Peptides derived from phage display libraries as potential neutralizers of Shiga toxin-induced cytotoxicity in vitro and in vivo

Authors


Abstract

Aims

To use the phage display technique to develop peptides with the capability to neutralize the cytotoxicity induced by Stx1 and Stx2 toxins produced by Shiga toxin-producing Escherichia coli (STEC).

Methods and Results

The phage display technique permitted the development of three peptides, named PC7–12, P12–26 and PC7–30, which bind to the globotriaosylceramide (Gb3) receptor for Shiga toxins produced by STEC. Moreover, these peptides were capable of competing efficiently with the Shiga toxins for binding to Gb3. The peptides described herein partially inhibited the Stx-induced cytotoxicity of cell-free filtrates of STEC O157 : H7 and purified Stx toxins in Vero cells. The inhibition of lethality induced by Stx toxins in mice indicated that peptide PC7–30 inhibited the lethality caused by Stx1 (2LD50) in mice.

Conclusions

The phage display technique permitted the development of peptides that inhibited the cytotoxicity induced by Stx toxins in vitro. Peptide PC7–30 inhibited the lethality of Stx1 in vivo; this molecule would be a promising candidate for the development of therapeutic agents for STEC-related diseases in humans.

Significance and Impact of the Study

The selection of Gb3, the common receptor for Stx1 and Stx2, may contribute to the development of efficient neutralizers for both toxins, and our approach would be an interesting alternative for the development of therapeutic molecules for the treatment of diseases caused by STEC strains.

Introduction

Shiga toxin-producing Escherichia coli (STEC), including the O157 : H7 serotype, are a subgroup of bacterial strains that produce the toxins Stx1 or Stx2 or one of their variants. These pathogens may cause diarrhoea and haemorrhagic colitis in humans, and diseases that are often complicated by subsequent potentially fatal systemic sequelae, such as neurological damage and haemolytic uraemic syndrome (HUS), the leading cause of acute renal failure in children (Kaper et al. 2004). It is estimated that there are 63 153 cases of O157 and 112 752 cases of non-O157 STEC illnesses in the USA annually (Scallan et al. 2011). Between May and July 2011, several European countries, particularly Germany, experienced one of the largest STEC-HUS outbreaks ever reported (Robert Koch Institute 2011). Stx1 and Stx2 are potent cytotoxins and constitute the main virulence factors in STEC strains. Stx is a multi-subunit AB5 protein complex in which the A subunit is responsible for the toxic effects and the B subunit promotes binding to specific eukaryotic cell types (Trachtman et al. 2012).

The initial step in the pathogenesis of STEC is the binding of the B subunit of Stx to the glycolipid (Gal(α1-4)Gal(β1-4)Glc(β1-1)Cer) (Gb3) receptor on the cell membrane. This step is followed by retrograde transport of the toxin via the Golgi apparatus to the endoplasmic reticulum before the toxin is translocated to the cytosol. In this process, the A subunit is cleaved, and the A1 fragment, which has RNA N-glycosidase activity, is released. This fragment cleaves a specific N-glycosidic bond in the 28S rRNA, causing protein synthesis to cease (Melton-Celsa and O'Brien 1998; Scheiring et al. 2010; Petruzziello-Pellegrini and Marsden 2012).

The natural history of STEC infection indicates that severe symptoms in HUS patients develop in approximately 1 week in most acute cases (Tarr et al. 2005). Many researchers have attempted to develop an effective treatment for Stx-mediated HUS, but a specific therapy has not yet been developed. Previous studies have focused on preventing severe Stx-mediated symptoms via the development of diverse neutralizing agents, such as molecules mimicking the Gb3 receptor (Tesh et al. 1993; Lingwood 1996; Miura et al. 2006) and anti-Stx IgY (Neri et al. 2011), among other alternatives. However, although there is indirect evidence that human vaccination against STEC may be effective in preventing STEC-mediated illness, at present, there are no human vaccines or therapeutics to treat human STEC infections (Kaper et al. 2004). Moreover, although STEC strains are generally susceptible to a variety of antibiotics, diverse studies have shown that the use of antibiotics negatively alters the outcome of STEC infections, increasing the incidence of HUS in humans (Grif et al. 1998). El Sayed Zaki and El-Adrosy (2007) confirmed that most strains are susceptible to different antibiotics in vitro but that these antibiotics could increase both the release of Shiga-like toxin and toxin synthesis during the induction of lysogenic toxin-producing bacteriophages at sublethal concentrations (Köhler et al. 2000).

Although diagnostic reagents have been developed for the early detection of Stx and different types of antibodies have been developed for potential passive immunization, it is unclear whether the administration of anti-Stx therapeutics would be effective after symptoms have developed in humans (Sauter et al. 2008).

Some peptides derived from library screenings often modulate the target's activity in vitro and in vivo and can be used in drug design and as alternatives to antibodies (Molek et al. 2011). Moreover, therapeutic peptides have several advantages over antibodies, such as lower manufacturing costs, higher activity per unit mass, generally lower immunogenicity and better organ or tumour penetration (Vlieghe et al. 2010).

Thus, in this study, we aimed to develop synthetic peptides that neutralize the cytotoxicity caused by Shiga toxins both in vitro and in vivo. These peptides bind to the Gb3 peptides and were selected using biopanning peptide phage display libraries expressing peptide-linked filamentous phage M13.

Materials and methods

Bacterial strains

The following strains were used: E. coli O157 : H7 EDL933 (stx1+/stx2+) and E. coli C600 K12 from the Bacterial Collection of Laboratório de Fatores de Virulência em Bactérias (Instituto de Biologia -UNICAMP-São Paulo- Brazil) and E. coli K-12 ER2738 (New England Biolabs, Ipswich, MA, USA).

Phage display libraries

The phage display peptide libraries used in this study were a loop-constrained heptapeptide library (Ph.D.™-C7C) and linear dodecapeptide library (Ph.D.™-12). These libraries were purchased from New England Biolabs.

Shiga toxins and globotriaosylceramide

Purified Shiga toxins (Stx1 and Stx2) were purchased from the Phoenix Lab at Tufts Medical Center (Boston, MA, USA), and globotriaosylceramide (Gb3) was purchased from Matreya LLC (Pleasant Gap, PA, USA).

Phage display panning procedures

Globotriaosylceramide (Gb3) was diluted in 0·1 mol l−1 NaHCO3, pH 8·6 (50 μg ml−1), and 150 μl was adsorbed to one well of a 96-well polystyrene microtitre plate (Nunc Maxisorp, Naperville, IL, USA) for each phage display library. The plate was incubated at 4°C overnight, and wells were then blocked by adding 300 μl of blocking buffer (0·1 mol l−1 NaHCO3, pH 8·6, 5 mg ml−1 bovine serum albumin (BSA)), and the plate was incubated for 2 h at 37°C. The plate was washed six times with TBST (50 mmol l−1 Tris-HCl pH 7·4, 150 mmol l−1 NaCl + 0·1% (v/v) Tween-20), and 100 μl of diluted phage libraries (2 × 1011 pfu for the Ph.D.C7C library and 1 × 1011 pfu for the Ph.D. -12 library) was added to each well. The plate was incubated for 1 h at room temperature with gentle agitation. Unbound phages were removed from the wells by washing 10 times with TBST. Bound phages were eluted with 100 μl of Gb3 (100 μg ml−1 of Gb3 in TBST), and after 10 min at room temperature in the Gb3 solution, the eluate was collected and stored at 4°C. The titres of eluted phages were determined, and phages were amplified according the manufacturer's instructions provided with the Ph.D. libraries (NEB). This panning procedure was repeated for a total of three rounds for both the Ph.D.-C7C and Ph.D. -12 libraries. For each subsequent round of panning, the input number of phages was the same as the first round. The stringency of selection was increased using 0·5% Tween-20 in TBS for subsequent rounds to reduce the frequency of nonspecific phage binding.

Isolation of phage M13 clones, DNA isolation and sequencing

One millilitre aliquots of an overnight culture of E. coli K12-ER2738, diluted 1 : 100 in Luria Bertani broth (LB) supplemented with tetracycline (20 μg ml−1), were dispensed in sterile tubes. Phage plaques from the titration plates of the final round of selection were stabbed with sterile pipette tips and inoculated into an E. coli ER2738 bacterial suspension. These tubes containing bacterial suspensions were incubated for 4·5 h at 37°C with vigorous shaking (250 rpm) and were then transferred to microcentrifuge tubes and centrifuged at 10 000 × g for 2 min to pellet bacterial cells. The supernatants were decanted, and DNA was isolated according to the manufacturer′s instructions. The quality of isolated DNA was assessed by agarose gel electrophoresis.

DNA samples were sequenced at the Laboratório de Biologia Molecular de Plantas e Genoma of the Centro de Biologia Molecular e Engenharia Genética (UNICAMP-São Paulo-Brazil). All amino acid sequences were aligned using MEGA software (ver. 5.2 (www.megasoftware.net)). After discarding repeated sequences, those sequences sharing at least three amino acids at the same position were selected for a phage ELISA-binding assay.

Phage ELISA-binding assay

One hundred microlitres of Gb3 (100 μg ml−1 in 0·1mol l−1 NaHCO3, pH 8·6) was added to each well of a 96-well microtitre plate (Nunc Maxisorp), and the plates were incubated overnight at 4°C. The plates were washed three times with PBST (PBS + 0·05% (v/v) Tween-20) and blocked for 1 h at room temperature with PBST containing 1·5% (w/v) BSA. Previously selected phage clones were amplified and titred according to the phage libraries' manufacturer′s instructions and diluted in PBST. One hundred microlitres of phage solutions (1·5 × 1012 pfu ml−1) was added to Gb3-sensitized wells, and the microplates were incubated for 1 h at room temperature. The microplates were washed three times in PBST, and 100 μl of mouse anti-M13 bacteriophage horseradish peroxidase-conjugated monoclonal antibody (GE Healthcare, Uppsala, Sweden), diluted to 1 : 2000 in PBST, was added to each well. The microplates were incubated for 1 h at room temperature and were then washed three times in PBST. Subsequently, 100 μl of o-phenylenediamine dihydrochloride (OPD) substrate solution (400 μg ml−1) was added to each well. The absorbance of the end product was measured at a wavelength of 492 nm using an ELISA microplate reader (Epoch; BioTek Instruments Inc., Winooski, VT, USA). Raw absorbance data were corrected by subtraction of the values obtained from an OPD-containing blank. This assay was performed in triplicate.

Synthesis of PC7–12, P12–26 and PC7–30 peptides

Life Technologies (Life Technologies, São Paulo, Brazil) synthesized all of the peptides for the in vitro and in vivo assays, according to the phage display libraries' manufacturer's instructions.

Competition ELISA

One hundred microlitres of Gb3 solution (100 μg ml−1) was added to each well of 96-well microtitre plates (Nunc Maxisorp), and the plates were incubated overnight at 4°C. The plates were washed three times with PBST and blocked for 1 h at 37°C with PBST containing 1·5% w/v BSA. One hundred microlitres of peptides dissolved in sterilized distilled water was twofold diluted in PBS until achieving 100, 50, 25 and 12·5 μg of each peptide in the wells. The microplates were incubated for 1 h at 37°C and were washed three times in PBST. After this step, Stx toxins (Stx1 or Stx2) were added at 500 pg per well. Subsequently, the plates were washed three times in PBST, 100 μl of the corresponding murine monoclonal antibody against Stx (1 : 1000) was added and the plate was incubated at 37°C for 2 h. Again, the plates were washed three times in PBST, and 100 μl of horseradish peroxidase-conjugated goat anti-mouse IgG monoclonal antibody (Southern Biotechnology, Birmingham, AL, USA) was added to each well. The microplates were incubated at 37°C for 1 h and were washed three times in PBST. After this, 100 μl of OPD substrate solution was added to each well. The absorbance of the end product was measured at 492 nm using an ELISA microplate reader (Epoch; BioTek Instruments Inc.). The results were expressed as percentage of competition (%competition), where ‘%competition’ was the decrease in the absorbance caused by the toxins in the absence of peptides. These results were described as follows: ‘competition’ if the %competition was >75%, ‘relative competition’ when the %competition was between 75 and 25% and ‘noncompetition’ when %competition was <25% (Xu et al. 2006). This assay was performed in triplicate, and the statistical analysis was conducted using two-way anova followed by the Bonferroni test with GraphPad Prism Software ver. 5.00 (GraphPad Software, San Diego CA, USA).

In vitro cytotoxicity and neutralization assays

Preparation of E. coli O157 : H7 EDL933 cell-free filtrate

The E. coli O157 : H7 EDL933 strain was cultured in tryptic soy broth (TSB) and incubated overnight at 37°C with shaking. The bacterial culture was centrifuged at 10 000 × g for 10 min, and the supernatant was filtered through a 0·22-μm pore-sized membrane filter (Xia et al. 2010).

Inhibition of cytotoxicity assay in cultured cells

MTT reduction cytotoxicity assay

The capacity of synthesized peptides to inhibit the cytotoxic activity of Stx toxins was assessed using the MTT (3-(4,5-dimethylthiazole)-2,5-diphenyltetrazolium bromide thiazole blue) cytotoxicity assay in Vero cells (African green monkey kidney cells). First, 96-well microtitre plates were seeded with approximately 104 Vero cells per well and incubated at 37°C for 24 h in a 5% CO2 atmosphere. The tissue culture medium (Eagle's minimum essential medium (EMEM) containing 10% foetal bovine serum) was discarded and replaced with 100 μl of fresh medium without foetal bovine serum. One hundred microlitres of twofold serial dilutions of peptides diluted in EMEM (200 μg per well) was added to wells of Vero cells, and the plates were incubated at 37°C for 2 h in a 5% CO2 atmosphere. Subsequently, 100 μl of Stx solutions (500 pg per well) and 100 μl of E. coli O157 : H7 EDL933 cell-free filtrate were added to each well containing peptide dilutions and incubated at 37°C for 43 h in a 5% CO2 atmosphere. After this step, 100 μl of MTT (2 mg ml−1 in PBS) was added to each well, and the plate was incubated for 3 h at the same temperature and in the CO2 atmosphere. Finally, one MTT solution was discarded, and 100 μl of isopropyl alcohol-hydrochloric acid 1 N (24 : 1 v/v) was added to each well. Absorbance was measured at 450 nm using an ELISA microplate reader (Epoch; BioTek Instruments Inc.). The E. coli O157 : H7 (stx1+/stx2+) cell-free filtrate was used as a positive control for cytotoxicity. The percentage of neutralization was calculated using the following formula (Schmitt et al. 1991):

display math

where A(peptide+toxin) is the absorbance of the well containing peptide plus toxin, A(toxin) is the absorbance of the well containing only toxin and A (nontreated cells) is the absorbance of the well containing only cells. This assay was performed in triplicate.

In vivo lethality assays

All of the in vivo lethality assays were performed in mice according to the guidelines of the Ethics Committee for Animal Experimentation (Institute of Biology – UNICAMP- São Paulo- Brazil).

Determination of the Shiga toxin 50% lethal dose in mice

The 50% lethal dose (LD50) values for the Stx1 and Stx2 toxins in Swiss mice were determined as previously described (22). Twenty male mice, each weighing 25–35 g, were separated into four groups (five animals per group) for the determination of LD50 for each toxin. Mice received intraperitoneal (i.p.) injections of Shiga toxins diluted in PBS (pH 7·4) at the following concentrations: 12·5, 125 and 1250 ng (for Stx1) and 1, 10 and 100 ng (for Stx2). One group inoculated only with PBS was included as the survival control. Mice were then monitored for survival over a period of 7 days. The LD50 was determined using the formula:

log LD50 = log(highest dose tested) + (log D)[(1/2) − (ΣR/N)], where D is the fold difference between successive doses, ΣR is the total number of dead mice and N is the number of animals per group (Rahal et al. 2011).

Lethality inhibition assay in mice

This assay was performed with Swiss mice weighing 25–35 g, separated into four groups as follows: survival control (inoculated only with PBS), Shiga toxin control (inoculated only with Stx1 or Stx2 toxins), peptide control (inoculated only with PC7–12 or P12–26 or PC7–30 peptides) and toxin + peptide group (received toxin and peptide). Each group included five mice. The groups that received toxin were inoculated with Stx toxin corresponding to 2LD50. The peptide control group received only 200 μg of peptide diluted in PBS. The survival of mice was monitored over a period of 7 days.

Statistical analysis

For the inhibition of cytotoxicity and lethality assays, the experimental and control groups were analysed with the Duncan multiple range test using the S.A.S. statistical software package (SAS Institute Inc 1986).

Results

Phage display panning procedures

The alignment of the predicted amino acid sequences of the inserts expressed by clones selected for the phage ELISA-binding assay is shown in Fig. 1.

Figure 1.

Amino acid sequences of inserts expressed by phage clones selected for the ELISA assay to determine whether these clones bind to the Gb3 receptor. The amino acids of the insert shared by the phage clones at the same position are shown in bold, and the amino acids belonging to the linker sequences for each phage library are underlined (Ph.D.C7C and Ph.D.12).

Phage ELISA Gb3-binding assay

The binding of the selected phage clones to purified Gb3 was assayed. Clones Ph.D.C7C-12, Ph.D.12–26 and Ph.D.C7C-30 exhibited strong binding to Gb3. In contrast, modest binding values were observed for the other selected clones (Fig. 2).

Figure 2.

Phage ELISA Gb3-binding assay. The phage clones P.h.D.C7C-12, P.h.D.12–26 and P.h.D.C7C-30 exhibited higher absorbance values, and the amino acid sequences of their displayed inserts were used for the synthesis of peptides tested in subsequent assays in vitro and in vivo. The error bars indicate the standard deviations from the means.

Based on the results of this assay, three peptides were synthesized and named PC7–12, P12–26 and PC7–30. These peptides corresponded with the amino acid sequences of inserts displayed by phages P.h.D.C7C-12, P.h.D.12–26 and P.h.D.C7C-30, respectively (Fig. 1).

Competition ELISA to determine binding to the Gb3 receptor

The results observed for each peptide competing with Stx1 and Stx2 are shown in Fig. 3(a,b), respectively. With the Stx1 toxin, we observed that peptide PC7–30 exhibited a higher competition percentage at the concentration of 50 μg per well than did peptides P12–26 (P < 0·05) and PC7–12 (P < 0·001). When the three peptides were assayed together against the Stx2 toxin, significant differences were not observed (P > 0·05) when examining competition percentages for peptides P12–26 and P7–30 at the concentration of 50 μg per well. All peptides exhibited similar competition percentages at the concentration of 100 μg per well.

Figure 3.

Competition ELISA for binding to Gb3. (a) Competition between peptides and the Stx1 toxin for binding to purified Gb3; (b) Competition between peptides and the Stx2 toxin for binding to Gb3. The error bars indicate the standard deviations from the means. image_n/jam12451-gra-0001.png PC7–12; image_n/jam12451-gra-0002.png P12–26; image_n/jam12451-gra-0003.png PC7–30.

The highest competition percentages for binding to Gb3 between the toxins and peptides were observed at peptide concentrations ranging from 25 to 100 μg per well.

Neutralization of the Stx cytotoxic activity in Vero cells

The Stx cytotoxic activity neutralization assay indicated that when Escherichia coli O157 : H7 cell-free filtrate was applied to Vero cells, the inhibition of cell death was similar for the three peptides at the different concentrations assayed (Fig. 4). The only exception was the inhibition caused by the peptide P12–26, which was higher than the inhibition caused by peptides PC7–12 and PC7–30 at the concentration of 50 μg per well (P < 0·0001). Peptide concentrations higher than 100 μg per well resulted in decreased cytotoxicity inhibition rates (data not shown).

Figure 4.

Neutralization of cytotoxic activity induced by the cell-free filtrate of Escherichia coli O157 : H7 in Vero cells. The error bars indicate the standard deviations from the means. ■ E. coli O157 + peptide PC7–12; image_n/jam12451-gra-0004.png E. coli O157 + peptide P12–26; □ E. coli O157 + peptide PC7–30.

When purified Stx toxins were assayed, higher inhibition percentages were observed for Stx2 when compared with Stx1 (P < 0·0001; Fig. 5). There were no significant differences for the inhibition of cytotoxicity among the peptides assayed with the Stx1 toxin. For Stx2 toxin, a higher inhibition value was observed for peptide PC7–30 at the concentration of 50 μg per well. However, a dose-dependent inhibitory response was not observed in this assay.

Figure 5.

Neutralization of cytotoxicity induced by purified Stx toxins in Vero cells. (a) Neutralization by peptide PC7–12; (b) Neutralization by peptide P12–26 and (c) Neutralization by peptide PC7–30. The error bars indicate the standard deviations from the means. ■ Stx1 toxin + peptide PC7–30; image_n/jam12451-gra-0004.png Stx2 toxin + peptide PC7–30.

Determination of the Shiga toxin 50% lethal dose in mice

According the methodology used, the theoretical LD50 values calculated for Stx1 and Stx2 in mice were 40 and 32 ng, respectively.

Inhibition of lethality induced by Stx in mice

The results for the in vivo inhibition of lethality using peptides PC7–12, P12–26 and PC7–30 are shown in Fig. 6(a,b,c) (for mice injected with Stx1 toxin) and Fig. 6(d,e,f) (for mice injected with Stx2 toxin).

Figure 6.

Inhibition of lethality induced by purified Stx toxins in mice. Each group of mice was injected with Stx toxins (2LD50) and 200 μg of each peptide as follows: (a) Stx1 toxin + peptide PC7–12; (b) Stx1 toxin + peptide P12–26; (c) Stx1 toxin + peptide PC7–12; (d) Stx2 toxin + peptide PC7–12; (e) Stx2 toxin + peptide P12–26 and (f) Stx2 toxin + peptide PC7–30. (a) image_n/jam12451-gra-0005.png Survival control, image_n/jam12451-gra-0002.png Stx1 toxin control, image_n/jam12451-gra-0006.png peptide PC7–12 control, image_n/jam12451-gra-0007.png Stx1 toxin + peptide PC7–12; (b) image_n/jam12451-gra-0005.pngSurvival control, image_n/jam12451-gra-0002.png Stx1 toxin control, image_n/jam12451-gra-0006.png peptide P12–26 control, image_n/jam12451-gra-0007.png Stx1 toxin + peptide P12–26; (c) image_n/jam12451-gra-0005.png Survival control, image_n/jam12451-gra-0002.png Stx1 toxin control, image_n/jam12451-gra-0006.png peptide PC7–30 control, image_n/jam12451-gra-0007.png Stx1 toxin + peptide PC7–30; (d) image_n/jam12451-gra-0005.png Survival control, image_n/jam12451-gra-0002.png Stx2 toxin control, image_n/jam12451-gra-0006.png peptide PC7–12 control, image_n/jam12451-gra-0007.png Stx2 toxin + peptide PC7–12; (e) image_n/jam12451-gra-0005.png Survival control, image_n/jam12451-gra-0002.png Stx2 toxin control, image_n/jam12451-gra-0006.png peptide P12–26 control, image_n/jam12451-gra-0007.png Stx2 toxin + peptide P12–26; (f) image_n/jam12451-gra-0005.png Survival control, image_n/jam12451-gra-0002.png Stx2 toxin control, image_n/jam12451-gra-0006.png peptide PC7–30 control, image_n/jam12451-gra-0007.png Stx2 toxin + peptide PC7–30.

Peptides PC7–12 and P12–26 partially inhibited the lethality caused by Stx1 in mice (P < 0·0023), whereas peptide PC7–30 abolished the lethality induced by Stx1 (P < 0·0007). In contrast, none of the peptides assayed inhibited the lethality caused by the Stx2 toxin in mice. Moreover, none of the peptides themselves induced lethality in mice.

Discussion

The use of peptides as therapeutic molecules is considered an important goal in medicine because peptides can mimic a variety of endogenous agents, leading to their highly selective binding to specific receptors (Mantzourani et al. 2008).

In this study, we describe the use of the peptide phage display technique for the development of three peptides, PC7–12, P12–26 and PC7–30, that exhibited the capacity to neutralize the cytotoxic effects induced by the Stx toxins. These peptides showed high specificity for their target and exhibited strong binding to the Gb3 receptor, leading to the inhibition of Shiga toxin-induced cytotoxicity. The use of phage display libraries to develop neutralizing agents for Stx-induced cellular damage was examined previously by Miura et al. (2006), who reported that peptides mimicking the Gb3 receptor resulted in strong neutralization of Shiga toxins in vitro. The use of phage display libraries to develop Stx neutralizers was also reported by Stone et al. (2007), who described the use of pentavalent antibodies (pentabodies) isolated from phage display libraries that bound to the wild-type Stx1-B subunit and neutralized the cytotoxicity caused by Stx1.

When the amino acid sequences of the peptides PC7–12, P12–26 and PC7–30 (Table 1) were compared against the NCBI protein database using blast (http://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins), no sequence homology with Shiga toxins was identified. According to Heins and Quax (2010), the chances of successful selection of peptide receptor modulators can increase when using a library of random peptides because peptide ligands that are not homologous to the primary structure of the natural ligand can be identified.

Table 1. Amino acid sequences of synthesized peptides PC7–12, P12–26 and PC7–30
PeptideAmino acid sequenceaPredicted molecular weight (Da)
  1. a

    The underlined letters represent the amino acids forming the specific linker sequences for each phage display library. The amino acids corresponding to the variable region of the phage display library are represented in bold.

PC7–12 AC PNNTISL CGGGS 1292·48
P12–26 SAPRHNVPDNPR GGGS 1617·69
PC7–30 AC PLTTKTL CGGGS 1308·52

Additionally, peptide-based Stx-neutralizers with clustered trisaccharides are effective in STEC infection experiments in mice; however, the clinical application of these molecules has been substantially hampered by the synthetic complexity of the trisaccharide moiety, which possesses very low affinity for Stx (Kd = 1 × 10−3 mol l−1; Nishikawa 2011).

The peptides described herein exhibited different competition percentages when assayed against the Stx toxins. All peptides could compete with the Stx1 toxin, while for the Stx2 toxin, three peptides had high %competition values, higher than 80% in the case of P12–26, confirming the specificity of this peptide for binding to the Gb3 receptor and its potential to inhibit cytotoxicity in Gb3-expressing cells.

Specific binding to the Gb3 receptor is a critical characteristic in developing neutralizing molecules for Stx-induced cytotoxic damage because cell cytotoxicity and likely in vivo pathological damage are mediated by the binding of Stx toxins to the Gb3 receptor (Lingwood 1996).

The results obtained in the competition ELISA assay confirm the validity of the selection of phage clones using immobilized Gb3 because all peptides tested displayed the capacity to compete with an Stx toxin for binding to Gb3 receptor. The peptide P12–26, expressed by the P.h.D.12–26 phage clone, displayed high competition percentages against the Stx1 and Stx2 toxins and exhibited a high level of inhibition of the cytotoxicity induced by the E. coli O157 : H7 EDL933 cell-free filtrate, demonstrating the specificity of this peptide for binding to Gb3. The importance of focusing attention on the Gb3 receptor as an important target for the development of neutralizing molecules against Stx-induced damage was highlighted by Silberstein et al. (2009). These authors proposed that the inhibition of Gb3 synthesis in target cells using a glucosylceramide synthase inhibitor could be a potential treatment for protection against the pathological effects of Shiga toxin producing HUS, suggesting that the Gb3 receptor may be an effective alternative target for developing therapeutic agents against STEC infections.

Yamada et al. (2006) described the development of antimicrobial peptides fused with Gb3-mimicking peptides with neutralizing activity against Stx1 (50 pg ml−1) and observed an increase in cell survival of approx. 10%, whereas we observed higher inhibition percentages, ranging from 11·49 to 16·62% for Stx1 and from 19·99 to 28·04% for Stx2, applying tenfold higher concentrations of Stx1 or Stx2.

Nishikawa et al. (2006) screened a tetravalent peptide library to select compounds that bind to the Stx2 B subunit and identified peptide motifs that bind to the trisaccharide-binding site 3 with high affinity and effectively inhibit Stx2 cytotoxicity. PPP-tet, a tetravalent peptide, almost completely protected mice from a fatal dose of E. coli O157 : H7.

The peptides described in this study partially protected Vero cells against the cytotoxic effects of Shiga toxins. This effect was higher for Stx2 in vitro, which is an interesting result because Stx2 is more closely associated with severe disease and the development of HUS than is Stx1 (Fuller et al. 2011). In primates, the administration of Stx2 alone can produce the symptoms of HUS, whereas the administration of Stx1 at the same dose does not cause this syndrome (Stearns-Kurosawa et al. 2010).

Another strategy to neutralize the cytotoxic effects of Stx toxins is the use of anti-Stx antibodies. Smith et al. (2009) showed that the 11E10 antibody partially neutralized the cytotoxic activity of Stx2 and some Stx2 variants; however, the neutralization of Stx1-induced cytotoxicity using this antibody was not described. Our results highlight the importance of developing neutralizing agents against both types of Stx toxins because outbreaks caused by STEC producing Stx1 alone in humans have occurred, and the incidence of infections caused by Stx1-producing E. coli is becoming increasingly important (Hashimoto et al. 1999; Sukhumungoon et al. 2011). Moreover, the effectiveness of treatments with antibodies against Stx can decline rapidly when these antibodies are administered after the second day of infection (Matise et al. 2001; Casadevall 2002). Furthermore, the use of polyclonal antisera against either Stx1 or Stx2 has the drawback that each specific antiserum does not cross-neutralize the other toxin (Strockbine et al. 1986; Wen et al. 2006). For these reasons, we believe that our strategy, together with other scientific approaches, may help to develop more effective molecules to overcome the limitations of passive immunization.

The peptides PC7–12, P12–26 and PC7–30 exerted protective effects against the in vitro cytotoxicity induced by purified Stx1 or Stx2 toxins, and these effects were confirmed when the cell-free filtrate of E. coli O157 : H7 (producing Stx1 and Stx2) was assayed in Vero cells. The inhibition of cytotoxicity induced by E. coli O157 : H7 cell-free filtrate was higher than the inhibition associated with each purified Stx toxin, suggesting that these peptides exert a protective effect against the two toxins. The peptide P12–26 exhibited the highest inhibition rates, suggesting that its linear structural conformation may be advantageous for inhibiting the interaction between the toxins and the Gb3 receptor in vitro, while the other two loop-constrained PC7–12 and PC7–30 peptides showed lower inhibition rates in Vero cells.

The administration of the Stx2 toxin (2LD50) killed all of the mice in the Shiga toxin control group; however, when Stx1 (2LD50) was applied, only three of five mice died following injection of this toxin. These divergent results between the theoretical LD50 for Stx1 and the expected lethality in vivo may be related to the fact that Stx1 and Stx2 bind to Gb3Cer with similar specificity, though the toxins exhibit differential association and dissociation properties. The Stx2 toxin binds to Gb3Cer more slowly than Stx1. However, once bound, Stx2 is difficult to dissociate in vitro using surface plasmon resonance-based analysis (Nakajima et al. 2001). These differential binding properties may affect the expected toxicities of Stx1 and Stx2 in vivo (Louise and Obrig 1995; Müthing et al. 2009).

The peptide PC7–30 prevented lethality in mice injected with Stx1 (2LD50) but did not exhibit activity against lethality induced by Stx2 (2LD50). These differences in biological activity could be associated with the specific binding characteristics of each Stx toxin to the Gb3 receptor. Previously, it was reported that the interactions between Stx1 and Stx2 and the Gb3 receptor are clearly distinguishable (Nyholm et al. 1996). Moreover, mutational analysis of the StxB subunit indicated that Stx1 requires all of the trisaccharide-binding sites (sites 1, 2 and 3) for high-affinity binding to Gb3 (Soltyk et al. 2002; Nishikawa et al. 2005), whereas Stx2 requires the interactions with sites 1 and 3 but not site 2 (Nishikawa et al. 2005). We speculate that the peptide PC730 would interact with binding site 2 of the Gb3 receptor because this molecule inhibits the lethality caused by Stx1 but does not alter the lethality induced by Stx2. These differences are consistent with the results of the ELISA competition assay in which this peptide also displayed higher competitive values against Stx1 toxins.

According to Gallegos et al. (2012), it is not clear whether Gb3 is the main factor mediating Stx binding to host cells, and in vitro binding affinities may not correlate with cellular or in vivo toxicity. We believe that the in vitro and in vivo differences in biological activity of the peptides tested may also be associated with their structural properties. Camacho et al. (2008) suggested that linear peptides are inherently unstable and noted that many peptides with pharmacological activities have been proven to be ineffective in vivo. Additional factors affecting the biological activity of peptides include their rapid renal clearance or may be associated with their small size, poor metabolic stability and biodegradability as a result of enzymatic degradation by proteases and peptidases of the blood, liver and kidney (Olmez and Akbulut 2012). Based on these factors influencing the in vivo response to Stx toxins, the in vivo loss of biological activity of the linear peptide P12–26, which displayed significant inhibition of Stx2 toxicity in vitro but did not inhibit lethality in vivo, may be explained. Nevertheless, some of the limitations on peptides as therapeutic agents may be overcome by introducing structural modifications to develop analogous molecules with improved pharmacological characteristics (Lien and Lowman 2003; Funke and Willbold 2009). Short peptides generally display high selectivity and specificity for their targets and have low systemic toxicity, and peptides that bind to target receptors with reasonable affinity and specificity can be considered to be good alternatives to antagonistic antibodies (Molek et al. 2011).

The use of antibodies against Stx to neutralize cytotoxicity was also reported by Sauter et al. (2008). These authors showed that the administration of the anti-Stx2 11E10 antibody could prevent the progression to HUS in mice exposed to Stx2. However, these authors observed that the protective activity of this antibody was dependent on the time of administration.

The loop-constrained peptide PC7–30 abolished the lethality caused by Stx1 in mice, in contrast with the linear peptide P12–26, which exhibited partial inhibition, suggesting that the linear structure is more unstable and that its half-life might be shorter in vivo. This hypothesis is supported by the fact that some peptides have a short half-life in vivo, on the order of minutes, and the biological activity of a peptide directly depends on its stability in serum (McGregor 2008; Vlieghe et al. 2010). One method to improve biological activity would be to incorporate D-enantiomeric amino acids in the synthetic peptides (Ladner et al. 2004; Funke and Willbold 2009). According to Ladner et al. (2004), to have a biological effect, a therapeutic peptide must either cover the active site of the target or distort the target enough to deprive it of its usual activity.

We consider our approach to be an interesting alternative method to identify novel neutralizing agents against Stx-induced toxicity because both Stx1 and Stx2 and all of their described variants, including Stx2e, which binds preferentially to Gb4, bind to the Gb3 receptor. The Gb3 molecule is considered to be the exclusive receptor for the Stx toxins in vivo and mediates the tissue damage and pathological manifestations caused by these toxins (Okuda et al. 2006). The approach described herein is likely to lead to new research challenges because the manner in which Shiga toxins bind their glycolipid receptors is extremely complex (Peter and Lingwood 2000).

Acknowledgements

We would like to thank Prof. Aricio Xavier Linhares for his technical support. This study was supported by Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP).

Conflict of interest

No conflict of interest is declared.

Ancillary