Preclinical evaluation of engineered L-asparaginase variants to improve the treatment of Acute Lymphoblastic Leukemia

Highlights

• •When compared to the wild-type EcA, KHY-17 and KHYW-17 EcA variants had comparable asparaginase activity at 37 °C. Both the variants showed approximately 24-fold lower glutaminase activity.
• •The triple mutant KHYW-17 exhibited higher catalytic activity. The cytotoxicity of both the variants against leukemic cell lines was significantly greater when compared to the WT-EcA.
• •Both the variants produced approximately three times less IgG and IgM than WT-EcA in vaccinated mice, suggesting that these variants were less immunogenic. Additionally, pharmacokinetic study in mice displayed that the mutants had significantly longer half-life in comparison to WT-EcA.
• •In Balb/C mice, single and repeated doses demonstrated no additional toxicity up to 5000 IU/kg and 1600 IU/kg, respectively. Likewise, these variations have demonstrated better efficacy in the ALL xenograft mouse model when compared to wild-type EcA. In mice treated with KHY-17 and KHYW-17 variants, we observed 80–90 % reduction in leukemic cells, compared to 40 % reduction in mice treated with wild-type EcA.
• •The survival rate of the mice treated with KHY-17 and KHYW-17 individually was 90 % as opposed to 10 % in the WT-EcA group. Notably, the binding of KHYW-17 was 2-fold lower by pre-existing anti-EcA antibodies present in serum samples obtained from primary and relapsed ALL patients undergoing asparaginase therapy.

Abstract

Introduction

Escherichia coli l-asparaginase (EcA), an integral part of multi-agent chemotherapy protocols of acute lymphoblastic leukemia (ALL), is constrained by safety concerns and the development of anti-asparaginase antibodies. Novel variants with better pharmacological properties are desirable.

Methods

Thousands of novel EcA variants were constructed using protein engineering approach. After preliminary screening, two mutants, KHY-17 and KHYW-17 were selected for further development. The variants were characterized for asparaginase activity, glutaminase activity, cytotoxicity and antigenicity in vitro. Immunogenicity, pharmacokinetics, safety and efficacy were tested in vivo. Binding of the variants to pre-existing antibodies in primary and relapsed ALL patients’ samples was evaluated.

Results

Both variants showed similar asparaginase activity but approximately 24-fold reduced glutaminase activity compared to wild-type EcA (WT). Cytotoxicity against Reh cells was significantly higher with the mutants, although not toxic to human PBMCs than WT. The mutants showed approximately 3-fold lower IgG and IgM production compared to WT. Pharmacokinetic study in BALB/c mice showed longer half-life of the mutants (KHY-17- 267.28±9.74; KHYW-17- 167.41±14.4) compared to WT (103.24±18). Single and repeat-doses showed no toxicity up to 2000 IU/kg and 1600 IU/kg respectively. Efficacy in ALL xenograft mouse model showed 80–90 % reduction of leukemic cells with mutants compared to 40 % with WT. Consequently, survival was 90 % in each mutant group compared to 10 % with WT. KHYW-17 showed over 2-fold lower binding to pre-existing anti-asparaginase antibodies from ALL patients treated with l-asparaginase.

Conclusion

EcA variants demonstrated better pharmacological properties compared to WT that makes them good candidates for further development.

Graphical abstract

Introduction

Ever since 1960s, Escherichia coli derived l-asparaginase (EcA) has been an invariable component of nearly every ALL protocol [1]. It made a significant difference to relapse free survival (RFS) and overall survival (OS) in both adults as well as pediatric ALL. EcA hydrolyses free blood l-asparagine (Asn) to l-aspartate and ammonia. Thus, it leads to inhibition of protein synthesis, cell cycle arrest in G0/G1 phase, and apoptosis of leukemic blasts which depend on extraneous Asn-due to inherent deficiency of asparagine synthetase [2], [3], [4], [5]. l-asparaginase treatment regimen increased the survival rate of ALL patients to over 80 % [6]

The success of EcA, notwithstanding several drawbacks, has come to the fore over the years. Number of doses of EcA varies from 4 to 8 depending on risk stratification. However, clinical hypersensitivity is often reported following repeated administration [7]. Also, l-asparaginase related toxicities and major side-effects including hypersensitivity, pancreatitis, hepatotoxicity and thromboembolism besides its immunogenicity and short half-life may compromise clinical outcomes because of failure to maintain treatment intensity [5,8].

Having derived from a bacterial source, EcA often recognized as a foreign substance by human immune system, resulting in the production of anti-drug antibodies, which neutralize the enzyme in up to 60 % of cases known as “silent inactivation” [9,10].

Presently, other formulations such as PEG-l-asparaginase (PEGylated l-asparaginase) and Erwinia chrysanthemi l-asparaginase (Erwinial-asparaginase) are also available [9,11]. Although, PEG-asparaginase has longer half-life compared to EcA, hypersensitivity reactions to PEG-asparaginase are more usual than the other side-effects linked with it. Moreover, silent inactivation in PEG-asparaginase treated patients is reported [11]. Published reports suggest that PEG-asparaginase treatment shows moderate to severe toxicities including hepatotoxicity, pancreatitis, thromboembolism, hyper-bilirubinemia and transaminitis [12], [13], [14]. Complications due to first-line asparaginase treatment are preferably switched to Erwinial-asparaginase to complete the treatment protocol and attain optimal event-free survival [9]. While Erwinial-asparaginase is better tolerated, it has lower therapeutic efficacy than EcA [15].

L-asparaginase also possesses glutaminase activity, which results deamidation of l-glutamine [16]. However, clinical consequences of glutaminase activity are uncertain. Observation in favor of glutaminase activity was reported by Wai Kin Chan et al. in 2014 who showed that it is necessary for durable anti-cancer activity against Asparagine Synthetase (ASNS) positive blasts [17]. More recently, the same group reported that side effects of wild-type l-asparaginase were more severe than the l-glutaminase-free l-asparaginase variant (L-ASP-Q59L) [18]. Several animal studies have shown that glutaminase activity may have deleterious effects on survival [19,20]. Complications such as ketonic hyperglycinemia, hypocholesterolemia, glycosuria, hepatotoxicity, prolonged bleeding time, pancreatitis and neurotoxicity have been attributed to glutaminase activity of l-asparaginase [16,[21], [22], [23], [24]]. Thus, novel l-asparaginase enzyme with better pharmacological profile may improve the overall outcomes of ALL.

Based on our previous studies and to address these issues, we have targeted certain amino acids namely lysine, tyrosine and tryptophan present at the subunit interfaces and B-cell epitopes responsible for stability, glutaminase activity and immunogenicity and used protein engineering approach to replace them with other amino acids for the construction of l-asparaginase variants with improved properties. Over 1000 EcA variants were screened for desirable therapeutic properties. The main aim was to improve the stability and reduce immunogenicity and glutaminase activity without compromising the asparaginase activity of the variants [5,25]. Here, we report the activity, thermostability, cytotoxicity, pharmacokinetics, immunogenicity, antigenicity, safety and efficacy of KHY-17, a double mutant, and its corresponding triple mutant, KHYW-17.

Materials and methods

Construction, expression and purification of ECA variants

E. coli asparaginase (EcA) encoding ansB gene cloned into pTEW1 (ansB-pTEW1) plasmid [26] was used as a template for site-directed mutagenesis (Quick Change XL II kit, Agilent Technologies, Santa Clara, CA) using gene specific primers (Supplementary Table 1). Resultant PCR products were used to develop ansB-pET28a construct and were confirmed by sequencing (Supplemental methods). After stringent screening from more than 100 mutants, a double mutant, K288H/Y176F (KHY-17) and a triple mutant, K288H/Y176F/W66Y (KHYW-17) were selected. The constructs were expressed in E. coli BL21 (DE3) for purification by chromatofocussing [27] (Supplemental methods). The confirmed EcA variants were screened for asparaginase activity, thermal stability, glutaminase activity, enzyme kinetics and antigenicity.

Determination of asparaginase activity, thermal stability, enzyme kinetics and glutaminase activity

Asparaginase activity of EcA variants and wild-type were measured as described previously [27]. Thermal stability of EcA constructs was determined by performing asparaginase activity assay at a temperature range from 55 °C to 70 °C. Kinetic properties measured by varying substrate concentrations of AHA in the range of 0–5 mM using the same indooxine assay [28]. Glutaminase activity was measured using a two-step discontinuous coupled enzymatic reaction as described previously [29,30] (Supplemental methods).

Anti-leukemic activity and antigenicity of ECA variants

Anti-leukemic activity of EcA wild-type and its variants were studied by MTT assay on Reh cell line (human B cell precursor cell line, ATCC) and PBMCs isolated from healthy individuals. Antigenicity of EcA enzymes was measured by indirect ELISA method as described previously [5,25] (Supplemental methods).

Immunogenicity study in mice

Female 6–8 weeks BALB/c mice were immunized with one primary and two booster doses of wild-type EcA, double mutant (KHY-17), and the triple mutant (KHYW-17). One week after first and second booster, blood samples were drawn; sera were prepared and stored. Serum antibody titer was determined by indirect ELISA method (Supplemental methods). Considering females are more sensitive to antigens, female gender was chosen in the investigation.

Pharmacokinetic study

Female BALB/c mice were administered with all the l-asparaginase formulations individually at 50 IU/Kg/mice concentration. Blood samples were collected at seven time points and asparaginase activity was measured using a 4-parametric non-linear regression equation as described by Sankaran et al. [31]. Detailed methodology is described in supplementary methods.

Toxicity study in mice

Single dose toxicity

27 female BALB/c mice randomized into 3 groups were further divided into 3 sub-groups containing 3 mice each. Mice were administered with 3 different l-asparaginase formulations through intravenous route at 500 IU/Kg, 1000 IU/Kg or 2000 IU/Kg concentration in each subgroup individually once and were observed 14 days for any clinical signs of toxicity or mortalities. Another 3 mice were administered with normal saline (vehicle control) and were observed for 14 days. On 15th day blood samples were collected for biochemical and hematological investigations. Simultaneously mice were sacrificed and vital organs were collected for histopathology.

Repeat dose toxicity

90 BALB/c mice divided into 3 groups having 3 sub-groups of 10 mice (5 male and 5 female) each were administered with three different l-asparaginase formulations at 400, 800 or 1600 IU/Kg concentration through intra-peritoneal route five days a week for 4 weeks. Animals were observed for any physical signs of toxicity or mortality. As vehicle control, 10 mice were administered normal saline in the same dosing schedule. On day 29, mice were sacrificed and blood samples were collected for biochemical and hematological investigations, and vital organs were collected for histopathology.

Efficacy study in leukemia xenograft model in immunocompromised mice

Xenograft model was developed by intravenous administration of Jurkat (human ALL cell line, ATCC) cells in NOD/SCID mice (Supplemental methods). Once the% of human CD45+ve cells was found to be ≥1 %, mice in each group were administered with the respective l-asparaginase formulation at a dose of 800 IU/Kg intra-venous 4 times at an interval of 72 h [32]. No intervention for mice in control group.% of human CD45+ve cells was examined once a week to evaluate drug efficacy. Further, mice were observed till day 21 for mortality.

Antibody binding study in blood samples from all patients

Blood samples from pediatric ALL patients having different conditions were collected. Samples from healthy individuals served as control. Serum was prepared from ALL patients’ blood samples and stored. Binding of pre-existing antibodies present in ALL patients’ sera to EcA variants was determined by indirect ELISA method (Supplemental methods).

Statistical analysis

Significant differences between the groups were determined by one-way analysis of variance. All the statistical calculations were performed using Dunnett's Multiple Comparison Test and Bonferroni as per the requirement with the help of GraphPad prism version 5.0. Significance was indicated as *** for p<0.001, ** for p <0.01, and * for p <0.05.

Ethics approval and consent to participate

Animal experiments were performed in accordance to the CPCSEA (Committee for the Purpose of Control and Supervision of Experiments on Animals) guidelines and approval of Institutional Animal Ethical Committee (IAEC) was obtained. Blood samples from patients and healthy volunteers were collected after receiving the approval from the Institutional Ethical Committee (IEC). Written informed consent was obtained from the study participants.

The data generated in this study are available within the article and its supplementary data files.

Results

Previously, Mashburn et al., 1968 and Kwok et al., 2006 have identified two residues, Ser-122 and Tyr-181 at the dimer-dimer interface which affects the activity and stability of l-asparaginase [33,34]. Based on these observations, we have replaced Tyr-176 located near Ser-122 and Tyr-181. This mutation has resulted in reduced glutaminase activity of our mutant variants. In an attempt to target immunogenicity of l-asparaginase, we first identified several B-cell epitopes responsible for immunogenicity from previous literature [35]. Further we selected 5 amino acids namely Asn-55, Asp-204, Asp-255, Lys-229, and Lys-288 from these epitopes and replaced them individually using site directed mutagenesis. Thousands of colonies from different variant were screened for immunogenicity using indirect ELISA. Initially variants with single mutation did not show any significant reduction in immunogenicity. However, we could develop a less immunogenic mutant by combining the mutation, Y176F at the dimer interface and K288H in the B-cell epitope thus generating the double mutant K288H/Y176F (abbreviated as KHY-17). W66 is the lone tryptophan residue present in l-asparaginase [36]. In our previous study, we found single mutants Y176F and W66Y and the double mutant Y176F/W66Y showed comparable asparaginase activity with the wild type [25]. Here we have combined the mutation, W66Y with the double mutant, KHY-17 and have generated a triple mutant namely, K288H/Y176F/W66Y (abbreviated as KHYW-17).

Purity of the ECA variants

Upon purification with Ni-NTA, anion exchange and gel filtration chromatography, wild-type EcA, KHY-17 and KHYW-17 were at least 95 % pure as determined by SDS-PAGE (Fig. 1A).

Fig. 1.

In-vitro characterization of l-asparaginase variants. (A) SDS-PAGE analysis of purified wild-type l-asparaginase and its variants. (B) Asparaginase activity at 37 °C. The experiment was performed in triplicate. Mean ± S.D are shown here. (C) Thermal stability analysis at a temperature range of 55 °C to 70 °C. Enzyme kinetics was determined by gradually increasing the synthetic substrate (AHA) concentration at (D) 37 ⁰C and (E) 62 ⁰C at pH 7.0 using Lineweaver-Burk equation. This experiment was performed in duplicate. (F) Glutaminase activity of asparaginases (25 µg) after 10 min of incubation with l-Glutamine at 37 ⁰C. (G) Antigenicity was determined in-vitro by indirect ELISA using commercially available anti-EcA-IgG antibody. Each experiment was performed thrice; Mean ± SEM are shown here. Significance was determined by Dunnett's Multiple Comparison Test of one-way ANOVA analysis using GraphPad Prism 5, ***- P<0.001; **- P<0.01; *-P<0.05; ns-P>0.05.

Activity and stability of EcA variants

Firstly, we determined the effects of the mutations on l-asparaginase activity under physiological conditions (37 °C, pH 7.4). No significant difference was found between the l-asparaginase activity of both the variants (KHY-17 and KHYW-17) and the wild-type EcA (Fig. 1B). Interestingly, difference in stability among the variants and wild-type EcA was observed at higher temperatures ranging from 55 °C to 70 °C. KHY-17 was found to be stable up to 70 °C, whereas, marked decrease in enzyme activity was observed in case of wild-type and KHYW-17 beyond 55 °C. (P<0.05; Fig. 1C).

Kinetic properties of EcA variants at a higher temperature

The specificity constant V_max/K_m, a measure of catalytic efficiency (K_cat), was calculated to compare the kinetic properties of the variants. Analysis at physiological temperature (37 °C) revealed no significant difference between KHY-17 and wild-type EcA (Fig. 1D), however, KHYW-17 was found to be catalytically superior to both of them. In addition, at a higher temperature of 62 °C, KHY-17 exhibited significantly (∼2.46 fold) higher%K_cat values than the wild-type (P<0.05; Fig. 1E). Though not statistically significant, KHYW-17 showed almost 1.86-fold increase in catalytic property compared to the wild-type EcA. The kinetic properties are summarized in Supplementary Table 2.

Glutaminase activity of EcA mutants

Both the EcA variants (KHY-17 and KHYW-17) showed approximately 24-fold reduced glutaminase activity compared to the wild-type EcA (P<0.05; Fig. 1F), which mirrors its safety use.

Anti-leukemic effect of EcA mutants on all cell lines and PBMCs isolated from healthy individuals

We investigated the anti-leukemic effect of the EcA variants in the leukemic cell line Reh, a B-ALL cell line by performing MTT assay. The reduction in Reh cell viability compared with untreated cells was 49.66 % for KHYW-17, 30.08 % for KHY-17 and 20.85 % for wild-type EcA (Figure S1). No significant difference in the viability of peripheral blood mononuclear cells (PBMCs) from healthy individuals was observed after treatment with both the EcA variants and wild-type EcA individually (Figure S2), indicating that the EcA variants exhibit selective cytotoxicity towards leukemic cells.

Immunogenicity of EcA mutants

Enzyme inactivation by eliciting anti-drug antibody production and strong allergic reactions are the major obstacles faced in the asparaginase treatment of ALL. Here, we checked for binding of anti-EcA antibody to KHY-17 and KHYW-17 by indirect ELISA. We observed 17.9 % reduction in binding of KHY-17 to anti-EcA antibody which was not statistically significant, however KHYW-17 showed significant reduction (39.42 %) in antibody binding compared to wild-type EcA (P<0.05; Fig. 1G).

To investigate the immunogenicity of the EcA variants, we measured the production of IgG and IgM antibodies in a total of three mice per group immunized with wild-type EcA and the two EcA variants individually. In alignment with the in-vitro ELISA results, it was observed following the second and third immunization of KHY-17 and wild-type EcA that IgG and IgM production was 1.4–1.6-fold lower (Fig. 2A, C, E) and 2.3–3.0-fold lower (Fig. 2B, D, F), respectively in the KHY-17 treated group. Similar results were also obtained with KHYW-17 variant showing 2.3–2.5-fold reduction in IgG (Fig. 2A, C, E) and 3.1–3.4-fold reduction in IgM production (Fig. 2B, D, F) with respect to wild-type EcA treated group. The IgG and IgM titers with a serum dilution from 1:100,000 to 1:800,000 are shown in the supplementary Figure S3.

Fig. 2.

Determination of immunogenicity of wild-type EcA and its variants in mice. Microtiter plates were coated with wild-type EcA and variants (5 µg/ml) in duplicate. The titer of IgG and IgM antibodies was determined by preparing different dilutions of serum- (A and B) 1:20,000, (C and D)1:40,000, and (E and F) 1:80,000 from mice subcutaneously immunized with wild-type EcA and variants were prepared and added to the wells. Goat anti-mouse IgM or anti-mouse IgG and HRP-conjugated polyclonal rabbit anti-goat were used as primary and secondary antibodies respectively. The titer of antibodies was determined by indirect ELISA method. Significance was determined by Bonferroni post-test of two-way annova using Graph pad prism 5, ***- P<0.001; **- P<0.01; *-P<0.05; ns-P>0.05.

Pharmacokinetic properties of EcA mutants

The activity vs time curve of KHY-17, KHYW-17 and wild-type EcA is depicted in Fig. 3A, 3B and 3C, respectively. The PK parameters of the EcA variants and wild-type EcA are summarized in Fig. 3D. Though no significant changes were observed in terms of total exposure given by the Area Under the Curve (AUC), and clearance of KHY-17, KHYW-17 and wild-type EcA, the mutants showed significantly higher volume of distribution (wild-type EcA- 7.95 ± 0.27; KHY-17- 19.95 ± 1.27; KHYW-17- 14.04 ± 1.33) and consequently longer half-life (wild-type EcA- 103.24 ± 18; KHY-17- 267.28 ± 9.74; KHYW-17- 167.41 ± 14.4) compared to wild-type.

Fig. 3.

Pharmacokinetic study of wild-type EcA and its variants. Activity vs time curve of (A) KHY-17, (B) KHYW-17 and (C) Wild-type EcA. (D) Pharmacokinetic properties of wild-type EcA and its variants (Geometric Mean ± SEM). Cmax- Highest concentration of the drug in blood; AUC- Area Under Curve; HL- Half Life; Vz- Volume of distribution; Cl- Rate of Clearance; MRT- Mean Residence Time. Time vs activity curve was plotted and pharmacokinetic parameters were calculated using PUMAS software. Significance was determined with respect to wild-type; Significance was determined by Dunnett's Multiple Comparison Test using GraphPad Prism 5, P<0.001 - ***; P<0.01 - **; P<0.05 - *; P>0.05 – ns.

Toxicity study

Single dose toxicity study

Acute toxicity of wild-type asparaginase and the variants was studied in Balb/C mice. No significant difference was observed in the levels of Blood Urea Nitrogen (BUN), Uric Acid (URCA), Total Protein (TP), Total Bilirubin (TBI), Aspartate Aminotransferase (AST) and Alanine Transaminase (ALT) even at the highest dose level in both the wild-type and the mutants in the acute toxicity study. However, serum creatinine (CRE) levels in mice treated with 2000 IU/Kg (acute) of KHYW-17 was significantly higher compared to vehicle control albeit within the range described by Charles River Laboratories [37]. Albumin levels were significantly affected at all dose levels by all the three formulations. The biochemical parameters of mice treated with single dose of wild-type EcA and variants are summarized in supplementary Table 3.

An important finding related to hematological parameters of acute study was the dose dependant decrease in WBC count in KHY-17 group. No significant effect was observed on any other parameter by any of the formulations. The hematological parameters of mice treated with single dose of wild-type EcA and variants are summarized in supplementary Table 4.

Histopathology showed mild-moderate degeneration in liver at all doses of wild-type EcA. Similarly, both mutants caused degenerative changes in the liver, but the extent of degeneration was less in the mutant groups compared to the wild-type at each dose (Fig. 4A). No other significant histopathological findings were observed in any other tissue (Supplementary Figures S4A, S5A, S6A, S7A, S8A and S9A).

Fig. 4.

Haematoxylin and Eosin (H&E) staining for histopathological examination of liver, collected from mice treated with EcA wild-type and variants.(A) In single dose (acute) toxicity study, wild type EcA and its variants were administered intravenously at a concentration of 500 IU/Kg, 2000 IU/Kg and 5000 IU/Kg and were observed for 14 days. The organs were collected on day 15. (B) In repeat dose or sub-acute toxicity, 400 IU/Kg, 800 IU/Kg and 1600 IU/Kg of designated l-asparaginase formulations were administered through intra-peritoneal route five days a week for 4 weeks. The organs were collected on day 29. All the images were taken using Carl Zeiss Axio Imager Z1 upright microscope with 20Χ magnification and 0.5 numerical aperture. Zen 3.0 software was used for data acquisition.

Repeat dose toxicity study

Repeat dose (sub-acute) toxicity was conducted on 100 Balb/C mice. All biochemical parameters were within the range in sub-acute toxicity study, details are summarized in Supplementary Table 5. The hematological parameters of mice treated with multiple intraperitoneal doses of wild-type EcA and the variants are summarized in Supplementary Table 6. No significant abnormalities were observed in any of the groups. Minimal to mild degree of degenerative changes in liver was observed in all the three formulations, consistent with the findings of acute toxicity study (Fig. 4B). No significant histopathological findings were made in any other organ (Supplementary Figures S4B, S5B, S6B, S7B, S8B and S9B). None of the animals suffered loss of weight due to the interventions, and the pattern of weight gained during the course of observation was not different between the groups (Supplementary Figure S10).

Efficacy study of EcA variants in xenograft all mouse model

We performed efficacy study in ALL xenograft mouse model constructed by injecting 5 × 10⁶ Jurkat (human ALL) cells. The percentage of leukemic burden was calculated by determining the percentage of human CD45 positive cells in mice blood using FACS. All mice in the untreated group died within 8 days due to leukemia burden (Fig. 5B). Notably, the mice treated with EcA variants showed about 60–70 % reduction in human CD45+ cells in peripheral blood stream compared to delay in progression but no reduction in human CD45+ cells in mice treated with wild-type EcA 3 days after 2nd dose. Moreover, the mice in the untreated group showed almost 100 % increase in the relative % of human CD45+ cells during this time period. Interestingly, it is worthy of attention that after final dose, 80–90 % reduction in human CD45+ cells were observed in mice treated with EcA variants compared to 40 % reduction in mice treated with wild-type EcA (Fig. 5A). A representative Figure to measure human and mouse CD45 positive cells by flow cytometry in the peripheral blood of leukemic mice treated with EcA variants, wild-type and the untreated group are shown in Supplementary Figures S11 and S12, respectively.

Fig. 5.

Efficacy of wild-type EcA and variants in cell line derived leukemic xenograft model. (A) Relative percentage of human CD45+ve cells by flow cytometry to determine the efficacy of wild-type EcA and variants in cell line derived leukemic xenograft model. All the mice in the untreated group were dead by day 8. (B) Kaplan meier survival analysis of the xenografted mice being treated with wild-type EcA and variants compared to the untreated group. (C) Pre and post treatment change in weight of all four groups of xenografted mice. Mean ± SEM are shown here. Significance was determined by Bonferroni post-test of two-way annova using Graphpad prism 5, ***- P<0.001; **- P<0.01; *-P<0.05; ns-P>0.05.

We also monitored the survival of mice till day 18 post treatment. Among the treatment groups, the survival rate was lowest in the group treated with wild-type EcA treatment (10 % survival rate), where 9 out of 10 mice were dead. Interestingly significantly higher survival rate was seen in groups treated with EcA variants. 90 % survival rate (9 out of 10 mice) was observed in the groups treated with both KHY-17 and KHYW-17 individually, in contrast, all 10 mice died by day 8 in the untreated group (Fig. 5B).

Daily physical monitoring of animals during and after therapy revealed some significant symptoms in the untreated group and the group treated with wild-type EcA such as weight loss, hind limb paralysis and hair loss. No such symptoms were observed in the mice treated with EcA variants. Initially weight loss was observed in all groups but eventually mice in the EcA variant treatment group recovered (Fig. 5C).

Binding of pre-existing antibodies in all patients to EcA mutants

Silent inactivation by circulating pre-existing anti-EcA antibodies remains a major challenge in the treatment of relapse and primary ALL patients receiving EcA therapy. Using an indirect ELISA method, we studied the binding of wild-type EcA and variants to pre-existing asparaginase antibodies in the sera of ALL patients. ALL patients were grouped as the one who were yet to receive l-asparaginase therapy (group B, n = 4), who were on l-asparaginase therapy for less than 7 days (group C, n = 3), primary ALL patients receiving l-asparaginase therapy for at least 60 days (group D, n = 14), and relapsed ALL cases (group E, n = 6). Healthy donors (group A, n = 4) were included as control. Patient demographics are described in Supplemental Table 7. Presence of anti-asparaginase antibodies is clearly unexpected in group A and B since they have not received any asparaginase therapy. Participants in group C are also unlikely to carry anti-asparaginase antibodies as they have started the treatment for less than a week. As expected, neither of the EcA variants or wild-type EcA showed detectable binding to IgG and IgM in sera obtained from healthy individuals, group A (Fig. 6A and F) or ALL patients in group B (Fig. 6B and G) and group C (Fig. 6C and H). However, Group D patients and relapsed patients of Group E are being anticipated to produce antibodies against the drug due to long-term treatment protocol. Interestingly, triple mutant KHYW-17 showed approximately 2–3-fold less binding to IgG and IgM antibodies in sera obtained from ALL patients in Group D compared with wild-type EcA (P<0.05; Fig. 5D and I). Furthermore, the same triple mutant KHYW-17 showed approximately 2-fold less binding to IgG and IgM antibodies in sera obtained from relapsed ALL patients (Group E) compared to wild-type EcA (P<0.05; Fig. 6E and J). In contrast, the antibody binding efficacy of the double mutant, KHY-17 and wild-type EcA were comparable in both the group D and Group E patients (Fig. 6D, I, E and J). Considering all data, we conclude that triple mutant KHYW-17 is less immunogenic and the best drug candidate for future clinical trials.

Fig. 6.

Binding of wild-type EcA and variants with pre-existing antibodies present in sera obtained from primary and relapsed ALL patients. Microtiter plates were coated in duplicate with wild-type EcA and variants (0.2 µg/well). Appropriate dilution of serum obtained from healthy individuals (A and F, n = 4), ALL patients without asparaginase therapy (B and G, n = 4), Primary ALL patients who were given asparaginase therapy for less than 7 days (C and H, n = 3) and more than 60 days (D and I, n = 14) and relapsed ALL patients (E and J, n = 6) were added to the wells. The binding of serum IgG (A-E) and IgM (F-J) with wild-type EcA and variants were determined as described in methods section. Each experiment was performed three times. Mean ± SEM are shown here. Significance was determined by Dunnett's Multiple Comparison Test of one-way ANOVA analysis using GraphPad Prism 5, ***- P<0.001; **- P<0.01; *-P<0.05; ns-P>0.05.

Discussion

Substitution of selected amino acid residues at dimer-dimer interfaces and within the immunogenic epitope regions of the wild-type EcA significantly increased thermal stability, reduced immunogenicity and glutaminase activity, and enhanced anti-leukemic activity while retaining l-asparaginase activity in variants.

Both KHY-17 and KHYW-17 showed comparable l-asparaginase activity and kinetic properties to wild-type EcA at 37 °C. However, at higher temperatures (55 ⁰C to 70 ⁰C), KHY-17 was most stable and the stability of KHYW-17 was comparable to wild-type EcA at temperatures of 55 °C and above, indicating that the third mutation reduced stability at higher temperatures. Higher thermal stability of KHY-17 could be extremely useful in hot humid countries of Asia and Africa, particularly when maintaining cold chain is difficult or not feasible due to frequent power outages. Knowing that conformational changes at subunit interface or EcA active site restrict the binding of glutamine's extra CH₂ group resulting in reduced glutaminase activity [5], we introduced a specific mutation, Y176F in both EcA variants, and both of them showed significantly diminished glutaminase activity while retaining their native l-asparaginase activity, which may confer safety attributes to the mutants. According to previous studies, the side effects of glutaminase activity are reduction in synthesis of proteins such as insulin, albumin, serum thyroxine binding globulin, serum sex hormone binding globulin, protein-S, protein-C and fibrinogen, leading to hypocholesterolaemia glycosuria, ketonic hyperglycaemia, prolonged bleeding and thrombosis [24,38]. ALL patients undergoing asparaginase therapy suffer venous thromboembolism at a reported incidence of 13 % to 36.7 % [39], [40], [41] and hyperglycemia is reported in up to 23 % [42]. In the single dose toxicity study, we clearly observed a relative albumin sparing effect in the KHY-17 group, compared to the wild-type EcA. Of course, all reported adverse events of wild-type EcA did not manifest in the experimental animals probably due to interspecies differences in the occurrence of these reactions. However, the noticeable albumin sparing effect of the double mutant may translate into a clinically meaningful benefit in terms of glutaminase-related adverse events which needs to be validated in a prospective clinical trial. Due to its bacterial origin, l-asparaginase induces anti-drug antibody formation leading to its silent inactivation in ALL patients [43]. We found that combining the mutations, Y176F and K288H reduced antigenicity. Here, KHYW-17 produced significantly less IgG and IgM antibodies in immunized mice. Furthermore, it showed significantly less binding to pre-existing anti-EcA antibodies in primary and relapsed ALL patient's undergoing prolonged EcA therapy, perhaps due to conformation changes of antibody binding regions over the incorporation of mutations in the EcA epitope regions. These findings have important clinical implications, as EcA administered to many sensitive and relapsed ALL patients are neutralized by pre-existing antibodies, thereby making the therapy ineffective. Thus, it is envisaged that mutant KHYW-17 can be used in both primary and relapsed ALL.

It is desirable to have EcA formulations with longer half-lives to avoid frequent administration and thereby prevent immunogenicity and achieve durable remissions [44]. KHY-17 and KHYW-17 mutants had significantly longer half-lives compared to the wild-type EcA. The change in half-life was mainly brought about by an increase in the volume of distribution of the mutants compared to wild-type, while the clearance was unaffected. A similar behavior in humans would result in longer duration of time for which plasma asparaginase activity will remain above 100 IU/L, which is the target threshold trough concentration for asparaginase. This would mean that even a less frequent administration of the asparaginase mutants will have similar effect on PD markers such as asparagine depletion compared to 72 hourly administrations of the wild-type. Also, since the volume of distribution is larger, the peak plasma concentrations achieved with the mutants are significantly less compared to the wild-type (Fig 3A). This is likely to reduce the incidence of dose dependent toxicities such as thrombosis, pancreatitis, hypersensitivity (subclinical and clinical), liver dysfunction and encephalopathy [45]. However, the reason for this pharmacokinetic behavior, just with change in 2–3 amino acid residues, is not clear.

Prolonged native asparaginase therapy causes life-threatening hepatotoxicity or microvesicularsteatosis by disrupting pivotal mitochondrial protein synthesis that participates in hepatic β-oxidation of fatty acids and secretion of very low-density lipoproteins (VLDL particles) from endoplasmic reticulum (ER) and Golgi apparatus of the hepatocytes [46,47]. These also lead to minimal to moderate changes in hepatocytes which were seen in histopathological examinations. Low levels of albumin in mice treated with EcA and its variants is in line with clinically known phenomenon of asparaginase activity [48]. Interestingly, the effect was more pronounced in the single dose toxicity study wherein the formulations were administered intravenously (i.v.), compared to the repeat dose toxicity study with intraperitoneal (i.p.) administration. Only the highest doses in the repeat dose study caused noticeably hypoalbuminea, statistically significant only in the KHYW-17 arm, although KHY-17 was not too far behind. We hypothesize that the higher maximum concentrations achieved on i.v. dosing could possibly explain this finding, although we do not have pharmacokinetic data to support this hypothesis. We resorted to i.p. administration for the repeat dose toxicity study to avoid risk of phlebitis in the tail vein and being unable to complete the i.v. schedule over a 28-day period. Incidentally, decrease in albumin levels correlates with asparaginase activity and is a predictor of overall survival and disease free survival in ALL [48].

Not surprisingly, in the cell line derived leukemia xenograft model, we found both KHY-17 and KHYW-17 showed better efficacy in comparison to wild-type EcA. We observed 80–90 % reduction in human CD45+ cells in the groups treated with EcA variants compared to just 40 % in the group treated with wild-type EcA. The groups treated with both the EcA variants showed 90 % survival rate compared to just 10 % survival rate of the group treated with wild-type EcA. Comprehensively both the EcA mutants emerged as promising candidates for future clinical trials and ALL therapy due to their remarkable in vivo efficacy property against ALL xenograft model.

Conclusion

Our targeted protein engineering approach has led to the generation of EcA variants with reduced immunogenicity, glutaminase activity, inability of the variants to bind to pre-existing antibodies, and longer serum half-lives. Considering all findings, the EcA variant KHYW-17, emerges as a promising candidate for the treatment of primary and relapsed ALL. Further studies on toxicity, efficacy, and safety under GLP conditions and subsequent clinical trials are underway to establish the clinical utility of the EcA variant KHYW-17 as a viable option for the treatment of ALL.

Funding

This work was supported by DST-SERB Grants, Government of India (File No. SB/SO/HS/203/2013 and CRG/2019/001593/BHS), BIG-BIRAC Grant of DBT, Government of India (BT/TEMP2159/BIG-07/15) and BRNS-DAE Grant (Sanction No. 37(1)/14/22/2018-BRNS) to Avinash Sonawane. ICMR-SRF and UGC—NET (JRF and SRF) fellowships awarded to Soumika Sengupta and Mainak Biswas, respectively is highly acknowledged.

CRediT authorship contribution statement

Soumika Sengupta: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Project administration, Methodology, Investigation, Formal analysis, Conceptualization. Mainak Biswas: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Methodology, Formal analysis, Data curation, Conceptualization. Khushboo A. Gandhi: Writing – review & editing, Validation, Software, Methodology, Formal analysis, Data curation. Saurabh Kumar Gupta: Writing – review & editing, Visualization, Validation, Methodology, Data curation. Poonam B. Gera: Writing – review & editing, Visualization, Validation, Formal analysis. Vikram Gota: Writing – review & editing, Writing – original draft, Validation, Supervision, Resources, Project administration, Investigation, Formal analysis, Data curation, Conceptualization. Avinash Sonawane: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Resources, Project administration, Investigation, Funding acquisition, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.