinetics and intracellular 1-B-D-arabinofuranosylcytosine metabolism with clinical response in adult acute myelogenous leukemia.Blood. 1987; 69: 1134-1140PubMed Google Scholar, 3Rustum YM Preisler HD Correlation between leukemic cell retention of 1-beta-D-arabinofuranosylcytosine 5′-triphosphate and response to therapy.Cancer Res. 1979; 39: 42-49PubMed Google Scholar] and others dismissing it [4Gruber A Liliemark E Tidefelt U et al.Pharmacokinetics of mitoxantrone, etoposide and cytosine arabinoside in leukemic cells during treatment of acute myelogenous leukemia: Relationship to treatment outcome and bone marrow toxicity.Leukemia Res. 1995; 19: 757-761Crossref PubMed Scopus (8) Google Scholar, 5Koehl U Hollatz G Rohrbach E et al.Pharmacology of intracellular cytosine-arabinoside-5′-triphosphate in malignant cells of pediatric patients with initial or relapsed leukemia and in normal lymphocytes.Cancer Chemother Pharmacol. 2007; 60: 467-477Crossref PubMed Scopus (4) Google Scholar]. This single-centre, observational study used a previously validated biosensor technology [6Anderson E Conway M Alloush H et al.Investigation and verification of a bioluminescent biosensor for the quantitation of ara-CTP generation: A biomarker for cytosine arabinoside sensitivity in acute myeloid leukaemia.Biosensors Bioelectronics. 2014; 52: 345-353Crossref PubMed Scopus (4) Google Scholar] to assess intracellular ara-CTP concentration in circulating blasts over the initial 72 hours post-treatment. This was used to predict remission status in AML patients (n = 26) after treatment with low-, standard-, and high-dose (LD, SD, HD) cytarabine. Samples were gathered under informed consent (15/WM/0415) from non-M3 AML participants (≥18 years) receiving cytarabine therapy (n = 26) (Table 1). Patients received one of three dosing levels as judged by the consultant hematologist according to current UK guidelines (SD: 200 mg/m2 infused in divided doses twice daily; HD: 1.5–3 g/m2 infused in divided doses twice daily; LD: 20 mg twice daily by subcutaneous injection) [7Milligan DW Grimwade D Cullis JO et al.Guidelines on the management of acute myeloid leukaemia in adults.Br J Haematol. 2006; 135: 450-474Crossref PubMed Scopus (123) Google Scholar]. Fresh peripheral blood samples (≤4 mL) were collected pre-treatment (t = 0) and post-treatment (t=2, 4, 8, 12, 24, 48, and 72 hours ± 30 min) commencement, with standard-of-care assessment of remission/non-remission status (CR/NR) performed at day 28 as per current UK guidelines. Where a sampling time point coincided with an infusion time point, the blood sample was removed pre-infusion.Table 1Patient characteristicsCharacteristicCytarabine regimenLow doseStandard doseHigh dosen7136Median age, y (range)77 (63-79)67 (47-76)41 (22-55)Gender Male6101 Female135Ethnicity White, British6114 White, non-British002 Not known120FAB M0–1322 M2143 M4–5031 M6–7000 MDS/therapy-related MDS210 Not known130Cytogenetic risk group Favorable113 Intermediate661 Adverse052 Not known010Prior MDS430Presenting WBC count, × 109/L8.89 (2.26–88.4)8.76 (0.63–118.52)9.175 (2.98–24.48)Dose, mg20 (20–20)200 (160–220)4100 (3900–4600)Relevant co-medication Daunorubicin13/13 Mylotarg12/135/6 FLAG-IDA6/6 Tosedostat2/7 Lenalidomide2/7Body surface area, m22.11 (1.74–2.26)2.02 (1.64–2.22)2.02 (1.74–2.31)Bone marrow burden at presentation, %24.7 (3.8–91.8)42 (21.3–90.0)65.3 (11.0–85.0)Peripheral blasts at presentation, %8.6 (0.2–97.1)25.3 (0.7–88.0)54.2 (30–78.4)Day 28 bone marrow burden, %10.75 (0.1–72) 6Anderson E Conway M Alloush H et al.Investigation and verification of a bioluminescent biosensor for the quantitation of ara-CTP generation: A biomarker for cytosine arabinoside sensitivity in acute myeloid leukaemia.Biosensors Bioelectronics. 2014; 52: 345-353Crossref PubMed Scopus (4) Google Scholar0.90 (0.12–60) 12Elliott MA Litzow MR Letendre LL et al.Early peripheral blood blast clearance during induction chemotherapy for acute myeloid leukemia predicts superior relapse-free survival.Blood. 2007; 110: 4172-4174Crossref PubMed Scopus (57) Google Scholar1.75 (0.4–2.5) 5Koehl U Hollatz G Rohrbach E et al.Pharmacology of intracellular cytosine-arabinoside-5′-triphosphate in malignant cells of pediatric patients with initial or relapsed leukemia and in normal lymphocytes.Cancer Chemother Pharmacol. 2007; 60: 467-477Crossref PubMed Scopus (4) Google ScholarValues are expressed as number of patients or median (range).FLAG-IDA=fludarabine. cytarabine, granulocyte–macrophage colony-stimulating factor, idarubicin; FAB=French–American–British classification; WBC, white blood cell; MDS=myelodysplastic syndrome. Open table in a new tab Values are expressed as number of patients or median (range). FLAG-IDA=fludarabine. cytarabine, granulocyte–macrophage colony-stimulating factor, idarubicin; FAB=French–American–British classification; WBC, white blood cell; MDS=myelodysplastic syndrome. Peripheral blood mononuclear cells (PBMC) fractions were isolated from whole blood samples by density gradient centrifugation (Histopaque-1077, Sigma-Aldrich, Gillingham, UK) within 1 hour of withdrawal. PBMCs were pelleted at 300g (5 min), washed in RPMI-1640 medium (10 mL), re-suspended in red cell lysis buffer (0.15 mmol/L ammonium chloride, 0.01 mmol/L potassium bicarbonate, 0.001 mmol/L EDTA, pH 7.2–7.4) for 5 min, and washed in RPMI-1640 medium (without phenol red) (10 mL). Lysates were prepared and stored at –80°C until biosensor analysis as per Alloush et al. [8Alloush HM Anderson E Martin AD et al.A bioluminescent microbial biosensor for in vitro pretreatment assessment of cytarabine efficacy in leukemia.Clin Chem. 2010; 56: 1862-1870Crossref PubMed Scopus (15) Google Scholar]. leukemia-associated immunophenotype (LAIP)-positive absolute counts were determined on baseline peripheral blood and every 24 hours thereafter for 72 hours. Clearance of peripheral blasts was calculated by conversion of daily blast count to logarithmic scale, and subtraction from baseline. Ara-CTP stock solution (10 mmol/L, Jena Biosciences, Jena, Germany) was diluted (0–0.5 µmol/L) in cell lysate as per Alloush et al. [8Alloush HM Anderson E Martin AD et al.A bioluminescent microbial biosensor for in vitro pretreatment assessment of cytarabine efficacy in leukemia.Clin Chem. 2010; 56: 1862-1870Crossref PubMed Scopus (15) Google Scholar]. Limits of detection (LOD) and quantitation (LOQ) were calculated from the standard deviation of the blank (n = 6) as per Shrivastava et al. [9Shrivastava A Saxena P Gupta VB Spectrophotometric estimation of tamsulosin hydrochloride by acid-dye method.Pharm Methods. 2011; 2: 53-60Abstract Full Text PDF PubMed Google Scholar]. Results indistinguishable from background were reported as <LOD and removed from the analysis. Analysis performed used GraphPad Prism 7 (GraphPad, La Jolla, CA, USA) including two-way analysis of variance (ANOVA) with Tukey's (Figure 1A) and Sidak's post hoc tests and analysis of co-variance (ANCOVA) of body mass index (BMI) versus sensitivity index (SI%). Linear regression of log10 reduction in day 28 bone marrow (BM) blast burden versus diagnostic BM (flow cytometry) was compared with SI% values, with the Pearson coefficient (r) calculated at each time point assessable (SD cohort). Feasibility was calculated as the percentage of samples successfully taken versus patients (n) in the group. Any patient suffering treatment-related mortality was removed from subsequent feasibility assessments. Measurement of circulating blast intracellular ara-CTP concentration post-treatment (expressed as SI%) was possible in all treatment groups over the initial 24 hours (Figure 1A), with the greatest discrimination observed between HD (max n = 6) and SD (max n = 11) patients (p < 0.05 at all time points collected). Peripheral LAIP counts were sufficient for ara-CTP analysis from a single vacutainer (4 mL, ≥2 × 106 blasts) in all patient groups over the initial 24 hours (LD: 100%, SD: 80%, HD: 77%) (Supplementary Figure E1A, online only, available at www.exphem.org), with a dose-dependent decrease in LAIP over the initial 72 hours of therapy (Figure 1B). Limitations to assessment were high rates of nighttime sampling across all treatment groups; decreasing peripheral burden in HD patients (Figure 1B); a high proportion of LD patients with results below the limit of detection of the assay (4/7); and high rates of discharge of LD patients at 48 hours (1/7) and 72 hours (2/7) (Supplementary Figure E1A), as these patients can self-administer cytarabine once stabilized. For a proportion of patients (SD: max n = 9; HD: max n = 5), it was possible to calculate the intracellular ara-CTP concentration (ng/mL) (Figure 1C) using a standard curve across a therapeutically achievable ara-CTP range (0–500 nmol/L) (Supplementary Fig. E1B). The LOD for the assay was 1.58 ng/mL, and the LOQ was 4.81 ng/mL. HD patients achieved the Css of ara-CTP from t = 4 hours post-infusion, with a tmax at 4 hours, and a 111-fold higher area under the curve (AUC) compared with SD patients (Figure 1D). In comparison, SD patients achieved the Css 2 hours earlier but with a less prolonged duration (HD, 4 hours, SD, 2 hours) and 27-fold lower concentration than HD patients (Figure 1C). These results indicate similar trends and timings to previous assessment of ara-CTP by HPLC-MS/MS [10Liang D Wang W Jiang X Yin S Simultaneous determination of 1-β-d-arabinofuranosylcytosine and two metabolites, 1-β-d-Arabinofuranosyluracil and1-β-d-arabinofuranosylcytosine triphosphate in leukemic cell by HPLC-MS/MS and the application to cell pharmacokinetics.J Chromatogr B. 2014; 962: 14-19Crossref PubMed Scopus (8) Google Scholar], but required 1/10th the cell density necessary for HPLC analysis, albeit lower concentrations of ara-CTP (ng/mL) were observed herein. This represents the first quantitative analysis of post-treatment intracellular ara-CTP levels in AML patients using the biosensor, conducted within a working clinical environment. Initial analysis of prediction of outcome at day 28 versus ara-CTP accumulation was performed at each time point for the SD cohort (Figure 2A–E). The most striking correlation was observed at t = 4 hours post-treatment commencement, where a higher SI% correlated with day 28 log10 bone marrow blast reduction (Figure 2A) (n = 6, p = 0.0113). Correlation was possible in only 6 of 7 patients at this time point, as 1 patient was not assessable for burden at day 28 (Supplementary Figure E2, online only, available at www.exphem.org). A significant relationship was notable at t = 8 hours (n = 4, p = 0.0280) (Figure 2C); however the n value limits interpretation. No significant correlation was observed with t = 2 (Figure 2A) or t = 12 hours post-treatment commencement SI% (Figure 2D) or AUC0–12 hours (Figure 2E). For any future study, infusion time +4 hours may derive the most predictive time point as the majority of SD patients exhibited declining ara-CTP levels after the 4-hour time point (Figure 1C). Concurrent LAIP assessment (0–72 hours) versus day 28 outcome performed in SD patients (n = 7) (Figure 2F) replicated previous findings from day 14 outcome correlation versus circulating blast burden decline over the initial days post-treatment [11Gianfaldoni G Mannelli F Baccini M Antonioli E Leoni F Bosi A Clearance of leukaemic blasts from peripheral blood during standard induction treatment predicts the bone marrow response in acute myeloid leukaemia: A pilot study.Br J Haematol. 2006; 134: 54-57Crossref PubMed Scopus (45) Google Scholar, 12Elliott MA Litzow MR Letendre LL et al.Early peripheral blood blast clearance during induction chemotherapy for acute myeloid leukemia predicts superior relapse-free survival.Blood. 2007; 110: 4172-4174Crossref PubMed Scopus (57) Google Scholar]. Herein, non-standard flow sampling (t = 24, 48, and 72 hours) had a <50% mean completion rate (Supplementary Figure E1A), limiting correlation notably for HD and LD cohorts. Intervention with additional training of night staff improved the completion rate substantially after interim identification of the issue (0% vs. 62.5%).Figure E.2Feasibility data (standard dose patients). 1, lysate produced; 0, no lysate produced; <LOD, below limit of detection; *below limit of quantitation.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure E.3Feasibility data (high dose patients). 1, lysate produced; 0, no lysate produced; <LOD, below limit of detection; n.c., not calculable (no 12 hour result); w.d., withdrawn/no longer able to verbally consent (removed from feasibility assessment).View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure E.4Feasibility data (low dose patients). 1, lysate produced; 0, no lysate produced; <LOD, result below limit of detection; *patient discharged from hospital; MDS, myelodysplastic syndrome; AML, acute myeloid leukaemia.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Recruitment exceeded targets in all treatment groups (Supplementary Figure E5A, online only, available at www.exphem.org), with no patients electing to withdraw from the study; however, sampling was dependent on both time of treatment start (Supplementary Figure E5B) and, consequently, the proportion of sampling by day versus night staff (Supplementary Figure E5C). As expected, LD patients showed the highest proportion of start times between 7 AM and 7 PM, as advance notice of treatment commencement was possible in this group (typically 1 week). Unexpectedly, the majority of SD and HD patients commenced treatment between 7 PM and 1 AM (∼50%), with a significant proportion starting between 1 and 7 AM (SD: 8%, HD: 16%). This, in turn, led to a high rate of night sampling in these treatment groups (Supplementary Figure E5C) and challenges associated with training and continuity of night staff. As well as affecting sampling rates for flow cytometry, the presence of analytical staff frequently prompted sampling by clinical night staff, potentially inflating the true feasibility of peripheral sampling at night. A further study investigating key time points is planned, benefiting from the feasibility observations reported herein. Interestingly, high median body surface area (BSA) of patients was noted in this study (Table 1), with body mass index (BMI) indicating a large proportion of patients in the overweight or obese category (Supplementary Figure E5D). Whilst no significant association was observed between BMI and SI% herein, a higher BMI was associated with lower initial intracellular ara-CTP accumulation in SD patients (p = 0.0169, n = 10). No association was observed in the LD (n = 3) or HD (n = 6) cohort. This is an interesting finding given the argument for dose capping in obese patients [13Tavitian S Denis A Vergez F et al.Impact of obesity in favorable-risk AML patients receiving intensive chemotherapy.Am J Hematol. 2016; 91: 193-198Crossref PubMed Scopus (18) Google Scholar]. Inter-individual variation in ara-CTP levels has previously been identified as contributing to clinical variability in response [14Lamba JK Genetic factors influencing cytarabine therapy.Pharmacogenomics. 2009; 10: 1657-1674Crossref PubMed Scopus (93) Google Scholar]. Previous groups have attempted to retrospectively predict response using expression analysis of key genes (e.g., hENT1, deoxycytidine kinase) involved in cytarabine uptake, metabolism, and mitochondrial signalling [15Lamba JK Crews KR Pounds SB et al.Identification of predictive markers of cytarabine response in AML by integrative analysis of gene-expression profiles with multiple phenotypes.Pharmacogenomics. 2011; 12: 327-339Crossref PubMed Scopus (22) Google Scholar, 16Abraham A Varatharajan S Karathedath S et al.RNA expression of genes involved in cytarabine metabolism and transport predicts cytarabine response in acute myeloid leukemia.Pharmacogenomics. 2015; 16: 877-890Crossref PubMed Scopus (30) Google Scholar, 17Yan H Wen L Tan D et al.Association of a cytarabine chemosensitivity related gene expression signature with survival in cytogenetically normal acute myeloid leukemia.Oncotarget. 2017; 8: 1529-1540PubMed Google Scholar]. While valuable prediction tools, a phenotypic assessment of post-treatment ara-CTP accumulation, incorporating systemic and intracellular handling of cytarabine, could allow more rapid and specific treatment tailoring, identifying those at risk of both adverse reaction and treatment failure. The authors thank participants and staff at the Bristol Haematology & Oncology Centre for the operational management of this work. Patient samples were gathered under informed consent from participants at University Hospitals Bristol NHS Foundation Trust (15/WM/0415), approved by the West Midlands, South Birmingham Research Ethics Committee (UK). The Escherichia coli HA1 whole-cell biosensor is protected under patent (PCT/GB2009/00196) held jointly by Randox Laboratories Ltd (Belfast) and UWE. This work was funded by a grant from Above & Beyond. This work was funded by a charitable grant from Above and Beyond, Bristol.

Referência(s)