Introduction

Acute lymphoblastic leukemia (ALL) is a devastating hematological malignancy with a high incidence rate in pediatric patients whose ages are under 14 years.^1–3 Polyethylene glycol-conjugated L-asparaginase (PEG-asp) combined with other chemotherapeutic drugs is commonly prescribed during the induction therapy for childhood ALL patients aiming to achieve a complete response,^4,5 However, with a wide range of applications of PEG-asp in treating childhood ALL patients, several unexpected anaphylaxes and adverse reactions are reported, such as skin rash, laryngeal edema, and risk of asphyxia.^6,7 These PEG-asp-related anaphylaxes may cause discontinuation of the therapy and reduced therapeutic efficacy, leading to a high relapse risk.^8,9 Hence, it is critical to identify some novel biomarkers to predict PEG-asp-related anaphylaxes and adverse reactions in childhood ALL patients, which might further improve their overall management.

L-asparaginase (L-asp), the unconjugated form of PEG-asp, is initially identified in bacteria as a key enzyme responsible for converting L-asparagine to L-aspartic acid.^4,5 L-asp is reported to be linked with the absence of PEG- asp-related anaphylaxis in childhood ALL patients due to its interaction with the immune receptors (such as human leukocyte antigen (HLA) or T cell receptor (TCR), etc.) on T cells.¹⁰ While limited studies investigate the association of L-asp activity with PEG-asp-related adverse reactions, only one exploratory clinical study reports that L-asp activity and anti-L-asp antibody are related to PEG-asp-related anaphylaxis risk but they fail to estimate the PEG-asp-related non-anaphylaxis adverse event risk.¹¹ However, its sample size is relatively small, also the clinical factors relating to PEG-asp-related anaphylaxis risk are not analyzed.

Therefore, L-asp activity and anti-L-asp antibody were detected in 170 childhood ALL patients who received PEG-asp-containing therapy in the current study, aiming to explore their values in estimating PEG-asp-related anaphylaxis risk and PEG-asp-related non-anaphylaxis adverse reaction risk as well as to identify potential clinical factors related to PEG-asp-related anaphylaxis in childhood ALL patients.

Methods

Patients

From April 2019 to May 2021, this study recruited a total of 170 childhood ALL patients. The inclusion criteria were set as follows: (1) confirmed as ALL per the guideline of children ALL,¹² (2) less than 16 years old, (3) scheduled to undergo treatment with PEG-asp-contained Chinese Children’s Cancer Group (CCCG)-ALL-2015 treatment strategy,¹³ and (4) willing to comply with the study protocol. Patients with the following conditions were excluded for inclusion: (1) had immunodeficiency diseases, and (2) complicated with cancers or other hematological malignancies. The study was permitted by Ethics Committee. For patients aged ≥8 years, a written informed consent was obtained from both the patient and their guardian; meanwhile, for patients aged <8 years, the written informed consent was obtained from the guardian.

Data collection and treatment

Demographics, disease characteristics, and allergic history were collected for analysis. Risk stratification of patients was assessed as per the guidelines and classified as low risk (LR), intermediate risk (IR), and high risk (HR).¹² The patients underwent treatment with PEG-asp contained CCCG-ALL-2015 protocol.¹³ The regimens of PEG-asp-contained treatment were as follows: (i) dexamethasone (DEX) + vincristine (VCR) + daunorubicin (DNR) + PEG-asp, (ii) 6- mercaptopurine (6-MP) + VCR + DNR + PEG-asp, and (iii) DEX + VCR + arabinoside cytosine (Ara-C) + PEG-asp. PEG-asp (H20090015, Jiangsu Hengrui Pharmaceutical Co. LTD, China) was administrated intramuscularly 2000 U/m² on the 6th day and the 26th day.

Sample processing

On the 7th day after treatment initiation, peripheral venous blood samples of patients were gained, and then the plasma samples were isolated by centrifugation to evaluate the activity of L-asp and anti-L-asp antibody. Plasma L-asp activity was detected by photometry method, which is in line with a previous study.¹¹ In brief, the diluted plasma samples were mixed with Tris-HCl and L-asparagine and then incubated at 37°C for 30 min. Sequentially, trichloroacetic acid was added and the mixture was centrifuged. The supernatant was mixed with redistilled autoclaved water and Nessler’s reagent, and then incubated at 25°C for 15 min. Next, the optical density (OD) value (405 nm) was obtained, and the calibration curve was plotted. Based on the calibration curve and the sample OD value, the plasma L-asp activity was calculated. The plasma anti-L-asp antibody was examined by a commercial enzyme-linked immunosorbent assay kit (catalog number: EY3616, Shanghai yiyan bio-technology Co. Ltd., China) per the instructions.

Assessment

The adverse reactions of patients were recorded. Later, two pharmacists independently assessed and attributed the PEG-asp-related adverse reactions per the instruction of PEG-asp based on the World Health Organization-Uppsala Monitoring Centre (WHO-UMC) system.¹⁴ In detail, the adverse reactions were classified as PEG-asp-related adverse reactions when the causality of the adverse reactions was certain and probable; the he adverse reactions did not classify as PEG-asp-related in the case of causality of possible, unlikely, conditional, and unassessable. Moreover, a third pharmacist was invited if there was a disagreement. After the assessment, the PEG-asp-related adverse reactions were further classified into anaphylaxis adverse reactions (rash, dyspnea, anaphylactic shock, laryngeal edema, etc.) and non-anaphylaxis adverse reactions (nausea and vomiting, pancreatitis, diarrhea, elevated transaminases, headache, and etc.). Besides, the management of adverse reactions was documented as well.

Statistics

The data were analyzed using SPSS 20.0 (IBM Corp., USA). Graphs were made using GraphPad Prism 6.01 (GraphPad Software Inc., USA). Comparisons of L-asp activity and anti-L-asp antibody in patients with or without adverse reactions were analyzed using Wilcoxon rank-sum test. The abilities of L-asp activity and anti-L-asp antibody in comparing different patients were assessed using receiver operating characteristic (ROC) curves. The adverse reaction rates in patients with different disease characteristics were compared using Fisher’s exact test. Correlations of two variables were analyzed using linear-by-linear test or Spearman’s rank correlation test. Independent factors for PEG-asp-related anaphylaxis were assessed by forward- stepwise multivariate logistic regression analysis. P <0.05 was considered significant.

Results

1 Childhood ALL patients’ characteristics

The mean age of childhood ALL patients was 6.5±3.7 years. There were 55 (32.4%) girls and 115 (67.6%) boys. In addition, 146 (85.9%) and 24 (14.1%) were classified as De novo ALL patients and relapsed ALL patients, separately. In terms of their risk stratification, 99 (58.2%), 62 (36.5%), and 9 (5.3%) patients were classified as LR, IR, and HR, respectively. Furthermore, 151 (88.8%) patients did not report an allergic history, while the remaining 19 (11.2%) had an allergic history. The detailed clinical information was shown in Table 1.

PEG-asp-related anaphylaxis and non-anaphylaxis adverse reactions

Among 170 childhood ALL patients, 27 (15.9%) of them occurred a PEG-asp-related adverse reaction (Table 2). Specifically, there were 17 (10.0%) patients who experienced PEG-asp-related anaphylaxis including 10 (5.9%) cases of rash, 4 (2.4%) cases of dyspnea, 2 (1.2%) cases of anaphylactic shock and 1 (0.6%) case of laryngeal edema. Moreover, 14 (8.2%) childhood ALL patients were identified with PEG-asp-related non-anaphylaxis adverse reactions including 4 (2.4%) cases of nausea and vomiting, 4 (2.4%) cases of pancreatitis, 3 (1.8%) cases of diarrhea, 2 (1.2%) cases of elevated transaminases, and 1 (0.6%) case of headache. Furthermore, there was no PEG-asp-related death reported.

After occurring PEG-asp-related adverse reaction, 23 (85.2%) of 27 childhood ALL patients were switched to another drug, while the remaining 4 (14.8%) patients did not change (Table 3).

Correlation of L-asp activity and anti-L-asp antibody with PEG-asp-related anaphylaxis as well as non-anaphylaxis adverse reaction

L-asp activity was negatively related to anti-L-asp antibody in childhood ALL patients (P<0.001, Figure 1). Elevated L-asp activity was associated with the absence of PEG-asp-related anaphylaxis (P<0.001, Figure 2A), PEG-asp-related non-anaphylaxis adverse reaction (P=0.004, Figure 2B), and total PEG-asp-related adverse reaction (P<0.001, Figure 2C). Moreover, a higher anti-L-asp antibody was linked with the occurrence of PEG-asp-related anaphylaxis (P<0.001, Figure 2D), PEG-asp-related non-anaphylaxis adverse reaction (P=0.030, Figure 2E), and total PEG-asp-related adverse reaction (P<0.001, Figure 2F).

Figure 1 L-asp activity was negatively related to the anti-L-asp antibody. Spearman’s correlation test was applied for statistical analysis.

Figure 2 L-asp activity and anti-L-asp antibody were related to PEG-asp-related anaphylaxis. Comparison of L-asp activity between childhood ALL patients with PEG-asp-related anaphylaxis and patients without PEG-asp-related anaphylaxis (A); between patients with PEG-asp-related non-anaphylaxis adverse reaction and patients without PEG-asp-related non-anaphylaxis adverse reaction (B); between patients with total PEG-asp-related adverse reaction and patients without PEG-asp-related adverse reaction (C). Comparison of anti-L-asp antibody between childhood ALL patients with PEG-asp-related anaphylaxis and patients without PEG-asp-related anaphylaxis (D); between patients with PEG-asp-related non-anaphylaxis adverse reaction and patients without PEG-asp-related non-anaphylaxis adverse reaction (E); between patients with total PEG-asp-related adverse reaction and patients without total PEG-asp-related adverse reaction (F). Wilcoxon rank-sum test was applied for statistical analysis.

Further ROC curve analyses were performed to explore the association of L-asp-activity and anti-L-asp antibody with total PEG-asp-related adverse reaction risk. Then, it was observed that L-asp activity was related to PEG-asp-related anaphylaxis risk with an area under curve (AUC) of 0.955 (95% confidence interval (CI): 0.921-0.989), PEG-asp-related non-anaphylaxis adverse reaction risk with an AUC of 0.730 (95% CI: 0.601-0.859), and total PEG-asp-related adverse reaction risk with an AUC of 0.857 (95% CI: 0.778-0.935) (Figure 3A). Moreover, the anti-L-asp antibody was linked with PEG-asp-related anaphylaxis risk with an AUC of 0.905 (95% CI: 0.836-0.974), PEG-asp-related non-anaphylaxis adverse reaction risk with an AUC of 0.675 (95% CI: 0.541-0.810), and total PEG-asp-related adverse reaction risk with an AUC of 0.801 (95% CI: 0.710-0.892) (Figure 3B). According to the values of AUCs, it was considered that L-asp activity and anti-L-asp antibody exhibited superior predictive values in estimating PEG-asp-related anaphylaxis risk than PEG-asp-related non-anaphylaxis adverse reaction risk.

Figure 3 L-asp activity and anti-L-asp antibody estimated PEG-asp-related anaphylaxis risk. The value of L-asp activity (A) and anti-L-asp antibody (B) in estimating PEG-asp-related anaphylaxis risk, PEG-asp-related non-anaphylaxis adverse reaction risk, and total PEG-asp-related adverse reaction risk by ROC curve analyses. ROC curve analyses were applied for statistical analyses.

After adjustment using multivariate logistic regression model, a higher L-asp-activity (odds ratio (OR): 0.992, P=0.004) was shown to be independently related to a reduced PEG-asp-related anaphylaxis occurrence, while a higher anti-L-asp antibody (OR: 1.022, P=0.003) was independently related to an increased PEG-asp-related anaphylaxis occurrence in childhood ALL patients (Table S1).

Correlation of clinical factors with PEG-asp-related anaphylaxis and non-anaphylaxis adverse reactions

De novo (vs relapse) disease status (P=0.135), higher risk stratification (P=0.087), and presence of allergic history (P=0.103) seemed to correlate with higher PEG-asp-related anaphylaxis rate but without statical significance, while T-ALL (vs. B-ALL) (P=0.706) was similar between childhood ALL patients with or without PEG-asp-related anaphylaxis (Figure 4A–D). Moreover, disease status, immunophenotype, risk stratification, or allergic history did not relate to PEG-asp-related non-anaphylaxis adverse reaction rate (Figure 4E–H) or total PEG-asp-related adverse reaction rate (all P>0.05) (Figure 4I–L).

Figure 4 De novo (vs. relapse), higher risk stratification, and allergic history exhibited trends related to PEG-asp-related anaphylaxis risk. Association of disease status, immunotype, risk stratification, allergic history, and PEG-asp-related anaphylaxis rate (A–D); PEG-asp-related non-anaphylaxis adverse reaction rate (E–H); total PEG-asp-related adverse reaction rate (I–L). Fisher’s exact test and linear-by-linear test were applied for statistical analyses.

Discussion

Different formulations of L-asp have been applied in treating ALL patients, which includes the native E. coli asparaginase, Erwinia asparaginase, and PEG-asparaginase.¹⁵ Although PEG-asp is reported to be less likely to cause anaphylaxis when compared with the other two formulations, PEG-asp-related anaphylaxis still emerges as one critical challenge for pediatric ALL patients.^16,17 It is estimated that the PEG-asp-related anaphylaxis rate ranges from 7.7% to 44.0% in childhood ALL patients, among which rash, dyspnea, and anaphylactic shock are often reported.^9,11,18,19 In terms of PEG-asp-related non-anaphylaxis reactions, the incidence rate is quite low, among which pancreatitis and gastrointestinal symptoms are commonly observed.^11,19 In the present study, it was observed that 15.9% of childhood ALL patients developed PEG-asp-related adverse reactions, among which 10.0% of them developed PEG-asp-related anaphylaxis and 8.2% of them developed PEG-asp-related non-anaphylaxis adverse reactions. The possible reason for these findings was that: L-asp might induce PEG-asp-related anaphylaxis in childhood ALL patients through the overstimulation of T cells, alteration of antigen-antibody interaction, and impairment of lysosomal protease activity.^20–22

Limited studies explore the value of L-asp activity and anti-L-asp antibody in predicting PEG-asp-related adverse reactions in childhood ALL patients. One study reports that higher L-asp activity is related to the absence of PEG-asp-related anaphylaxis in childhood ALL patients.¹¹ Moreover, the measurement of anti-L-asp antibody at week 7 of continuation therapy predicts the occurrence of PEG- asp-related anaphylaxis.¹⁸ In terms of their correlation with non-anaphylaxis adverse reactions, neither L-asp activity nor anti-L-asp antibody is related to PEG-asp-related non- anaphylaxis adverse reactions.¹¹ In the current study, it was observed that L-asp activity and anti-L-asp antibody displayed better capabilities to predict PEG-asp-related anaphylaxis risk than PEG-asp-related non-anaphylaxis adverse reaction risk in childhood ALL patients. Moreover, anti-L-asp antibody and anti-L-asp activity independently estimated the PEG-asp-related anaphylaxis risk in childhood ALL patients. The possible reasons to explain these findings were: (a) L-asp mediated the production of anti-L-asp antibody where the latter one could bind to immunoglobin (Ig) G and IgE receptors on the basophils and further promote their activation causing hypersensitivity, thus higher anti-L-asp antibody was related to the occurrence of PEG-asp-related anaphylaxis in childhood ALL patients.²¹ (b) PEG-asp-related anaphylaxis might induce some non-anaphylaxis adverse reactions indirectly in childhood ALL patients through the overstimulation of the gastrointestinal tract causing symptoms like nausea and vomiting etc. Importantly, L-asp activity exhibited more sensitivity than anti-L-asp antibody to estimate PEG-asp-related anaphylaxis risk in childhood ALL patients in the current study, which provided a novel and feasible method to estimate PEG-asp-related anaphylaxis risk in childhood ALL patients.

Also, another interesting finding in the current study was the weak relationship of De novo (vs. relapse) disease status, higher risk stratification, and presence of allergic history with higher PEG-asp-related anaphylaxis rate in childhood ALL patients, while no statistical significance was observed. The possible reason for this finding was: (a) relapsed childhood ALL patients might be applied for PEG-asp therapy prior to the treatment and they were desensitized to the PEG-asp, thus they were at low risk of occurring PEG-asp-related anaphylaxis. (b) childhood ALL patients in the HR group required extra dosages of PEG-asp than those in the LR or IR group, thus these patients were exposed to a high amount of PEG-asp and were more likely to develop PEG-asp-related anaphylaxis. (c) childhood ALL patients with allergic history to antibiotics might be associated with overreact immune system and uncontrolled immunomo-dulation, thus linking with a high risk of developing PEG- asp-related anaphylaxis.^23,24

The current study discovered several novel findings compared to one previous study addressing the same issue,¹¹ which included (i) a weak relationship of De novo (vs. relapse) disease status, higher risk stratification, presence of allergic history with higher PEG-asp-related anaphylaxis rate in childhood ALL patients; (ii) the value of L-asp activity and anti-L-asp antibody as an independent factor to estimate PEG-asp-related anaphylaxis risk in pediatric ALL patients; (iii) a larger sample size to assess the PEG-asp-related anaphylaxis occurrence in pediatric ALL patients. However, some limitations still existed in this study. For instance, PEG-asp was also applied for adult ALL treatment, while the value of L-asp activity and anti-L-asp antibody in estimating PEG-asp-related anaphylaxis risk in adult ALL patients could be determined in future studies. Also, further studies with a larger sample size were needed to perform more subgroup analyses among pediatric ALL with different characteristics (such as treatment regimens). Furthermore, further studies could investigate the correlation of L-asp activity and anti-L-asp antibody with the response as well as survival in childhood ALL patients.

In conclusion, L-asp activity and anti-L-asp antibody exhibited superior predictive values in estimating PEG- asp-related anaphylaxis risk than PEG-asp-related non- anaphylaxis adverse reaction risk in childhood ALL patients. Our study indicates the monitoring of L-asp activity and anti-L-asp antibody may be useful for early identification and prevention of PEG-asp-related anaphylaxis in childhood ALL management.