Background
L-asparaginase (EC 3.5.1.1.)
E. coli type II (EcA) has been widely used for acute lymphoblastic leukemia treatment for more than 30 years. The mechanism of antileukemic activity is believed to be directly related to the hydrolysis of L-asparagine and subsequent significant depletion of L-asparagine concentration in blood and death of cells that are not able to express asparagine synthetase or have low level of expression [
1]. It is known that hypersensitivity, including several allergic reactions and even anaphylactic shock, are among the most dangerous side effects of EcA treatment [
2]. Moreover, even if a patient doesn’t have severe hypersensitivity symptoms, the development of anti-EcA antibodies could minimize the efficacy of treatment due to the alteration of pharmacokinetics and elimination the enzyme from the blood.
A number of approaches to prevent the immunogenicity of L-asparaginases has been investigated extensively: chemical modifications of epitopes to reduce the immunogenicity, site-directed mutagenesis of amino acid residues to diminish immunogenicity without reduction of enzymatic activity, consequent use of L-asparaginases with different immunogenic properties. The most common way is a chemical modification of the protein, for instance, conjugation with polyethylene glycol (PEG). It was shown that pegylation significantly increases the half-life (T
1/2) of EcA in human serum, from 1.24 ± 0.17 to 5.73 ± 3.24 days [
3,
4]. However, if the patient is allergic to native EcA, the T
1/2 of pegylated EcA also decreases, which is due to identical antigenic epitopes [
5]. Besides pegylation, several approaches have demonstrated the ability to decrease immunogenicity, namely encapsulation into liposomes 158–180 nm [
6], immobilization into a biocompatible polyethylene glycol-albumin hydrogel [
7], formulation of L-asparaginase load poly(lactide-to-glycolide) nanoparticles [
8], chemical modification by N, O-carboxymethyl chitosan [
9], etc. For example, encapsulation of palmitoyl-asparaginase into liposomes increases the T
1/2 of native EcA for at least eight times [
10]. Conjugation with low-immunogenic and non-toxic proteins, for instance silk fibroin, or encapsulation into red blood cells in vitro could be used for hiding the epitopes of L-asparaginases and, therefore, increase of T
1/2 [
11‐
14].
Epitope mapping and subsequent production of antigenically modified enzymes can be considered as the second efficient method to minimize immunogenicity. It has been proved that the major antigenic epitope of
Erwinia chrysanthemi (ErA) is
282GIVPPDEELP
287, and replacement Pro
285 with Thr
285 has led to 8-fold decrease of the immunogenicity of the native enzyme [
15]. However, the immune response for large proteins is usually complex, and antibodies produced are usually polyclonal, hence the replacement of one amino acid can’t prevent the formation of antibodies against modified protein in hypersensitive mice previously treated with native one.
The third method that has been used is the consecutive administration of L-asparaginases with different antigenic properties. It is known that anti-EcA antibodies do not interfere with the ErA pharmacokinetics, which is the reason why ErA can be effectively used in patients previously treated with EcA [
16‐
18]. Administration of new L-asparaginases, that do not have cross-reactivity with EcA and ErA, could be an effective approach for treatment of hypersensitive patients who have received multiple doses of EcA and/or ErA. It has been proven that
Helicobacter pylori L-asparaginase has different immunogenic properties from EcA in mice [
19]. During the last 7 years we’ve obtained and evaluated the enzymatic and anticancer properties of a few recombinant L-asparaginases from different origin, namely
Wollinella succinogenes type II (WsA),
Yersinia pseudotuberculosis type II (YpA),
Erwinia carotovora type II (EwA), and
Rhodospirillum rubrum type I [
20‐
22].
The aims of this study were the evaluation of the immunogenicity in mice and cross-reactivity between these L-asparaginases and EcA, and elucidation of its structural basis based on analysis of three-dimensional structures.
Methods
Structure-based multiple sequence alignment of EcA, WsA, EwA, YpA and RrA was constructed using PROMALS3D [
23]. Three-dimensional structure of the WsA, EwA, YpA and RrA proteins were modeled by I-TASSER [
24]. Epitopes experimentally verified for ErA [
15] were projected on the structure-based alignment using sequence-to-profile alignment method implemented in Clustal Omega [
25]. Bioinformatics prediction of epitopes were made by Discotope [
26], ElliPro [
27] and EPSVR [
28]. UCSF Chimera [
29] was used for visualization of the 3D structure of enzymes. Solvent accessibility of the tetramer was calculated by DSSP [
30].
Reagents
L-asparagine (Reanal, Hungary); Na2HPO4, NaH2PO4, KCl (Serva, Germany); NaCl (Merck, Germany); Tween-20, NaHCO3, Na2CO3 (Sigma-Aldrich, USA). We used standard buffers: PBS, pH 7.4; carbonate-bicarbonate buffer 0.1 M, pH 9.5, citrate-phosphate buffer, 0.1 M, pH 5.0, PBS-Tween 0.05 %.
Enzymes (antigens)
We used the commercially available lyophilized EcA preparation (Medak, Germany, 10 000 IU per vial); lyophilized recombinant EwA, that is similar to ErA (Additional file
1: Figure S1,
http://purl.org/phylo/treebase/phylows/study/TB2:S18796) [
22], lyophilized recombinant YpA [
20], lyophilized recombinant RrA, obtained from IBMC RAMS [
21], and lyophilized recombinant WsA, obtained from GosNIIgenetika.
Animals
Female C57Bl6j 8–12 weeks old mice were used for the in vivo studies. Mice were kept in animal facility of N.N. Blokhin Cancer Research Center of RAMS. All animal studies were carried out using procedures in compliance with EU (European Convention for the Protection of Animals Kept for Experimental and other Scientific Purposes, Strasbourg, 1985; 86/609/EEC and 2010/63/EU) directives on the protection of animals used for scientific purposes and according to institutional policy on the care and use of laboratory animals. The animal studies were approved by the local ethics committee of I.I. Mechnikov Institute of vaccine and sera, the decision from 26/01/2015.
Immunogenicity studies
To evaluate the IgM immune response, groups of mice (five mice per group) received one i.v. injection of 500 μg of each preparation. To evaluate the IgG immune response 300 μg of each L-asparaginase were administered i.v. 3 times, at 2-week intervals. 0.9 % sodium chloride solution was injected in a separate groups of animals as controls (five mice per group). Blood samples were collected 7 days after the last immunization. Plasma samples for ELISA assay were obtained by centrifugation at 400 g at 4 °C for 10 min and stored at −80 °C until analysis. Then the samples were incubated for 10 min, centrifuged at 10,000 rpm (Eppendorf 1500) at 24 °C and were used immediately for experiments.
Serum Ab responses were determined using an ELISA kit (Maxisorb, Nunc). Briefly, standards, controls and pre-diluted samples of 100 μl of different L-asparaginases in carbonate-bicarbonate buffer, 5 μg/ml, were added into the wells of a 96-well plates and incubated at + 4 °C for 12 h. The plate was washed several times for 2 min with 300 μl phosphate buffer saline containing 0.05 % Tween 20 (PBS-Tween).
Serial dilutions (1:50 to 1:256,000) of mouse plasma were prepared. Following a 1 h incubation at 37 °C, the wells were washed as described above and the residual activity was measured. Serum Ab binding was detected using polyclonal secondary goat Ab to mouse total IgG or IgM. We used pre-diluted GoatAnti mouse IgG or GoatAnti mouse IgM, HU ADS biotin conjugate (Invitrogen, Cat# M30115) 1:10000, 100 μl in each well, and streptavidine conjugate of horseradish peroxidase STREPTAVIDIN HRP (AbD Serotec) 1:10000, according to manufacturer’s instructions. After suitable washing, standard buffer for ELISA was added and the plate was incubated for 15 min at room temperature. The reaction was stopped by adding 50 μm of 1.8 M H2SO4 before measuring the optical density at 450 nm using a plate reader Multiscan FC.
For IgG response the serum was considered negative (−) if geometric mean of the titers (GMT) was <50, positive: 51–600 (+), 601–30,000 (++), and >30,000 (+++). For IgM response the serum was considered negative (−) if GMT was <50, positive: 51–125 (+), 126–300 (++), and >300 (+++).
Statistical analysis
Geometric mean of the titers (GMTs) and the GMT ratios with corresponding 95 % confidence intervals (CIs) were calculated by taking the antilog of the mean of the logе-transformed data (assuming that logе-transformation of the titers follows a normal distribution).
To compare immunogenicity of different antigens we used GMT of each antigen reacting with the serum sample from mice immunized with the same antigen. For instance, EcA + anti-EcA vs YpA + anti-YpA. To prove statistical validity of cross-immunogenicity between groups we used GMT of each antigen reacting with the serum sample from mice immunized with the different antigen vs antigen reacting with the serum sample from control mice (no immunization).
SPSS 21 software was used for statistical analysis. One-way ANOVA was used to compare immunogenicity of antigens. Post hoc Dunnett’s T3 test was performed to assess differences between the individual groups. Calculations began with the logarithmic transformation of the antibody titers. P value <0.05 was considered statistically significant.
Discussion
We performed a comparative analysis of structure and sequence of the bacterial L-asparaginase epitope regions, which were predicted by bioinformatics methods and projected from the experimentally known epitope regions of the homologous proteins. Particularly, we analyzed whether the presumable epitope regions differ among considered enzymes and whether these differences are associated with known structure and sequence elements that influence immunogenicity. Analysis of the secondary structure showed that the secondary structure is highly conserved among considered enzymes and most of the presumable epitopes regions are located in loops that is in accordance with structural studies [
32]. Protrusion and solvent accessibility of the considered loops – two properties that are known for affecting immunogenicity – are sufficiently similar among the enzymes, except for N-terminal active site loops of the RrA, which are much shorter than the same loops of other enzymes. We also found a moderate differences among enzymes in a length of the inter-domain linker region.
Analysis of differences in amino acids associated with increased immunogenicity – tyrosine, tryptophane and charged amino acids – showed more differences than the structural analysis. Among these differences the most correlated with experimental data is the difference among enzymes in enrichment of tyrosines and charged amino acids in the multi-subunit interface region. However, structural properties of this region – a presence of two beta-strands and a moderate solvent accessibility – is not typical for epitopes. Apart from the predicted epitopes the N- and C-terminal ends of protein could be also considered as putative epitopes when they are long enough to possess enough protrusion and flexibility – two characteristic epitope properties. We speculate that a long N-terminus of YpA enzyme (see sequence alignment in Fig.
2), which is also enriched at the beginning with charged aminoacids and tyrosine (KYIK), could be a reason of higher immunogenicity of YpA in comparison with other considered enzymes. Overall, although the recognized structural and sequence differences in putative epitope regions among five considered L-asparaginases do not fully explain experimental observation of the immunogenicity of the enzymes, the performed analysis set the foundation for further research in this direction when more data are available.
There are two approaches to describe the cross-reactivity of antibodies: either L-asparaginase could be considered as an antigen for immunization or the enzyme is defined as a target for antibodies. The obtained data showed that after single and threefold administration all investigated L-asparaginases displayed moderate immunogenicity with low cross-reactivity. The antibodies produced in mice after EcA administration have the worst capacity for development of cross-reactivity with other enzymes. This phenomenon could be explained mainly by low immunogenicity of EcA, confirming the current position of EcA as a first-line L-asparaginase for acute lymphoblastic leukemia treatment. Any other enzyme, preferably WsA, YpA or RrA could be used as a second-line treatment without a risk of significant alteration of pharmacokinetics due to antibodies formation. However, EcA could be a target for antibodies produced in mice previously immunized with any other L-asparaginase. It could be explained both by higher immunogenicity of these enzymes and presence of similar epitopes in EcA and other enzymes, that was shown by analyses of three-dimensional structures. The affinity of cross-binding with EcA and other L-asparaginases vary directly as the average immunogenicity of each enzyme: YpA showed the highest cross-reactivity, that can be explained by more than 74 % homology of the sequences of EcA and YpA. The differences in cross-reactivity of each L-asparaginase evaluated by these two different approaches could be explained by different location of antigenic epitopes (internal or external epitopes).
The clinical relevance of our findings can be evaluated after comprehensive preclinical investigation of safety and efficacy of obtained enzymes and consequent clinical trials. The clinical pathway requires comparative trials of efficacy and safety of new L-asparaginases in patients with acute lymphoblastic leukemia, previously treated with EcA.
Conclusions
The performed study showed different immunogenic properties of L-asparaginases and confirmed the possibility of replacement of EcA with L-asparaginase from different origin. YpA was most immunogenic enzyme compared to EcA, EwA, WsA and RrA in murine model. Based on these data, the preferable enzyme for the second line treatment is WsA, RrA or EwA. Further investigations of the epitopes and immunogenicity of novel L-asparaginases are needed to assume their therapeutic applications or create the mutated proteins with minimized immunogenicity in order to increase the efficacy of leukemia treatment.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
VSP conceived the study; MDK performed bioinformatics analysis; VSP and IND carried out the immunoassay, MPV and SSA participated in data analysis and interpretation. All authors have read and approved the final manuscript.