Activity and toxicity of intramuscular 1000 iu/m2 polyethylene glycol‐E. coli L‐asparaginase in the UKALL 2003 and UKALL 2011 clinical trials

Summary In successive UK clinical trials (UKALL 2003, UKALL 2011) for paediatric acute lymphoblastic leukaemia (ALL), polyethylene glycol‐conjugated E. coli L‐asparaginase (PEG‐EcASNase) 1000 iu/m2 was administered intramuscularly with risk‐stratified treatment. In induction, patients received two PEG‐EcASNase doses, 14 days apart. Post‐induction, non‐high‐risk patients (Regimens A, B) received 1–2 doses in delayed intensification (DI) while high‐risk Regimen C patients received 6–10 PEG‐EcASNase doses, including two in DI. Trial substudies monitored asparaginase (ASNase) activity, ASNase‐related toxicity and ASNase‐associated antibodies (total, 1112 patients). Median (interquartile range) trough plasma ASNase activity (14 ± 2 days post dose) following first and second induction doses and first DI dose was respectively 217 iu/l (144–307 iu/l), 265 iu/l (165–401 iu/l) and 292 iu/l (194–386 iu/l); 15% (138/910) samples showed subthreshold ASNase activity (<100 iu/l) at any trough time point. Older age was associated with lower (regression coefficient −9.5; p < 0.0001) and DI time point with higher ASNase activity (regression coefficient 29.9; p < 0.0001). Clinical hypersensitivity was observed in 3.8% (UKALL 2003) and 6% (UKALL 2011) of patients, and in 90% or more in Regimen C. A 7% (10/149) silent inactivation rate was observed in UKALL 2003. PEG‐EcASNase schedule in UKALL paediatric trials is associated with low toxicity but wide interpatient variability. Therapeutic drug monitoring potentially permits optimisation through individualised asparaginase dosing.


K E Y W O R D S
acute lymphoblastic leukaemia, children, PEG-asparaginase I N TRODUC TION L-asparaginase (ASNase) is a critical drug in the treatment of acute lymphoblastic leukaemia (ALL). Intensive ASNase treatment in contemporary ALL treatment protocols is associated with improved survival outcomes. 1 Therapeutic formulations of ASNase are sourced from bacteria and intensive administration is limited by short half-lives and immune inactivation. Polyethylene glycol conjugation of E. coli-derived ASNase (PEG-EcASNase) extends the half-life of the native E. coli enzyme fourfold (to an average 5.5 days). 2 PEG-EcASNase is substantially less immunogenic than its native counterpart, decreasing two-to tenfold the risk of hypersensitivity observed with intensive use of the native enzyme 3,4 and is the ASNase formulation of choice in contemporary ALL treatment protocols.
Despite three decades of worldwide use of PEG-EcASNase in ALL protocols, considerable variability is observed in dose, schedule, and route of administration of the drug. Treatment schedules vary, with PEG-EcASNase administered either intermittently or continuously (10-15 doses at two-week intervals) 3,5-7 as part of intensive dose schedules. The dose of PEG-EcASNase used varies from 1000 to 2500 iu/m 2 or more. [8][9][10][11][12][13] More recently, individualised PEG-EcASNase dosing 14 based on therapeutic drug monitoring has been proposed as a cost-effective strategy 15 to address the considerable intra-and interpatient variability observed with fixed-dose schedules.
In the UK, PEG-EcASNase was introduced as part of risk-adapted therapy in the UKALL 2003 and UKALL 2011 trial protocols for treatment of newly diagnosed ALL. PEG-EcASNase in both trials was administered intramuscularly at a unit dose of 1000 iu/m 2 . In both trials, an accompanying substudy monitored ASNase-associated toxicity and therapeutic drug activity and reported serological reactivity to asparaginase. In this report, we present and discuss findings from the asparaginase monitoring studies.

PATIE N TS A N D M ETHODS
ASNase monitoring was performed as part of substudies within the trial protocols, UKALL 2003 (ISRCTN07355119; October 2003-June 2011) and UKALL 2011 (ISRCTN64515327; April 2012-December 2018), for treatment of children and adolescents (1-24 years old) with newly diagnosed ALL. Enrolled patients were treated with risk-adapted chemotherapy regimens of increasing intensity (Regimens A, B, C) administered in five sequential treatment phases [induction, consolidation, interim maintenance, delayed intensification (DI), maintenance]. Trial findings have been reported previously [16][17][18][19] and details of risk stratification are provided in the Data S1.

Asparaginase treatment
In both clinical trials, PEG-EcASNase was administered intramuscularly at 1000 iu/m 2 /dose. Patients treated on Regimens A/B received two doses in induction (treatment days 4 and 18) and one dose in DI (treatment day 4). An additional post-induction dose was administered in patients randomised to a second DI in UKALL 2003. Regimen C patients received two additional doses each in consolidation (treatment days 16 and 44), interim maintenance (treatment days 3 and 23) and delayed intensification (DI) (treatment days 4 and 43). Regimen C patients in UKALL 2003 received a total 12 doses of PEG-EcASNase (induction, two; post-induction, ten, including two doses during each of two interim maintenance and DI blocks) and in UKALL 2011, a total of eight doses (induction two; post-induction, six) (Table S1).  (Table S1), testing for ASNase-associated antibodies (UKALL 2003 substudy alone) and observation for ASNaseassociated toxicities [Common Terminology Criteria of Adverse Events (CTCAE) v4.0 grade ≥3], of clinical hypersensitivity, thrombosis and pancreatitis. Approval for the substudies was obtained as part of ethics approval for the clinical trials. A pragmatic sampling strategy was used for monitoring post-treatment ASNase activity, with sample collections timed to coincide with venous access for other clinical indications. ASNase activity assays were performed centrally (University of Manchester) using the aspartate-βhydroxamate/indooxine method reported previously. 21,22 Testing for ASNase-associated antibody was performed using indirect enzyme-linked immunosorbent assays to detect antibody reactivity in plasma to PEG-EcASNase alone, to E. coli ASNase (EcASNase) alone or to both, using the assay protocol reported by the Dutch Childhood Oncology Group ALL-10 asparaginase study team. 6

Study definitions
Plasma samples obtained 14 ± 2 days following PEG-EcASNase treatment were considered informative i.e. suitable for trough ASNase activity measurements. 10 Trough activity levels of 100 iu/l or more were considered to represent satisfactory ASNase activity. 23 When analysed by treatment phase, ASNase activity was deemed adequate if satisfactory trough activity was observed with PEG-EcASNase treatment in induction (with either one or both PEG-EcASNase doses) and DI (in case of two courses, with the latter course). Asparaginase-associated antibody reactivity was reported as either ASNase-reactive (reactive to both PEG-and native EcASNase) or PEG-reactive (reactive to PEG-EcASNase but not to native EcASNase). Silent hypersensitivity referred to all patients with ASNase-associated antibody reactivity alone, without clinical hypersensitivity. Silent inactivation referred to the subset of patients with silent hypersensitivity who experienced a concomitant decline in ASNase activity to subthreshold levels (<100 iu/l).

Statistics
Continuous variables are represented as median [with interquartile range (IQR)] values. Groups with continuous variables were compared using the Mann-Whitney or Kruskall-Wallis tests as appropriate. Categorical variables were compared using the chi-squared or the Fisher exact tests as appropriate. The influence of covariates (age, sex, sampling time point, treatment regimen, substudy) on serial ASNase activity measurements was analysed using the generalised estimating equations model. 24 This approach allows handling of repeated measures that contain missing observations 25 and the analysis used an exchangeable correlation structure that assumes a fixed correlation for all pairs of repeated measurements. Modelling was performed combining observations from both substudies as well as separately for each substudy, in each case with and without considering two-factor covariate interactions Data (S1). Statistical significance for all analyses was set at p ≤ 0.05. Analysis using generalised estimating equations was performed using the R software programme (https:// www.r-proje ct.org). Other analyses were carried out using the SPSS statistical package (v23.0, IBM Corp) and represented graphically using the GraphPad Prism software (v9.2, GraphPad Software).

R E SU LTS
A total of 1112 patients were enrolled in the ASNase substudies (UKALL 2003, 423; UKALL 2011, 689). Patient cohorts in the substudies were matched in key prognostic characteristics, including age, sex, immunophenotype, presentation white-blood-cell count, cytogenetics, and minimal residual disease risk (MRD) groups ( Table 1). The significantly lower proportion of Regimen C patients in the UKALL 2003 substudy (25% vs. 42% in the UKALL 2011 substudy) arose from the randomised allocation in UKALL 2003 to treatment intensification (Regimen C) versus continuation on Regimens A/B in patients with high end-of-induction MRD levels (day 29 MRD ≥0.01%).

Trough asparaginase activity is satisfactory in most patients
Based on the risk group distribution and sampling protocol in each substudy, the targeted number of plasma samples for measurement of trough ASNase activity was 4363 (UKALL 2003, 1743; UKALL 2011, 2620) (   (Table 2). Lowering the threshold trough activity to 50 iu/l halved the proportion of samples with subthreshold ASNase activity at TP1-IND (7%, from 15%) but did not substantially affect subthreshold proportions at TP2-IND and TP-DI time points (11% vs. 16%, TP2-IND; 11% vs. 14%, TP-DI).
In 116 (10%) of 1112 substudy patients, serial trough plasma ASNase activity measurements in the induction and post-induction treatment phases were summarised and categorised as 'adequate' or 'inadequate' (Table 3). Sustained adequate ASNase activity was observed in 94 (81%) patients. Seven (6%, including five Regimen C) patients with adequate ASNase activity in induction experienced inadequate ASNase activity post induction, possibly suggesting silent immune inactivation. In seven (6%) other patients, trough ASNase activity was persistently inadequate in both treatment phases. Of note, eight (7%) with inadequate trough ASNase activity in induction experienced adequate trough ASNase activity post induction, without switch to an alternative ASNase formulation.
Opportunistic sampling meant that in some patients, samples for ASNase activity measurement were obtained prior to trough time points (i.e. days 7-11 post dose). These pre-trough measurements were combined with trough time-point estimations (post-dose days 12-16) to develop time-course plots of ASNase activity. Figure 1

Age and treatment phase influenced ASNase activity
The influence of select covariates (age, sex, assay time point, treatment regimen) and their interactions on serial ASNase activity measurements was modelled using generalised estimating equations (Table S3). In both ASNase substudies, age significantly influenced ASNase activity. Older age was associated with lower ASNase activity, indicated by the negative regression coefficient for age in both the UKALL 2003 (−9.45; p = 0.00045) and the UKALL 2011 (−9.57; p < 0.0001) substudies. In the UKALL 2011 substudy alone, ASNase activity levels were influenced significantly by the assay time point, with significantly higher levels observed post induction (regression coefficient 28.21; p < 0.0001) ( Table 4).  (Table S4). Rates of other significant ASNase-associated toxicities (pancreatitis, thrombosis) were low and did not differ between the two substudies ( ASNase-associated antibody testing performed in the UKALL 2003 substudy alone identified antibody reactivity in 10 (71%) of 14 patients with clinical ASNase hypersensitivity, including eight with ASNase-reactivity ( Figure 2). Sixteen (11%) of 149 patients without clinical ASNase hypersensitivity who underwent testing were antibody-reactive, 10 (7%) of whom showed concomitant decrease in serial ASNase activity, indicating silent ASNase inactivation. In three patients with sustained ASNase activity, PEG-directed reactivity alone was observed.

DISCUS SION
The asparaginase monitoring substudies indicate that the PEG-EcASNase treatment regimen in UKALL 2003 and UKALL 2011 was associated with satisfactory trough ASNase activity in ~85% of evaluable patients through the induction and post-induction phases of treatment. The proportion with satisfactory trough activity in the UKALL cohorts is lower than that reported in other ASNase monitoring studies (Table S5). [9][10][11]14,26,27 In the two studies that also administered PEG-EcASNase at 1000 iu/m 2 , treatment schedules, administration routes and assay methodologies varied, making direct comparisons difficult. The Nessler and MAAT (Medac asparaginase activity test) assays have been reported to overestimate asparaginase activity. 28,29 The increase in activity with time 5,30-32 and age 31 has been observed previously in other studies. An important study limitation was the proportion of targeted post-dose samples that was either not collected (44%) or collected at non-trough time points (61%) ( Table S2). The practice of minimum sampling, seeking to match sample collection with routine clinical care and venous access, accounted for this shortcoming and is a real-world challenge when conducting multicentre research studies of this nature in paediatric patients. 33 Rates of ASNase-associated hypersensitivity and other associated toxicities were low, especially compared to clinical trials using higher and more frequent doses of PEG-EcASNase. 26,27,30,34 Higher activity levels are reported to be associated with increased toxicity 30 and as reported previously and observed in this study, the more frequent administration of PEG-EcASNase is associated with an increase in hypersensitivity rates (Table S5). We speculate that the higher hypersensitivity rate in the UKALL 2011 substudy is related potentially to the randomised steroid treatment in induction, where patients with the shorter dexamethasone pulse received the second dose of PEG-EcASNase administered without steroid cover. 35 Higher post-induction ASNase activity, particularly significant in the UKALL 2011 substudy, has been reported previously 14,31 but its basis is uncertain. Disease-related factors 36 could potentially accelerate ASNase clearance in induction and account for lower ASNase activity during this treatment phase.  Induction, adequate: ASNase activity ≥100iu/l at any or both induction trough time points.
The survival implication of the comparatively lower rates of therapeutic ASNase activity observed in the UKALL substudies is uncertain and will be examined when follow-up matures in the UKALL 2011 trial cohort. Of note, introduction of PEG-EcASNase in UKALL 2003 was considered a key contributor to improved survival outcomes observed in the trial. 18 Equally, inadequate exposure to ASNase, either from premature discontinuation due to toxicity, 34 silent inactivation 37 or treatment with a substandard product, 38 has been reported to be associated with poorer outcomes. 1 The use of generalised estimating equations to examine the influence of covariates on serial measurements allowed identification of age and treatment phase as factors that independently influenced ASNase activity in the UKALL substudies. Treatment phase variability (with higher clearance of PEG-EcASNase during the induction phase) and age are now included as covariates in population pharmacokinetic models of PEG-EcASNase. 31 Standard fixed doses of PEG-EcASNase are associated with high asparaginase activity levels 14 and intensive fixed-dose schedules may result in higher rates of hypersensitivity (Table S5) and other ASNase-associated toxicities. 30 The threshold for therapeutic asparaginase activity  (defined variously as ≥100 iu/l or less), 1 can be achieved in the majority of patients with doses as low as 450 iu/ m 2 when accompanied by therapeutic drug monitoring. 32 Collectively, these observations support the argument for introducing therapeutic drug monitoring using a standardised assay to individualise the dose and choice of ASNase formulation for the treatment of ALL, 4

C ON F L IC T OF I N T E R E S T S
Medac GmbH provided drug and standards for the laboratory assays and performed the antibody studies. Servier provided funding for laboratory studies. Vaskar Saha is a recipient of speaker and consultancy fees from Medac GmbH and Servier.

AU T HOR C ON T R I BU T ION S
The funders and sponsors of the study had no role in study design, data collection, data analysis, data interpretation, writing of the report, or the decision to submit the paper for publication. All authors approved the final version of the manuscript. Jasmeet Sidhu: data curation, formal analysis, writing-original draft preparation, visualisation, writing:review and editing. Ashish Narayan Masurekar: conceptualization, formal analysis, investigation, methodology, data curation, validation, writing: original draft preparation. Manash Pratim Gogoi: methodology, data curation, formal analysis, visualisation. Caroline Fong: investigation, methodology, data curation, validation.