Unsatisfactory quality of E. coli asparaginase biogenerics in India: Implications for clinical outcomes in acute lymphoblastic leukaemia.

BACKGROUND: The biotherapeutic asparaginase is a cornerstone of therapy in acute lymphoblastic leukaemia (ALL). With limited access to the original native Escherichia coli-derived asparaginase (EcASNase), a variety of EcASNase biogenerics are used in low-middle-income countries (LMICs). The variable quality of these biogenerics potentially influences clinical outcomes. PROCEDURE: Seven biogeneric EcASNases (P1-P7) marketed widely in India were evaluated, with P2 as an exemplar for in vivo monitoring. Therapeutic activity of P2 (10,000 IU/m2 /dose, intramuscular, every 72 hours) was monitored during induction therapy, and drug-related toxicities recorded. Molecular identity, purity and in vitro drug activity of seven biogenerics were characterised using multimodal analyses, and findings compared with reference EcASNase (R).
RESULTS: In patients (N = 62) receiving P2, subtherapeutic asparaginase activity (<100 U/L) was observed in 66% (46/70) of trough timepoints (72 hours postdose) during induction. Twelve patients (19%), 11 with high-risk ALL, developed hypersensitivity. Isoforms of EcASNase were identified in all seven biogenerics. All generic products contained impurities with batch-to-batch variability. These included high levels of protein aggregates and host cell protein contamination. In vitro assays of EcASNase activity and leukaemia cell line cytotoxicity were not discriminatory.
CONCLUSIONS: Our findings confirm widespread concerns over the unsatisfactory quality and therapeutic activity of native EcASNase biogenerics marketed in LMICs. Appropriate use of these products requires monitored studies to identify clinical suitability and determine appropriate dosing and schedule. For large parts of the world, assured access to high-quality asparaginases remains an unmet therapeutic need.

Introduction

Acute lymphoblastic leukaemia (ALL) is the commonest cancer of childhood. In high-income countries, contemporary treatment protocols achieve cure rates approaching 90%.[1-3] Remarkably, this progress has been largely achieved using cytotoxic drug combinations discovered nearly half a century ago. The drug asparaginase (ASNase) is a critical component of multidrug treatment in ALL. Unique among cytotoxic drugs, ASNase is a therapeutic enzyme. The drug catalyses the breakdown of asparagine and to a lesser extent, glutamine—the resulting systemic depletion of these amino acids is relatively toxic to ALL lymphoblasts.[4,5] The pharmacodynamic activity of ASNase thus depends on achieving sustained systemic depletion of asparagine. Current therapeutic formulations of ASNase are derived from bacterial strains of Escherichia coli and Erwinia species. ASNase extracted from E. coli (EcASNase) is the formulation used most widely in clinical practice. The reference EcASNase formulation (Kyowa Hakko, Japan) is derived from a modified high asparaginase-producing strain of E. coli. EcASNase has a short half-life (22–38 hours) and requires frequent dosing to maintain sustained asparagine depletion.[6] As EcASNase is a non-human protein, repeated dosing of the drug can elicit a hypersensitivity response. The hypersensitivity may be subclinical or clinically overt, and is associated with ASNase-reactive antibodies that antagonise drug activity.[7]

Polyethylene glycol (PEG) conjugation of EcASNase substantially addresses the therapeutic limitations of the native molecule. The PEG-conjugated formulation (PEG-EcASNase) has a longer half-life (85–190 hours), requiring less frequent administration (typically every 14 days).[8] The longer half-life of PEG-EcASNase ensures sustained amino acid depletion for extended durations, necessary for the drug’s therapeutic activity.[6] PEG-EcASNase is also associated with lower rates of hypersensitivity.[9,10] The superior pharmacological attributes of PEG-EcASNase permits intensification of ASNase therapy, making it the formulation of choice in affluent economies. The switch to PEG-EcASNase has resulted in a decline in the supply of the standard formulation of native EcASNase worldwide.

PEG-EcASNase is expensive. Erwinia-derived ASNase is even more expensive, has a shorter half-life than EcASNase (13–22 hours), and its use is restricted to patients with hypersensitivity to EcASNase.[7] Additionally, there is a global shortage of Erwinia ASNase. Thus in low- and middle-income countries (LMICs), native EcASNase is the only affordable ASNase formulation. In these countries, limited availability of the reference native EcASNase product has led to the manufacture and supply of a number of EcASNase biogenerics. These biogenerics are not necessarily therapeutically equivalent to the reference product.[6,11-13] The E. coli source of ASNase in these products are different, and the resulting EcASNase have amino acid sequence variations leading to variations in their substrate affinity, enzyme activity, stability and antigenicity. Differences in manufacturing processes also influence the quality of EcASNase biogenerics and their therapeutic activity.

Expanding on these observations, this paper reports the findings of a comprehensive examination of the physical, chemical, and in vitro biological characteristics of seven biogeneric EcASNases marketed widely in India. We then focus on one of these formulations, investigating asparaginase activity serially in a cohort of patients during the induction treatment phase and examining rates of drug-associated toxicity.

Methods

Patients and asparaginase treatment

Between 1 February 2017 and 31 March 2018, consecutive patients aged 1–18 years with newly diagnosed ALL were enrolled in the study. Patients were stratified into risk groups (T-cell ALL; standard-, intermediate- and high-risk B-cell precursor ALL) based on patient characteristics, disease features and treatment response. Treatment was based on the Indian Collaborative Childhood Leukaemia Group Study ALL 2014 (ICiCLe-ALL-14) treatment protocol (Clinical Trials Registry-India CTRI/2015/12/006434). In all risk groups, EcASNase was administered intramuscularly every 72 hours at 10,000 IU/m2/dose. All patients received EcASNase during induction (four doses in standard-risk ALL, eight doses in others) and delayed intensification treatment phases (four doses in all risk groups). Patients with high-risk B-cell precursor ALL and T-ALL received additional EcASNase during the consolidation phase (eight doses). Through the study period, one brand (P2) of native EcASNase was administered uniformly according to the protocol. Plasma asparaginase activity was evaluated in peripheral blood samples collected 48 and 72 hours after EcASNase administration during the induction treatment phase. The study was approved by the institutional review board and patients/families gave written informed consent for participation.

Analysis of asparaginase activity in vivo

Peripheral blood samples collected for measuring asparaginase activity were maintained at room temperature and processed within 24 hours.

The resulting plasma aliquots were stored immediately at −80°C and tested in batches. Asparaginase activity in banked plasma samples was analysed using a modified indooxine assay.[14] Replicates of plasma aliquots were incubated with excess L-aspartic acid ß-hydroxamate (AHA; Sigma) at 37°C in 96-well microtitre plates. Hydroxylamine generated from ASNase hydrolysis of AHA was condensed with 8-hydroxyquinoline (Sigma) and the resulting indooxine quantified spectrophotometrically at 710 nm (Spectramax, Molecular Devices). A trough asparaginase activity ≥100 IU/L was considered adequate.[7] The dynamic linear range of the assay was 100–1000 IU/L. For ASNase activity levels below the dynamic range, plasma samples were retested using a modification of the assay with a dynamic range of 5–100 IU/L. The assay standard curve was developed using serial dilutions of the reference asparaginase in pooled human plasma. Assay performance was monitored by including plasma samples with known asparaginase activity (low, mid and high activity) as calibrators.

In vitro ASNase cytotoxicity assays

The in vitro cytotoxic activity of six EcASNase products (PI–P6) was evaluated against that of the reference formulation (R, Medac GmbH). Two ALL cell lines, REH and SUP-B15, were used for cytotoxicity studies. Cell lines were authenticated, maintained in antibiotic-free growth medium (RPMI-1640) supplemented with 10% foetal bovine serum (FBS, Gibco) and tested routinely for mycoplasma contamination. Assessment of cytotoxic activity was performed in 100 μl reaction volumes by adding serial 10-fold dilutions of ASNase products (20 μl, 0.001–1000 IU/ml) to 80 μl of cell line suspensions (REH, 50,000 cells/well; SUP-B15, 35,000 cells/well) in triplicates, followed by incubation for 72 hours (37°C, 5% CO2). Proportion of live cells was determined by measuring the metabolic activity of intact cells using a microtitre plate-based WST-1 (4-[3-(4-Iodophenyl)-2-(4-nitro-phenyl)-2H-5-tetrazolio]-1,3-benzene sulfonate) cell proliferation assay (Sigma). In the assay, mitochondrial dehydrogenases in intact cells reduce the water-soluble tetrazolium reagent WST-1 to formazan, which when measured at 440 nm provides an estimate of the proportion of live cells. Proportion of live cells in each well was estimated after incubation with WST-1 for 4 hours. Background absorbance (cell-free growth medium with WST-1) and untreated wells (with phosphate-buffered saline as vehicle control) were used as controls. Half-maximal inhibitory concentrations (IC50) from three separate experiments were derived from a four-parameter nonlinear regression model using the GraphPad Prism software (v8·2·0, www.graphpad.com). Additional details are provided in the accompanying Supporting Information.

Physicochemical characterisation of commercial native E.coli ASNase products

Seven EcASNase products marketed commercially in India (brands P1– P7) were compared with the reference therapeutic EcASNase product (R, Medac GmbH). Where feasible, at least two different manufacturing batches were analysed for each brand (total 12 samples). Quality of the marketed brands was determined by investigating a range of physicochemical attributes. Drug quantity and activity were estimated, respectively, by measuring protein content (bicinchoninic protein assay, Thermo Scientific) and indooxine generation (described previously). Variations in protein characteristics of the test brands were examined using four orthogonal techniques—capillary isoelectric focussing, size exclusion chromatography (SEC) with multiangle light scattering (SEC-MALS), reversed phase high-performance liquid chromatography (RP-HPLC) and sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE). Amino acid sequence of EcASNase biogenerics and the reference products were determined by peptide mass fingerprinting and compared with the theoretical EcASNase sequence. Host cell protein (HCP) content and bacterial endotoxin levels were analysed as measures of product contamination. Finally, suitability for administration was assessed both by evaluating ease of drug reconstitution (appearance of the lyophilised drug cake) and testing product osmolality for parenteral use. Details of analytical techniques are presented in the accompanying Supporting Information.

Statistics

Descriptive statistics were used to report all observations, including the median and interquartile range (IQR) for continuous data. Groups were compared using the Mann–Whitney U test (two groups) and oneway analysis of variance (ANOVA; more than two groups). Two-sided p-values less than .05 were deemed significant.

Results

Subtherapeutic asparaginase activity of an EcASNase biogeneric

Over the 12-month study period, 65 children with newly-diagnosed ALL received the EcASNase biogeneric P2. Three died during the induction phase, and 62 patients were included in the final analysis. The clinical characteristics, treatment response and EcASNase-associated toxicities are shown in Table 1. A total 117 post-dose plasma samples were obtained from the 62 patients during the induction treatment phase. Forty-seven (40%) samples were collected at the 48-hour post-dose timepoint ( post-dose 4 in SR patients; post-dose 7 in IR/HR patients) and the rest (70 samples) 72 hours after EcASNase administration (trough level; post-dose 2 in SR patients; post-dose 2 & postdose 4 in IR/HR patients).. At 48 hours postdose, the median ASNase activity was 149 IU/L (IQR, 93–217 IU/L), declining to 58 IU/L (IQR, 29– 139 IU/L) at 72 hours post-dose (Figure 1). Inadequate activity levels (<100 IU/L) were observed in 14 of 47 (30%) 48 hours post-dose and in 46 of 70 (66%) at the trough time of 72 hours post-dose.

Table 1

Clinical characteristics of the P2 EcASNase patient cohort (N = 62)

	N	%
Age (years)
Median (IQR)	2.8 (3.9–6.1)
Range	1.4–18.2
Sex
Male	41	66
Female	21	34
Immunophenotype
Precursor B	62	100
WBC (× 109/L)
<50	45	73
≥50	17	27
Median (IQR)	22.3(7.1–51.5)
Complete remission at end of induction
Yes	58	94
No	4	6
Minimal residual disease at end of induction
Low (<0.01%)	44	71
High (≥0.01%)	18	29
Risk groups[a]
Standard risk	16	26
Intermediate risk	10	16
High risk	36	58
Asparaginase-associated toxicities
Hypersensitivity	12	19
Pancreatitis	2	3
Thrombosis	0	0

Note: All hypersensitivity reactions were CTCAE grade 3. For pancreatitis, one patient each for CTCAE grade 3 and CTCAE grade 4.

Abbreviations: CTCAE, Common Terminology Criteria for Adverse Events; IQR, interquartile range.

Risk groups: Standard risk (SR) = age ≥1 to <10 years at diagnosis, presenting WBC <50 × 109/L, prednisolone good response, no high-risk cytogenetics, MRD <10-4 end of induction and no CNS disease. Intermediate risk (IR) = age ≥10 years at diagnosis or presenting WBC ≥50 × 109/L or bulky/testicular disease, prednisolone good response, no high-risk cytogenetics and no CNS disease. High risk (HR) = with high-risk cytogenetics or CNS disease or prednisolone poor response or MRD ≥10-4 end of induction.

Figure 1

Subtherapeutic asparaginase activity in patients administered the P2 EcASNase biogeneric.

Box-plot representation of plasma asparaginase activity at trough (72-hour post-dose, n = 70 samples) and 48-hour post-dose timepoints (n = 47 samples) during the ALL induction treatment phase in patients (N = 62) administered P2 EcASNase intramuscularly at 10,000 IU/m2/dose every 72 hours. Adequate asparaginase activity (≥100 IU/L, horizontal dotted line) was observed in 34% (24/70) of trough timepoint samples and 70% (33/47) of 48-hour postdose samples (Mann–Whitney U test of significance, p =.00002).The interquartile ranges are indicated by boxes, whiskers indicate values 1.5 times the upper and lower quartiles, filled circles indicate outliers, median values are indicated by horizontal bold lines within boxes and ’+’ within the boxes indicate mean values

Twelve patients (19%), 11 with high-risk disease, developed clinical hypersensitivity (Common Terminology Criteria for Adverse. Events version 5 (CTCAE) grade 3), requiring discontinuation of further asparaginase treatment. Clinical hypersensitivity in all high-risk patients were associated with the additional EcASNase doses during the consolidation treatment phase (EcASNase doses 9-13).

Shortcomings in critical quality attributes of EcASNase biogenerics marketed in India

Seven EcASNase biogenerics marketed widely in India (brands P1–P7) were examined for drug identity, specific activity, purity and quality.

Table 2 (and Table S1) summarises the findings of these physicochemical studies. No one brand met the minimum expected requirements for all characteristics. Batch-to-batch variability, examined in four brands, additionally demonstrated inconsistency in quality attributes between manufacturing batches. In five of seven brands, drug strength, evaluated by measuring specific activity per milligram of vial protein content, was lower than expected (<220 IU of ASNase activity/milligram of protein).

Table 2

Analysis of critical quality attributes of native EcASNase biogenerics marketed in India

		Native EcASNase biogenerics marketed in India							Reference
		P1	P2	P3	P4	P5	P6	P7	EcASNase
Manufacturing batches tested		2	2	1	2	2	1	1	1
	Expected
Drug strength
Specific activity (IU/mg)	>220	220.2	243.2	157.8	211.2	189.8	154.0	162.0	256.3
Activity per 10,000 IU vial (IU)	10,000 ± 1500	9381.1	7446.3	5295.5	11,696.7	9215.5	5666.5	7080.5	9556.4
Drug identity
Molecular weight, SEC (kDa)	134-136	135.8	136.2	135.9	134.6	135.6	136.6	135.5	134.6
Isoelectric point (pI)	5·37 ± 0.30	5.78	5.75	5.79	5.84	5.81	5.60	5.86	5.46
Drug purity
Aggregates by SEC (%)	<5%	6.0	10.1	17.0	6.6	12.9	8.0	5.8	0.8
Purity by RP-HPLC (%)	>90%	93.8	88.5	79.4	92.0	85.4	91.8	92.6	99.2
Host cell protein (ppm)	<100	9.0	338.5	613.0	85.4	232.5	123.9	320.6	2.6
Bacterial endotoxin (EU/ml)	<50	2.0	1.3	5.1	1.2	0.4	11.0	41.0	0.2
Suitability for parenteral use
Osmolality (mOsm/kg)	250–600	178.5	11	129	106.5	104.5	557	68	279

Note: Values outside acceptable limits are in bold font.

Abbreviations: EcASNase, Escherichia coli L-asparaginase; EU/ml, endotoxin Units/ml; kDa, kilodalton; ppm, parts per million; RP-HPLC, reversed phase high-performance liquid chromatography; SEC, size exclusion chromatography.

Identity of the therapeutic product was determined using a combination of ASNase activity assay, SDS-PAGE analysis, SEC, isoelectric point (pI) determination and peptide mass fingerprinting. In all brands, the ~36 kilodalton (kDa) band of the EcASNase monomer was detected on SDS-PAGE. SEC-MALS confirmed presence of the active enzyme tetramer (~136 kDa) (Figure 2A, B). The pI of the EcASNase biogenerics (5.6–5.86) differed from that of the reference standard (5.37 ± 0.30), suggesting amino acid sequence variations in the biogenerics. RP-HPLC corroborated this observation, with all brands demonstrating retention time peaks different from that of the reference molecule (Figure 2C). Peptide mapping, covering 98.2% of the reference ASNase amino acid sequence, indicated sequence variations for EcASNase biogenerics, suggested both by the reduced sequence coverage and distinct missing amino acid sequences (Table 3).[15] An 11-amino acid sequence (residues 252–262) was not detected in two biogeneric products, while a 22-amino acid sequence (residues 51–72) was not detected in one, suggesting differences in amino acid sequences in these segments. Both sequence segments represent potential EcASNase antigenic sites. Detailed sequencing of the P2 EcASNase used in clinical practice demonstrated amino acid substitutions at position 64 (D64N) and 252 (T252S), the latter within a potential antigenic region. Detailed analysis of amino acid sequence changes in the other EcASNase products was not carried out.

Figure 2

Multimodal analyses indicate impurities and amino acid sequence variations in EcASNase biogenerics.

Examination of EcASNase biogenerics P1–P7 and the reference EcASNase (R) using multimodal analytical techniques, including sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE, A), size exclusion chromatography (SEC, B) and reversed phase high-performance liquid chromatography (RP-HPLC, C). (A) Image of Coomassie-stained SDS-PAGE analysis of EcASNase products examined in reducing (‘+’) and nonreducing (‘-’) conditions confirms presence of the ~36 kDa EcASNase subunit in all products, and shows additional higher and lower molecular weight bands in the biogenerics, suggesting presence of protein impurities (M, molecular weight marker). (B) Overlay of chromatograms from SEC analyses of EcASNase products indicates that in addition to the primary peak of EcASNase, biogeneric products are characterised by additional smaller peaks with shorter retention times (inset, magnified view), suggesting the presence of higher molecular weight impurities, likely multimer aggregates of EcASNase (mAU, milli-absorbance unit). (C) Overlay of chromatograms from RP-HPLC analyses of EcASNase products indicate difference in retention times (~1 minute) between EcASNase biogenerics and the reference product, likely related to differences in hydropathicity owing to amino acid sequence variations in the biogeneric EcASNases. Additional smaller peaks with varying retention times are also observed in the biogeneric products, suggesting impurities (inset, magnified view). (AU, absorbance unit)

Table 3

Protein identification by peptide mass fingerprinting

EcASNase	Sequence coverage (%)[a]	Missing sequence regions
P1	91.1	51-72; 92–95; 302–303
P2	97.9	92–95;302–303
P3	97.9	92–95; 302–303
P4	94.5	92–95; 252–262; 302–303
P5	97.9	92–95; 302–303
P6	94.5	92–95; 252–262; 302–303
P7	97.9	92–95; 302–303
R	98.2	92–95;302–303

Note: Protein identification by peptide mass fingerprinting indicates presence of isoforms of EcASNase in biogeneric products, suggested both by the reduced sequence coverage and distinct missing amino acid sequences (highlighted in bold, P1, P4, P6) in these products. Amino acid sequence regions 92–95 and 302–303 were not detected in all samples, indicating technical limitations in the analysis

Abbreviations: EcASNase, E. coli L-asparaginase; P1–P7, biogeneric native EcASNases; R, reference EcASNase.

With respect to the theoretical reference amino acid sequence of E. coli L-asparaginase.

Purity of EcASNase biogenerics was analysed using a combination of SDS-PAGE, SEC-MALS and RP-HPLC along with assays for HCP content and bacterial endotoxin contamination. SDS-PAGE indicated presence of additional higher and lower molecular weight bands in all EcASNase biogenerics that were not observed in the reference EcASNase product (Figure 2A). While SEC showed a single peak in the reference sample, additional peaks (with shorter retention times, at 8.7 and 11.5 minutes) were observed in all EcASNase biogenerics, indicating presence of higher molecular weight aggregates (5–18%) (Figure 2B). Multiangle light scattering confirmed that these aggregates were multimers of the EcASNase tetramer (molecular mass >280 kDa). Analysis by RP-HPLC demonstrated product impurities in all EcASNase biogenerics (Figure 2C). HCP content was high in six of seven biogenerics, nearing upper tolerable limits in two and exceeding recommended thresholds in four brands (Table 2).

In addition to the above, inconsistences were observed in the appearance of the lyophilised cake in the biogeneric products, with potential deleterious consequences for drug reconstitution, recovery and stability. A homogenous lyophilised cake was observed in only two products, with the rest as either free-flowing or crystalline powder. Product reconstitution time ranged from 30 to 100 seconds, resulting eventually in a clear colourless solution, although one brand was observed to have particle threads after reconstitution. Although the osmolality of all EcASNase formulations was below 600 mOsm/kg, osmolality values were inconsistent. Most formulations were hypotonic (exceedingly low in two, <100 mOsm/kg), while one product had an osmolality close to the upper tolerable limit.

Analysis of in vitro cytotoxicity of EcASNases is not a discriminatory determinant of product quality and therapeutic activity

Malignant lymphoblast cell lines with relatively high (SUP-B15) and low (REH) sensitivity to EcASNase[16] were tested in the WST-1 assay to examine the differential cytotoxic activities of biogeneric and reference EcASNase products (Figure 3). In both cell lines, no differential sensitivities were observed to the EcASNase products, suggesting that conventional in vitro assays of cytotoxicity do not serve as surrogate determinants of therapeutic suitability of these compounds.

Figure 3

Unsatisfactory quality of EcASNase biogenerics is not identified by standard in vitro cytotoxicity studies.

Drug response profile of malignant lymphoblast cell lines with relatively high (SUP-B15) and low (REH) sensitivity in vitro to EcASNase, treated for 72 hours with EcASNase products (biogenerics P1–P6; reference product R) at serial 10-fold dilutions (0.001–1000 IU/ml) and assayed for viability using the WST-1 assay. Graph depicts pooled results from three separate experiments. Error bars represent standard errors of mean. ANOVA was used as the test for significance. Four-parameter logistic regression was used to obtain the dose response curve with the x-axis indicating log EcASNase concentration (log(U/ml)) and the y-axis indicating residual live cells as a percentage of untreated controls. 95% CI, 95% confidence interval; ANOVA, one-way analysis of variance; IC50, half maximal inhibitory concentration (IU/ml)

Discussion

Our findings reiterate concerns over the quality of biogeneric EcASNases marketed in LMICs.[12,13] None of the seven biogeneric formulations of native EcASNase marketed widely in India met the minimum manufacturing quality standards required for these products. These shortcomings in manufacturing quality conceivably impair the therapeutic activity of these biogeneric EcASNases as demonstrated by P2. Importantly, neither in vitro asparaginase activity nor cytotoxic activity of these products was a suitable surrogate determinant of poor therapeutic activity.

Observations with the P2 biogeneric serve as exemplar. Excess impurities were detected in vials of the P2 biogeneric (average purity, 89%), associated with high levels of protein aggregates (10%) and HCP contamination (~250 parts per million). Estimated activity of the product was ~25% lower than that specified in the product label. The unusually low osmolality of the batches tested renders the P2 product unsuitable for parenteral use. Unsatisfactory therapeutic activity was observed when the P2 biogeneric was administered to patients. Generic EcASNase dosed at 2500–5000 IU/m2 every 72-hours has previously been reported to produce trough levels of >100 IU/L in 95% of patients.[17] In contrast, nearly two thirds of patients administered intramuscular P2 EcASNase at standard doses (10,000 IU/m2/dose, every 72 hours) demonstrated inadequate drug activity at the 72-hour trough timepoint. Nearly one in five patients administered the P2 biogeneric developed clinical hypersensitivity, almost all of whom were treated on high-risk schedules that require intensified asparaginase therapy. Variations in amino acid sequence contribute to but do not probably account fully for the observed differences in therapeutic activity and antigenicity of the P2 biogeneric.

These observations have several implications. Asparaginase is primarily used early in the therapy of ALL. Inadequate asparaginase activity is likely to compromise the early response to therapy, a key prognostic determinant of survival in ALL.[18-21] This is especially so in patients administered anthracycline-free schedules of asparaginase, vincristine and corticosteroid during the induction treatment phase. The variable quality of generic agents, in particular biologicals such as asparaginase, marketed in LMICs[22] is potentially a significant determinant of the observed differences in outcomes between these countries and more affluent nations.[23,24]

Nearly one in five patients administered P2 EcASNase, the majority with high-risk disease, had to discontinue asparaginase treatment due to clinical hypersensitivity. Erwinase is unaffordable in our context. The North American Children’s Oncology Group reported higher risk of treatment failure in patients who required premature discontinuation of ASNase therapy, particularly in patients with high-risk disease who require intensive ASNase therapy.[25] ASNase discontinuation in these patients was predominantly due to clinical hypersensitivity. The Nordic Nordic Society of Paediatric Haematology and Oncology (NOPHO) group similarly reported increased risk of treatment failure in patients with shortened ASNase schedules, including in patients with silent ASNase inactivation identified through therapeutic drug monitoring.[26]

In our cohort, change to alternate-day dosing of P2 EcASNase (48-hour dose intervals) to obtain trough therapeutic levels would still be associated with inadequate drug activity in a third of patients. The additional drug doses with the alternate-day schedule are also likely to increase the risk of drug hypersensitivity and cost of treatment. Similar subtherapeutic activity has been observed with other biogeneric EcASNases in LMICs. Investigators from India have reported unsatisfactory drug activity in 75% of postdose plasma samples obtained from patients administered nine different biogeneric EcASNase formulations, with the proportion of subtherapeutic activity varying widely among the generic agents (from ~25% to ~100%).[12] Investigators in Brazil observed unsatisfactory asparaginase activity in vivo in a high proportion of patients administered a biogeneric EcASNase (97%) compared to the reference product (19%) with resultant inferior outcomes.[27] High level of HCP contamination (estimated at 19–37% of product content) was observed in biogeneric formulations, potentially accounting for the decreased drug bioavailability observed in patients and heightening the risk of immune reactions to EcASNase.[13] Recent availability of PEG-conjugated EcASNase biogenerics provides a potential therapeutic alternative. At least six biogeneric PEG-EcASNase brands are marketed in India currently. Switching to PEG-EcASNase biogenerics is not a ready remedy. Manufacture of these compounds is a challenge, requiring stable uniform PEG-conjugation of EcASNase, minimising polydispersible complexes and diol impurities.[28] Inadequacies in PEG-conjugation will compound existing shortcomings in the manufacture of native EcASNase biogenerics. The relatively high cost of PEG products and the limited stability of these liquid formulations are additional drawbacks. Added to this are the intrinsic variabilities associated with altered amino acid sequences of the EcASNase isoforms in these products. Clinical use of PEG-EcASNase biogenerics will therefore require stringent tests of product quality and monitored studies to determine suitability for therapeutic use and the appropriate dose and schedule.

In summary, our findings echo deep concerns that across the globe for many patients with ALL, assured access to high-quality affordable asparaginases continues to remain an unmet therapeutic need.

Supplementary Material

Additional supporting information may be found online in the Supporting Information section at the end of the article.