1 Introduction

l-asparaginase (ASNase) is considered a key therapeutic enzyme for the treatment of acute lymphoblastic leukemia (ALL) nowadays. This disease is recognized as the most frequent cancer among children [1], also affecting adults [2]. ALL consists of the overproduction of undifferentiated lymphocytes in the bone marrow, which can lead to death [3]. As part of the multiple therapy for the treatment of ALL, the enzyme ASNase has enabled remission rates over 90% to be reached [4]. Leukemic cells, unlike normal ones, are unable to synthesize asparagine in sufficient quantities, so they incorporate this amino acid from the circulation [5]. ASNase hydrolyzes the circulating amino acid, exerting its therapeutic action by starving the tumor cells [6]. There are currently four ASNase formulations available for clinical use from the bacterial source Escherichia coli [7,8,9] and Dickeya chrysanthemi [10]. Additionally, bacterial asparaginases present glutaminase activity, which has been reported to be responsible of several side effects in patients [11,12,13]. Notwithstanding their pharmacological effectiveness, these formulations generate a wide range of adverse effects in patients [14,15,16]. However, the side effects best documented for this biopharmaceutical likely correspond to immunogenicity and allergenicity [17,18,19] due mainly to its prokaryotic nature [20]. The occurrence of hypersensitivity reactions has been related to the generation of antibodies, either IgG [21,22,23,24] or IgE type [25,26,27]. On the other hand, some reports have indicated that ASNase from E. coli is more allergenic than its equivalent from D. chrysanthemi [28, 29]. The patient’s immune response to ASNase not only results in undesirable effects at a clinical level, but also in decreased enzyme activity [30]. This is because the antibodies generated as part of the humoral response contribute to the inactivation of the enzyme, thus decreasing its half-life in circulation [31, 32].

One of the strategies adopted to reduce the immunogenicity caused by ASNase has been the search for new microorganisms capable of producing the enzyme with a lower risk of causing side effects in patients [33,34,35]. In this regard, fungal sources have been explored for their eukaryotic nature, as their similarity to humans at cellular level could reduce the unwanted immunological reactions [36, 37]. Several reports [38,39,40,41] and comprehensive reviews [20, 42,43,44] have been published, revealing the potential of fungi as a feasible source of ASNase and presenting antitumor activity [45, 46]. In addition to being eukaryotes, they can produce ASNase with reduced glutaminase activity [47, 48]. In spite of a high antitumor activity having been demonstrated in vitro using fungal enzymes [49,50,51], the immunological implications in patients have been scarcely explored.

The use of bioinformatics tools for evaluating immunological responses from an in silico approach has been addressed in many studies in recent years [52,53,54,55,56,57]. Several have focused on explaining antigenic epitopes in the structure of various proteins, including ASNase, for subsequent modification and reduction of immunogenicity [58,59,60,61]. All this knowledge in the bioinformatics arena offers an opportunity to discover and develop new safer and more effective biopharmaceuticals.

In this work, we present the first in silico study that describes l-asparaginase from nine experimentally tested filamentous fungi species in terms of their predicted immunogenicity profile. Our objective is to show the structural features that could be responsible for generating an immunological reaction in patients, after the administration of fungal ASNase, also predicting for the first time the composition of the peptides that contribute to the allergenicity of fungal ASNase. These findings are of great significance in the development of therapeutic asparaginases from new sources.

2 Materials and methods

2.1 l-asparaginase sequence dataset

For this study, the amino acid sequences of nine fungal l-asparaginases were downloaded from the National Center for Biotechnology Information (NCBI) Protein Database (https://www.ncbi.nlm.nih.gov/protein) and the UniProtKB Database (https://www.uniprot.org/). Table 1 shows the filamentous fungi selected for the study, whose activity has previously been proven. Additionally, the amino acid sequence of Escherichia coli type II (EcA) l-asparaginase was downloaded from the Protein Data Bank (https://www.rcsb.org/). All the sequences appear in Supplementary Material A.1.

Table 1 Filamentous fungi used for the in silico immunogenicity analysis of l-asparaginases in the current research
Full size table

2.2 T-cell epitope prediction and epitope density determination

For T-cell epitope prediction the NetMHCII 2.3 Server (http://www.cbs.dtu.dk/services/NetMHCII/) was used [57], setting the default parameters of the program. The peptides predicted as “strong binder” (≤ 2%) and “weak binder” (≤ 10%) were selected as immunogenic epitopes. For epitope density computation, the relative frequency was used: \(f_{i} = n_{i} /N = n_{i} /\sum {_{j} n_{j} }\), where \(n_{i}\) is the number of immunogenic epitopes predicted, and \(N\) is the total number of epitopes determined by the program (immunogenic and non-immunogenic). The HLA-DRB1*01:01, HLA-DRB1*03:01, HLA-DRB1*04:01, HLA-DRB1*07:01, HLA-DRB1*08:01, HLA-DRB1*11:01, HLA-DRB1*13:01 and HLA-DRB1*15:01 alleles were used for the prediction because these are reference alleles with a high distribution worldwide [71, 72].

2.3 Prediction of epitope allergenicity

In order to perform the allergenicity prediction of the nine fungal asparaginases, the AllerTOP v. 2.0 server (http://www.ddg-pharmfac.net/AllerTOP/) was used [73]. For this purpose, each of the T-cell epitopes predicted as immunogenic for the HLA-DRB1*07:01 allele for the nine fungal and E.coli ASNase were evaluated using this program. The program evaluates the sequence of peptides and returns the results as “probable allergen” or “probable non-allergen.” The relative frequency of the predicted allergenic epitopes was calculated, dividing the number of immunogenic epitopes for the HLA-DRB1 * 07: 01 allele over the total immunogenic epitopes previously determined as described for each ASNase in Sect. 2.2.

2.4 Modeling, refinement and quality assessment of tertiary structures of fungal ASNase

The sequences of the nine fungal ASNases were used for homology modeling of their tertiary structures, with the Phyre2 program [74]. Subsequently, the models were refined with the GalaxyRefine tool [75], and the quality of the generated models was assessed with the MolProbity tool [76], available in Swiss-Model.

2.5 B-cell epitope prediction and mapping

We performed the B-cell epitope prediction (linear epitopes) by using the ElliPro server (http://tools.iedb.org/ellipro/) [77] from the IEDB Analysis Resource database. For this analysis, we used the monomeric structures of the nine fungal ASNase previously modeled. The epitope mapping in the modeled structures of fungal asparaginases was carried out with PyMOL (PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC).

2.6 Statistical analysis of the epitope density

A comparative analysis of epitope density of the l-asparaginases was performed by using a one-way ANOVA with Tukey’s test a posteriori. Probability values (p ≤ 0.05) were considered significant. The analysis was carried out with the GraphPad Prism package version 5.0 for Windows (GraphPad Software, San Diego, California, USA).

3 Results and discussion

The identification of immunogenic epitopes in the protein structure by computational tools has been a strategy used to achieve a reduction in the immunogenicity of several proteins [78,79,80]. It is also the case of ASNase from E. coli, where the use of bioinformatics techniques has allowed engineering the enzyme with the aim of obtaining an improved therapeutic agent [58, 81, 82]. On the other hand, the search for novel sources such as yeast and fungi has provided a new niche for obtaining less immunogenic ASNase [34, 83,84,85]. In our work, we address the problem of asparaginase immunogenicity through the evaluation of enzymes from nine filamentous fungi using computational tools. For this, we applied the measure known as relative frequency, since it has been used in previous studies to determine the immunogenic epitope density present in several proteins, correlating this measure with their immunogenicity level [86,87,88,89,90,91].

In assessing the degree of asparaginase immunogenicity, we found that there were no significant differences in the predicted immunogenicity between the nine fungal ASNases evaluated (Fig. 1). However, when comparing the relative frequencies of these asparaginases with that obtained for the enzyme from E. coli, interestingly, the predicted immunogenicity degree of fungal asparaginases was equivalent to the bacterial enzyme and even higher in the particular cases of Penicillium digitatum (Fig. 1). These results are remarkable: first, because they constitute the only report to date that refers to a study of the immunogenicity of fungal l-asparaginase, and second because they contradict expectations as to the possible safety of this enzyme derivative from eukaryotic cells. However, considering the analysis performed, an immune response analogous to that of the E.coli ASNase could be likely if the enzyme derived from one of these evaluated species were used. Therefore, although it is promising to obtain ASNase with an activity from the nine species of fungi reported, this activity will presumably be affected after recognition by the immune system, which could lead to subsequent neutralization, as in the case of bacterial ASNase. The predicted immunogenic T-cell epitopes are presented in Table B.1 as Supplementary Material.

Fig. 1
figure 1

Statistical analysis of the relative frequency of predicted immunogenic T-cell epitopes in fungal and bacterial asparaginases for eight alleles (HLA-DRB1*01:01, HLA-DRB1*03:01, HLA-DRB1*04:01, HLA-DRB1*07:01, HLA-DRB1*08:01, HLA-DRB1*11:01, HLA-DRB1*13:01 and HLA-DRB1*15:01). Significant differences were observed only between Penicillium digitatum versus Escherichia coli (p = 0.05). Error bars represent the standard deviation of the data set

Full size image

Since HLA-DRB1 alleles confer high-affinity binding to ASNase epitopes causing several allergic reactions [92, 93], we compared the relative frequency of T-cell epitopes for the eight alleles evaluated. Significant differences were observed between the HLA-DRB1*01:01 allele versus the HLA-DRB1*08:01, HLA-DRB1*11:01, HLA-DRB1*13:01 and HLA-DRB1*15:01 alleles, respectively (Fig. 2). In previous research, the HLA-DRB1*01:01 allele has been linked to the occurrence of hypersensitivity reactions associated with the administration of the antiretroviral drugs nevirapine and efavirenz, used in the treatment of patients infected with HIV [94,95,96]. This allele was also associated with the occurrence of hypersensitivity reactions producing liver damage after administration of the nonsteroidal anti-inflammatory lumiracoxib [97]. However, this is the first report that highlights the association between the HLA-DRB1*01:01 allele and the probable occurrence of immunological reactions against asparaginases, particularly those obtained from fungal sources. In our study, significant differences were also observed between the HLA-DRB1*04:01 allele and the HLA-DRB1*11:01 and HLA-DRB1*13:01 alleles, respectively (Fig. 2). HLA-DRB1*04:01 allele has previously been associated with an immune system response in the case of the therapeutic protein interferon-ß [98].

Fig. 2
figure 2

Relative frequency of predicted allergenic T-cell epitopes for the eight alleles evaluated and the nine fungal asparaginases, by using the AllerTOP v. 2.0 server. Significant differences were observed between the HLA-DRB1*01:01 allele versus the HLA-DRB1*08:01, HLA-DRB1*11:01, HLA-DRB1*13:01 and HLA-DRB1*15:01 alleles. In addition, significant differences were observed between the HLA-DRB1*04:01 allele versus the HLA-DRB1*11:01 and HLA-DRB1*13:01 alleles (p = 0.0010). Error bars represent the standard deviation of the data set

Full size image

By contrast, it has been well established that the HLA-DRB1*07:01 allele is associated with a high risk of hypersensitivity reactions after treatment with bacterial ASNase, possibly due to it being a high-binding allele [93, 99]. In our work, we performed a predictive analysis of the relative frequency of allergenic T-cell epitopes for the HLA-DRB1*07:01 allele in the nine fungal asparaginases, comparing them with the enzyme from E. coli. Surprisingly, a high epitope density was observed for fungal asparaginases in comparison with the bacterial one, underscoring the highest values for the Penicillium chrysogenum and Aspergillus terreus species (Fig. 3). The predicted allergenic T-cell epitopes are summarized in Supplementary Table B.2. Hypersensitivity reactions caused by E. coli ASNase have been widely studied in vivo due to their high incidence [100,101,102,103]. However, fungal ASNase has not been explored in this regard.

Fig. 3
figure 3

Relative frequency of allergenic T-cell epitopes for the HLA-DRB1*07: 01 allele for the ten microorganisms studied. A high epitope density was observed for fungal asparaginases

Full size image

B-cell epitope identification has been a widely used approach in peptide-based vaccine design and disease diagnosis [104,105,106]. Additionally, we identify the linear B-cell epitopes present in the nine fungal and E.coli ASNase. Previously, we modeled the monomeric three-dimensional structure of fungal ASNase, since these are not crystallized or reported. We present the predicted models in Fig. 4, as well as the quality assessment of these models. As can be seen in the Ramachandran plot, all the models present more than 95% of their amino acid residues in favorable regions, suggesting that the predicted models have a high quality. Supplementary Material C.1 summarizes the linear B-cell epitopes identified in the monomeric structure of the nine fungal and E. coli ASNase, while Fig. 5 shows the mapping of the B-cell epitopes identified. In general, the results show that the immunogenic B-cell epitope distribution in fungal ASNase as well as their length (Supplementary Table C.1) is very similar to the pattern in E.coli. This finding reinforces the fact that the immune response generated by the fungal ASNase could be very similar to that generated by E.coli in patients, as suggested from the results obtained in the prediction of T-cell epitopes. All these reports offer new insights about the choice of fungal ASNase as a possible alternative for the treatment of ALL because, although they are attractive as producers of the enzyme, the way in which the immune system of patients could respond to their administration is detrimental to its quality at industrial level.

Fig. 4
figure 4

Monomeric three-dimensional models predicted for the nine fungal ASNase and quality assessment using Ramachandran plots. All predicted models showed a quality higher than 95%

Full size image
Fig. 5
figure 5

Mapping of predicted B-cell epitopes in nine fungal asparaginases and E.coli

Full size image

4 Conclusions

This research is the first study to determine the immunogenicity of fungal l-asparaginases using an in silico approach, also showing the allergenic peptides predicted in the nine fungal enzymes studied. ASNase from fungi showed high immunogenicity patterns. This knowledge is important in the search for new sources of asparaginases and demonstrates the usefulness of bioinformatics tools in the discovery of immunological features for the design of safer biopharmaceuticals.