Cerebrospinal fluid asparagine depletion during pegylated asparaginase therapy in children with acute lymphoblastic leukaemia

Authors


Summary

L-asparaginase is an important drug in the treatment of childhood acute lymphoblastic leukaemia (ALL). Cerebrospinal fluid (CSF) asparagine depletion is considered a marker of asparaginase effect in the central nervous system (CNS) and may play a role in CNS-directed anti-leukaemia therapy. The objective of this study was to describe CSF asparagine depletion during 30 weeks of pegylated asparaginase therapy, 1000 iu/m2 i.m. every second week, and to correlate CSF asparagine concentration with serum L-asparaginase enzyme activity. Danish children (1–17 years) with ALL, treated according to the Nordic Society of Paediatric Haematology and Oncology ALL2008 protocol, standard and intermediate risk, were included. CSF samples were obtained throughout L-asparaginase treatment at every scheduled lumbar puncture. A total of 128 samples from 31 patients were available for analysis. Median CSF asparagine concentration decreased from a pre-treatment level of 5·3 μmol/l to median levels ≤1·5 μmol/l. However, only 4/31 patients (five samples) had CSF asparagine concentrations below the limit of detection (0·1 μmol/l). In 11 patients, 24 paired same day serum and CSF samples were obtained. A decrease in CSF asparagine corresponded to serum enzyme activities above 50 iu/l. Higher serum enzyme activities were not followed by more extensive depletion. In conclusion, pegylated asparaginase 1000 iu/m2 i.m. every second week effectively reduced CSF asparagine levels.

L-asparaginase is a cornerstone in the multi-drug treatment of childhood acute lymphoblastic leukaemia (ALL) (Pieters et al, 2011). It has been used in the treatment of ALL in different formulations, doses and treatment regimens for more than 30 years (Ertel et al, 1979). However, the right dose, formulation, and treatment regimen remain topics of active investigation (Asselin, 2012). Cerebrospinal fluid (CSF) asparagine depletion is presumed to be a marker of asparaginase effect in the central nervous system (CNS), and is the focus of this study.

L-asparaginase is an enzyme that hydrolyses asparagine to aspartic acid and ammonia and, to some extent, glutamine to glutamic acid and ammonia (Miller & Balis, 1969). In the extra cellular space this results in asparagine depletion and a reduction of glutamine. The lymphoblasts, which cannot effectively synthesize de novo asparagine, have their protein biosynthesis inhibited, and ultimately they undergo apoptosis (Asselin et al, 1989). The glutaminase activity possessed by the L-asparaginase enzyme enhances asparagine depletion (Panosyan et al, 2004). However, in the CNS glutamine is involved in complex metabolic pathways and disturbances in CSF glutamine concentration may possibly affect normal brain function (Albrecht et al, 2010). L-asparaginase is a large molecule and does not readily cross the blood brain barrier (Schwartz et al, 1970; Riccardi et al, 1981). Thus depletion of cerebrospinal fluid (CSF) asparagine presumably happens through diffusion.

Few studies have characterized pegylated asparaginase pharmacodynamics in the CSF of patients with de novo ALL (Avramis et al, 2002; Appel et al, 2003; Rizzari et al, 2006; Vieira Pinheiro et al, 2006). Insufficient CSF asparagine depletion is a concern, as the consequence may be ineffective killing of leukaemic cells in the CNS. Thus, it has been hypothesized that the CNS may serve as a sanctuary for leukaemic cells, and consequently this may increase the risk of CNS relapse despite other CNS-directed therapy (Avramis & Panosyan, 2005). In addition, truncated L-asparaginase treatment has been associated with increased risk of CNS relapse (Moghrabi et al, 2007; Sirvent et al, 2011).

In the Nordic countries, three bacterially derived L-asparaginase preparations are currently available: native L-asparaginase from Escherichia coli (Asparaginase Medac®; Medac GmbH, Hamburg, Germany), pegylated asparaginase from E. coli (Oncaspar®; Sigma-Tau Pharmaceuticals, Inc., Gaithersburg, MD, USA), and native erwinia asparaginase from Erwinia chrysanthemi [Erwinase®; EUSA Pharma (Europe) Ltd., Oxford, UK]. Pegylated asparaginase is the drug of choice in the current Nordic Society of Paediatric Haematology and Oncology paediatric treatment protocol (NOPHO ALL2008) for children and young adults below 45 years of age with de novo ALL. Pegylated asparaginase, 1000 iu/m2 per dose, is administered intramuscularly (i.m.). Compared with the other two L-asparaginase preparations pegylated asparaginase has the longest half-life (5·73 ± 3·24 d) and is thought to be the least immunogenic (Asselin et al, 1993).

The objective of this study was to describe the depletion of asparagine in CSF during 30 weeks of pegylated asparaginase therapy, 1000 iu/m2 i.m. every second week, and to correlate CSF asparagine concentration with serum L-asparaginase enzyme activity.

Materials and methods

Patient population

Danish children diagnosed with ALL, aged 1–17 years, and treated according to the standard- (SR) and intermediate risk (IR) groups of the NOPHO ALL2008 protocol were included. From January 2011 to February 2013, 31 patients entered the study. The study was approved by the Danish Data Protection Agency and the National Ethics Committee. Informed consent, according to the declaration of Helsinki, was obtained from the children′s parents or guardian.

NOPHO ALL-2008 treatment protocol

In the Nordic (Denmark, Norway, Sweden, Finland, Iceland) and Baltic (Lithuania, Estonia) countries all children diagnosed with ALL have been treated according to a common protocol which opened in 2008, (NOPHO ALL2008). The NOPHO ALL2008 therapy is previously described (Frandsen et al, 2014; Raja et al, 2014). Based on the characteristics of leukaemia at diagnosis and initial response to induction therapy [minimal residual disease (MRD) on days 15, 29] and to MRD on day 79 during consolidation, the patients are assigned to four risk groups, standard (SR), Intermediate (IR), high-risk-chemotherapy (HR-Chemo) or high-risk-stem cell transplant (HR-SCT).

In the SR and IR risk group treatment with intramuscular pegylated asparaginase, 1000 iu/m2 (Oncaspar®; Sigma-Tau Pharmaceuticals, Inc.; prior to August 2012: Medac GmbH) begins on day 30 (end of induction therapy), and continues every second week throughout the consolidation phase (five intramuscular injections of pegylated asparaginase). Prior to the delayed intensification, patients are randomized to receive pegylated asparaginase every second week (total of 15 doses, standard arm) or every 6th week (total of eight doses, experimental arm). Thus, all patients receive the first five pegylated asparaginase injections at 2-week intervals. Overall, the pegylated asparaginase treatment continues for 30 weeks.

Sample collection

During L-asparaginase treatment, CSF samples from 31 patients were obtained, when intrathecal therapy was scheduled. In 11 patients, blood samples were obtained simultaneously with CSF sample collection. Samples were drawn on ice, and stored at −80°C until analysis. In the SR group, CSF samples were collected on treatment days 29, 37, 58, 79, 122, 142 and 198 and in the IR group on days 29, 37, 58, 79, 92, 127, 156 and 212 after diagnosis. The time interval between pegylated asparaginase administration and CSF sample collection varied from 1 to 44 d. However, most samples were collected approximately 7 or 14 d after intramuscular administration of pegylated asparaginase. Samples collected on day 29 were obtained before pegylated asparaginase was administered and thus served as pre-treatment controls.

Amino acid analysis – asparagine and glutamine

Asparagine and glutamine levels in the CFS were analysed with ion pair liquid chromatography coupled with tandem mass spectrometric detection (Waters TQD; Waters Corporation, Milford, MA, USA). After mixing with internal standards (15N2 L-aspargine and 2,3,3,4,4 D5 L-glutamine, Cambridge Isotope Laboratories Inc., Tewksbury, MA, USA) and protein precipitation with perchloric acid, samples were separated with a 2·1 × 150 mm HSST-3 [High Strength Silica particles with C18 alkyl phase bonded using Water's trifunctional (T3) ligand technology] 1·7 μm column with a 2·1 × 5 mm Vanguard pre-column (Waters) using gradient water: acetonitrile mobile phases both with 0·1% perfluoroheptanoic acid. Calibrators in water were analysed in parallel. Multiple reaction monitoring (MRM) traces were 132·9–73·8/134·8–74·8 for asparagines and 146·9–83·8/151·8–88·9 for glutamines: cross-talk was observed, but with chromatographic separation a specific trace for each compound was obtained.

As an analyte-free biological matrix for spiking experiments could not be obtained, instrument sensitivity for aspargine was estimated from repeated injections (n = 5 pseudoreplicates) of four patient samples with low analyte levels. To account for variable matrix effects, lower limit of detection (LOD) and lower limit of quantification (LOQ) were estimated for each sample as 3 and 10 standard deviations of response (analyte: IS area ratio) divided by the slope of the standard curve. Mean (range), LOD and LOQ estimates (μmol/l) were 0·09 (0·08–0·11) and 0·33 (0·23–0·38). The asparagine levels in these samples [0·23 (0·15–0·30) μmol/ml], was used to determine an intra-day reproducibility [coefficient of variation (CV)] of 17 (8–24)%. All samples contained similar and high levels of glutamine. Indeed, injection volumes had to be reduced to 1 μl (vs. 3 μl for asparagine runs) to prevent chromatographic peak tailing. Hence, formal sensitivity testing was omitted and validation for glutamine analysis limited to inter-day variability.

Inter-day reproducibility was determined with 12 patient samples, each prepared and assayed on at least two separate days. From those samples the mean (range) CV for asparagine above LOQ was 2·9% (0·2–7·2). Mean (range) CV for glutamine was 1·9% (0·4–3·9). A subset of these samples was included as controls in all analytical runs.

Quantification of L-asparaginase activity using Nessler's reagent

Serum and CSF L-asparaginase enzyme activity was measured by a spectrophotometric method using Nessler's reagent as previously described (Albertsen et al, 2001), except that Erwinase standards were replaced with pegylated asparaginase spiked into blank CSF or serum [800, 600, 400, 200, 100, 50 and 25 international units/l (iu/l)]. One international unit is defined as the amount of enzyme required to catalyse the formation of 1 μmol of ammonia from asparagine in 1 min at 37°C. The inter-day assay CV, calculated from mean of triplicate analysis of the same sample (284 iu/l) 17 times over 1 year, was 7·4%. The intra-day assay CV, calculated from mean of triplicate analysis of the same sample (185 iu/l) ten times in 1 d, was 4·1%. The lower limit of detection was 25 iu/l.

Statistics

Descriptive statistics were calculated with SPSS statistical package for social sciences version 22.0 (SPSS, Chicago, IL, USA). Undetectable asparagine levels (below 0·1 μmol/l) were entered as 0 in calculations and graphs. Intervals in brackets after medians are observation ranges. Spearman′s rank was used for correlation calculations. Mann–Whitney U test was used for comparison of asparagine levels pre-treatment versus 1–15 d after pegylated asparaginase therapy. Given that some of the study subjects had an incomplete set of CSF samples, comparisons of mean glutamine concentrations were made by two sample (rather than paired) t-test assuming unequal variances. P-values, (two-sided) <0·05 were considered significant.

Results

Patient characteristics

We obtained 130 CSF samples from 31 patients; 17 males and 14 females. Median age was 4 years (1–16). Twenty patients were in the IR group and 11 in the SR group. Twenty-eight had B-cell precursor ALL and three had T-cell ALL. Patients contributed with a median of four samples (1–8). One patient (IR) had trisomy 21, and received an extra dose of pegylated asparaginase during induction on day 8 after diagnosis. Two samples from this patient obtained on day 15 and 29 were excluded, leaving a total of 128 samples for data analysis.

Asparagine depletion

The median CSF asparagine pre-treatment level was 5·3 μmol/l (2·2–12·0). This is within the normal range reported for healthy children this age (Gerrits et al, 1989). To estimate the degree and duration of CSF asparagine depletion, patient samples from different days of treatment were pooled, and divided into five time-point categories after the last given pegylated asparaginase dose (day 0, days 1–8, days 11–15, days 17–24, days 27–44) (Fig 1). Thus one patient could contribute with more than one sample at any time point. There was a marked decrease in CSF asparagine levels from 1 d after pegylated asparaginase treatment to approximately 3 weeks after treatment (days 17–24) (Fig 1). However, only 4/31 patients (a total of five samples) had CSF asparagine concentrations below 0·1 μmol/l (LOD) at any point during treatment. The raw plotted data on which Fig 1 is based is given in Figure S1A.

Figure 1.

Cerebrospinal fluid (CSF) asparagine concentration measured at different time points after administration of pegylated asparaginase. Number of samples (n) is indicated above each box-plot. Circles represent outliers and asterisks extreme outliers.

Most samples were collected between 1 and 15 d after pegylated asparaginase therapy (Fig 1). Based on the pharmacodynamics and pharmacokinetics of pegylated asparaginase we would expect asparagine depletion and asparaginase enzyme activity in serum for a minimum of 2 weeks (Avramis et al, 2002; Vieira Pinheiro et al, 2006). Thus, all following calculations on asparagine levels include only samples collected between 1 and 15 d after pegylated asparaginase therapy. In total, 91 samples from 29 patients were obtained within 15 d after pegylated asparaginase therapy. We found a statistically significant reduction in asparagine concentration between pre-treatment (n = 18) samples and pools (day 1–15) after pegylated asparaginase therapy (n = 91 samples), Mann–Whitney U test (< 0·001).

The distribution of CSF asparagine levels during asparaginase therapy is shown in Fig 2. The level of asparagine in most (83/91) samples were below 2·0 μmol/l. Four patients had samples with high asparagine concentration. Three of the four patients had corresponding serum samples from the same day + 2 d, all of which showed undetectable L-asparaginase activity. The median asparagine concentration measured at 10 time points varied from 0·6 to 1·5 μmol/l during treatment (Table 1). The total sample median for the entire treatment period was 1·0 μmol/l (<0·1–8·6). Figure 3 displays the distribution of CSF asparagine levels by patient. Except for four patients (marked with symbols in the figure), there was little inter-patient variability.

Table 1. Asparagine and glutamine concentration during treatment
Protocol dayNumber of samples (n)CSF asparagine (μmol/l)CSF glutamine (µmol/l)
Median (range)Mean (95% CI)
  1. Protocol day is days since diagnosis.

  2. a

    Protocol day 29 is pre-treatment. Except for the pre-treatment samples only samples obtained 1–15 d after last dose of pegylated asparaginase were included.

29a185·3 (2·2–12)459 (436; 481)
37171·1 (0·0–1·6)486 (442; 530)
58161·2 (0·1–8·6)460 (424; 496)
79170·8 (0·0–6·4)485 (443; 527)
92130·8 (0·0–6·3)469 (429; 508)
12230·9 (0·7–6·3)402 (146; 659)
12770·6 (0·2–0·9)465 (397; 533)
14251·5 (0·3–6·2)485 (342; 628)
15630·7 (0·3–1·0)501 (298; 703)
19841·3 (1·1–6·3)479 (312; 647)
21260·9 (0·2–1·1)539 (430; 648)
Figure 2.

Cerebrospinal fluid (CSF) asparagine concentration measured throughout asparaginase treatment. Each circle represents a sample. Sample outliers are marked with squares, triangles, crosses and asterisks with each symbol representing a patient. Same day (+2 d) serum samples were available for patients represented by squares, asterisks and crosses, and showed no detectable enzyme activity. No serum was available from the patient represented by triangles. Except for pre-treatment samples, only samples collected 1–15 d after the last pegylated asparaginase dose were included. * Protocol day 29 is pretreatment.

Figure 3.

Cerebrospinal fluid (CSF) asparagine concentration by patient. Each circle represents a sample. Sample outliers (patients 26, 27, 28 and 29) are marked with asterisks, crosses, squares and triangles as in Fig. 2. Dashed line shows the total median. Only samples obtained 1–15 d after pegylated asparaginase were included.

Glutamine concentration

Cerebrospinal fluid glutamine levels were within the normal range for healthy children (Gerrits et al, 1989) and no decrease was observed during treatment. No correlation between glutamine levels and days after pegylated asparaginase administration (0–44 d) were found (Spearman′s rank correlation coefficient rs = −0·120; = 0·178).

For samples obtained 1–15 d after last pegylated asparaginase, the mean pre-treatment and on treatment glutamine concentration was 459 μmol/l (95% CI, 436; 481; n = 18) and 478 μmol/l (95% CI, 461; 495; n = 91), respectively. There was no statistically significant difference between CSF glutamine levels pre-treatment versus on treatment, (t-test, = 0·17). CSF mean glutamine levels during treatment are shown in Table 1.

CSF asparagine concentration correlated to serum L-asparaginase enzyme activity

Eleven patients (24 samples) had both CSF and serum samples taken on the same day (samples obtained day 7–43 after pegylated asparaginase injections). Analysis of these samples showed a reduction in CSF asparagine corresponding to serum enzyme activity of more than 50 iu/l. A high serum enzyme activity was not followed by a total depletion, but a reduction in the same range, as seen with an enzyme level of 50 iu/l (Fig 4).

Figure 4.

Serum L-asparaginase enzyme activity, (iu/l), correlated with cerebrospinal fluid (CSF) asparagine level (μmol/l). Each circle represents a sample. CSF and serum samples (n = 24) were taken on the same day, n = 11 patients.

To verify the assumption that pegylated asparaginase does not cross the blood-brain barrier we also performed enzyme analyses of the CSF samples. As expected, we found no enzyme activity (95 samples analysed).

By December 2013 no patients had developed a CNS relapse (median follow up time 1·7 (0·9–3·7) years after diagnosis).

Discussion

As different L-asparaginase formulations have different pharmacokinetic and pharmacodynamic profiles, an appropriate treatment schedule and dose is extremely important to ensure asparagine depletion in serum and CSF (Albertsen et al, 2001; Panetta et al, 2009). However the threshold level of CSF asparagine depletion necessary to achieve anti-leukaemic activity in the CNS is not clear. The first step to approach this problem is the characterization of CSF asparagine depletion. The strength of this study is the collection of CSF samples during the entire treatment, i.e., after several doses of pegylated asparaginase. Thus, we were able to study the longitudinal depletion of CSF asparagine through a longer treatment period.

We found a marked decrease in CSF asparagine for approximately 3 weeks after the administration of pegylated asparaginase 1000 iu/m2 per dose. CSF asparagine depletion (median CSF asparagine concentration of 1 μmol/l) lasted throughout 30 weeks of pegylated asparaginase exposure (at 2-week intervals). In the correlation between serum enzyme activity and CSF asparagine levels we demonstrated a reduction in CSF asparagine corresponding to a serum enzyme activity above 50 iu/l. At this level a plateau was reached and complete depletion was not seen even at almost 20-fold higher serum enzyme activity (Fig 4). Thus our data strongly suggest that a higher dose of pegylated asparaginase would not lead to an enhanced depletion. In Figs 2 and 3 we marked samples from four patients with outliers. In three of four patients, serum enzyme activity was known to be undetectable explaining well the high asparagine level. The outliers appeared patient specific.

When comparing this study with the results of other study groups we wish to address four issues of importance for the degree of depletion, i.e., dose, route of administration, treatment schedule and asparaginase formulation.

Dose and route of administration

Two studies that used intravenous pegylated asparaginase, 1000 iu/m2 up-front demonstrated a degree of CSF asparagine depletion in agreement with our results (Appel et al, 2003; Rizzari et al, 2006). In the first study (Appel et al, 2003), CSF asparagine was measured after one dose of i.v. pegylated asparaginase given prior to induction. At five and 19 d after pegylated asparaginase administration they found that CSF asparagine had decreased, from a pre-treatment level of 5·1–1·58 μmol/l and 2·2 μmol/l, respectively. These results are very close to our findings (Table 1 and Fig 1). However, we found low levels of asparagine over a longer period after asparaginase treatment compared to the results reported Appel et al (2003). In the second study (Rizzari et al, 2006), two doses of pegylated asparaginase i.v. were given in induction and one in delayed intensification. The authors reported a marked decrease in CSF asparagine with the lowest mean level 2 d after the second dose in induction (1·08 μmol/l) and a slight increase again after 11 d (2·6 μmol/l). The shorter duration of CSF asparagine depletion after i.v. pegylated asparaginase in both of these studies compared to our results is most probably due to the different routes of administration.

Prior to these study reports, similar results had been reported from the Children's Cancer Group 1962 protocol (Avramis et al, 2002), only with a higher dose of pegylated asparaginase (2500 iu/m2) administered i.m. On this treatment CSF asparagine concentration fell from a median pre-treatment level of 2·3–1·1 μmol/l 4 d after pegylated asparaginase and 0·6 μmol/l at 25 d after pegylated asparaginase. This is in agreement with our results (Fig 1) but slightly lower for a longer time. The longer lasting depletion may be attributed to the higher dose and the lower levels may possibly reflect differences in methods of amino acid analysis as both their pre- and on-treatment levels are lower than those in the present study.

If the goal is complete asparagine, depletion defined as <0·2 μmol/l (LOD in many studies) it appears that there is a lack of depletion, regardless of dose or route of pegylated asparaginase administration. More interesting though is the level of depletion necessary to avoid CNS relapse.

Treatment schedule and asparaginase formulation

In this study we chose to include only CSF samples obtained 1–15 d after pegylated asparaginase treatment in the analysis of CSF depletion at different time points (Fig 3) for two reasons. First, based on the pharmacokinetics and pharmacodynamics of pegylated asparaginase we would assume sufficient enzyme activity in serum and a reduction in asparagine 1–15 d after the last dose of pegylated asparaginase (Avramis et al, 2002; Vieira Pinheiro et al, 2006). Secondly, the treatment interval in the protocol was every 2 or 6 weeks. Study reports of CSF asparagine concentration during treatment with native E. coli are not in agreement, but may trend towards a higher degree of depletion (LOD <0·2 μmol/l) than what is found with pegylated asparaginase (Ahlke et al, 1997; Woo et al, 1999; Avramis et al, 2002; Vieira Pinheiro et al, 2006). The same trend applies to Erwinase, though the studies are small and treatment schedules differ (Dibenedetto et al, 1995; Gentili et al, 1996). Again we cannot rule out small differences in methods of amino acid analyses, making comparisons difficult.

The small number of patients and thus few samples at some time points is a limitation of our study. However with the quite clear results, the reduction in CSF asparagine may be representative for a larger population. Furthermore, it adds to the results of others using pegylated asparaginase in different doses, and routes of administration (Avramis et al, 2002; Appel et al, 2003; Rizzari et al, 2006).

In this study we demonstrated that i.m. pegylated asparaginase 1000 iu/m2 per dose effectively reduced CSF asparagine levels in the vast majority of patients. Whether that translates into improved CNS outcomes was not addressed as the sample size was small and the follow-up too short. Two large studies, both of which used other CNS-directed therapy (i.e. methotrexate), have reported L-asparaginase therapy to be associated with CNS outcomes (Moghrabi et al, 2007; Sirvent et al, 2011). However CSF asparagine was not measured in any of these studies. Additional and large studies are needed to determine the association between poor CSF asparagine depletion and the risk of CNS relapse. It is an intriguing thought that this parameter could be used for more individualized therapy, assuming that the accuracy and precision of the amino acid determination is validated.

In conclusion we found that pegylated asparaginase 1000 iu/m2 i.m. lead to a marked decrease in CSF asparagine levels during the entire treatment period. However, only 4/31 patients (five samples) demonstrated complete depletion, i.e., below 0·1 μmol/l (LOD). The correlation between CSF asparagine and serum enzyme activity demonstrated that CSF asparagine decreased to a plateau level around a serum enzyme activity of 50 iu/l. Higher serum enzyme activities did not provide more extensive CSF asparagine depletion. This in, combination with the reports of other studies, strongly suggests that a higher dose of pegylated asparaginase will not provide better CSF asparagine depletion.

Acknowledgements

We sincerely thank the patients and their families for participating in this study. Furthermore, we thank laboratory technician Jane Hagelskjær Knudsen for her valuable assistance with the CSF and serum L-asparaginase enzyme analyses.

Author contributions

LTH had primary responsibility for manuscript preparation and data analysis. LTH, BKA and HS were responsible for study design and analysis of data. SR, TLF, RAR, HS, BKA and LTH participated in the acquisition of data. JN performed the amino acid analyses. All authors took part in manuscript preparation and approval of the final manuscript.

Conflict of interest

The authors declare no conflict of interest.

Funding

The study was financially supported by: The Danish Childhood Cancer Foundation, The Danish Cancer Society and the M.L. Jørgensen and Gunnar Hansen Foundation.

Ancillary