Oral Intoxication of Mice with Shiga Toxin Type 2a (Stx2a) and Protection by Anti-Stx2a Monoclonal Antibody 11E10

Abstract

Shiga toxin (Stx)-producing Escherichia coli (STEC) strains cause food-borne outbreaks of hemorrhagic colitis and, less commonly, a serious kidney-damaging sequela called the hemolytic uremic syndrome (HUS). Stx, the primary virulence factor expressed by STEC, is an AB₅ toxin with two antigenically distinct forms, Stx1a and Stx2a. Although both toxins have similar biological activities, Stx2a is more frequently produced by STEC strains that cause HUS than is Stx1a. Here we asked whether Stx1a and Stx2a act differently when delivered orally by gavage. We found that Stx2a had a 50% lethal dose (LD₅₀) of 2.9 μg, but no morbidity occurred after oral intoxication with up to 157 μg of Stx1a. We also compared several biochemical and histological parameters in mice intoxicated orally versus intraperitoneally with Stx2a. We discovered that both intoxication routes caused similar increases in serum creatinine and blood urea nitrogen, indicative of kidney damage, as well as electrolyte imbalances and weight loss in the animals. Furthermore, kidney sections from Stx2a-intoxicated mice revealed multifocal, acute tubular necrosis (ATN). Of particular note, we detected Stx2a in kidney sections from orally intoxicated mice in the same region as the epithelial cell type in which ATN was detected. Lastly, we showed reduced renal damage, as determined by renal biomarkers and histopathology, and full protection of orally intoxicated mice with monoclonal antibody (MAb) 11E10 directed against the toxin A subunit; conversely, an irrelevant MAb had no therapeutic effect. Orally intoxicated mice could be rescued by MAb 11E10 6 h but not 24 h after Stx2a delivery.

INTRODUCTION

Shiga toxin (Stx)-producing Escherichia coli (STEC) strains are food-borne pathogens with an estimated infectious dose of fewer than 50 organisms (1). While multiple STEC serotypes are associated with disease, illness associated with infection by E. coli O157:H7 accounts for over 63,000 of the 113,000 total STEC cases each year in the United States (2). Bovine and other ruminants are the natural carriers of STEC, and contamination of meat generally occurs during beef processing, with up to 40% of the outbreaks occurring from beef (3, 4). However, contaminated fresh produce is also responsible for both outbreaks and sporadic cases of STEC in the United States (4, 5).

Upon STEC infection, the most common disease manifestation is hemorrhagic colitis. In 5 to 15% of patients, a serious sequela of STEC infection, the hemolytic uremic syndrome (HUS), may occur (6, 7). The HUS is characterized by a triad of symptoms: microangiopathic hemolytic anemia, thrombocytopenia, and acute kidney failure (8). Currently, there is no vaccine to prevent or therapeutic agent to cure STEC infection, as antibiotics are contraindicated due to the potential upregulation of bacteriophage production of Stx (9).

An STEC strain may encode Stx1a (equivalent to Stx from Shigella dysenteriae type 1) and/or Stx2a, two antigenically distinct but biologically similar toxins (10, 11). Stx1a and Stx2a share about 57% amino acid homology, analogous crystal structures, and identical modes of action (12). Stx1a and Stx2a are AB5 toxins. The A subunit is responsible for the catalytic activity of the toxin molecule, and the B subunit, a homopentamer, is required for the toxin to bind to the Stx host cell receptor, globotriaosylceramide (Gb3) (54, 55). Once bound to its receptor, Stx undergoes retrograde transport through the cell; the enzymatically active portion of the toxin is then released into the cytoplasm, where it depurinates a single adenine residue from the 28S rRNA of the 60S ribosome (13, 14). This ribosomal injury results in the inhibition of protein synthesis and, ultimately, cell death (15). Although Stx1a and Stx2a have the same receptor and mode of action, epidemiologic studies indicate that strains that encode Stx2a are more likely to be associated with food-borne outbreaks and severe disease, such as the HUS, than are those that make Stx1a only or Stx1a and Stx2a (16–18).

Although no one animal model recapitulates all aspects of STEC pathogenesis, the capacity of the Stxs to cause disease has been demonstrated by either infection or intoxication models in mice, rats, pigs, baboons, and greyhounds (for reviews, see references 19 and 20). For example, oral infection with certain strains of STEC in mouse models or injection of mice with either Stx1a or Stx2a results in renal injury and death (reviewed in reference 21). Monoclonal antibody (MAb) against the toxin is able to protect those animals from disease and death (22, 23), findings that further establish a role for Stx in pathogenesis. Only a few studies that examine oral intoxication by Stx in animals have been reported. In two such investigations, purified Stx (specific toxin type[s] unknown, [24]) and Stx2a (25) were found to be lethal after intragastric (i.g.) gavage of infant New Zealand White rabbits. More recently, Rasooly et al. demonstrated that i.g. administration of 50 μg, but not 0.5 μg, of Stx2a is lethal in Swiss Webster mice (26). Here, we further defined the oral toxicity of the Stxs in mice. We found that although Stx1a was not lethal via the oral route, Stx2a has an i.g. 50% lethal dose (LD₅₀) of 2.9 μg. We further showed that the renal pathologies caused by Stx2a by both the i.g. and intraperitoneal (i.p.) routes of intoxication were similar. Finally, we protected mice from oral Stx2a intoxication by passive antibody transfer of MAb 11E10 against the Stx2a toxin A subunit.

(A portion of this work was presented at the 8th International Symposium on Shiga Toxin [Verocytotoxin] Producing Escherichia coli Infections, Amsterdam, The Netherlands, 6 to 9 May 2012 [27]).)

MATERIALS AND METHODS

Bacterial strains and growth conditions.

E. coli K-12 DH5α strains transformed with pLPSH3 (28) or pJES120 (29) encode Stx (equivalent to Stx1a and called Stx1a herein for simplicity) or Stx2a, respectively. Both strains were grown in Luria-Bertani (LB) broth or LB agar supplemented with 100 μg/ml ampicillin for maintenance of the recombinant plasmid.

Purification of Stx1a and Stx2a.

Both toxins were purified by affinity chromatography with 5 ml AminoLink coupling resin (Thermo Scientific) columns.

Column preparation.

MAb to the B subunit of either Stx1a or Stx2a was covalently bound to the column resin in pH 7.2 coupling buffer according to the manufacturer's instructions. MAb 13C4 (30) was purified from hybridoma supernatant by fast protein liquid chromatography over a HiTrap protein G high-performance (HP) 5-ml column (GE Life Sciences, Pittsburgh, PA), desalted (HiTrap desalting column; GE Life Sciences) into phosphate-buffered saline (PBS), and used for purification of Stx1a (approximately 7 mg MAb/column), while MAb BC5 BB12 (31) in ascites fluid, a gift from Nancy Strockbine, was diluted into PBS and used for Stx2a purification (approximately 13 mg MAb/column).

Cell lysate preparation.

An overnight culture of the E. coli K-12 strain that contained the plasmid that encoded the stx of interest was sedimented by centrifugation (5,000 × g) and the pellet resuspended at 4 ml per g in sonication buffer (50 mM NaPO₄, 200 mM NaCl). The resuspended cell pellet was disrupted by sonication. The cell lysate was sedimented by centrifugation (20,000 × g), and the supernatant was filtered prior to toxin purification.

Toxin purification.

Cell lysates were applied to the column in accordance with the manufacturer's instructions for sample application except as noted below. All column elutions were recovered via gravity instead of centrifugation. The cell lysate flowthrough was reapplied to the column for additional toxin binding. The column contents were washed consecutively with three column volumes each: sonication buffer, high salt buffer (0.5 M NaCl, 50 mM NaPO₄) to further remove nonspecific contaminants, and sonication buffer to prevent protein denaturation resulting from the high salt concentration. Toxin was eluted (0.1 M glycine, pH 2.8) in 1-ml fractions into 200 μl neutralization buffer (1 M Tris HCl, pH 9.5). The fractions that contained toxin were dialyzed against PBS in Slide-A-Lyzer dialysis cassettes (Thermo Scientific). Toxin activity was confirmed on Vero cells as previously described (32). The endotoxin level was less than 0.025 endotoxin units (EU) per μg Stx2a (data not shown) as determined by the Limulus amebocyte lysate chromogenic endpoint assay (Hycult). When necessary, 15-ml Millipore Amicon Ultra 30K concentrators were used in accordance with manufacturer's instructions to concentrate the toxin preparation. The protein concentration in the toxin preparations was determined with a bicinchoninic acid (BCA) assay (Thermo Scientific).

Specific toxin concentration determination.

To normalize for differences in purity among the toxin preparations, we used densitometry analyses of stained gels to determine the percentage of toxin in each preparation relative to the total protein in each sample as follows. Purified toxins were separated on a NuPAGE Novex 4 to 12% bis-Tris gel(s) (Invitrogen). The gel(s) was then stained with Oriole fluorescent gel stain (Bio-Rad) and scanned with the ImageQuant LAS 4000 system (GE Healthcare). The scanned image was analyzed with ImageQuant TL software (GE Healthcare) to determine what contribution to the total protein was made by the bands that comprised the A and B subunits of the toxins. The concentration of purified toxin was then calculated by multiplying the percentage of toxin in the preparation by the total protein concentration measured in the BCA assay. The densitometry analyses of the individual A and B subunit bands suggested that the proportion of A to B subunits in each preparation was at or close to the expected ratio of 1:5 and that the purity of the preparations ranged from 82 to 95% (data not shown).

Mice.

All animal studies were approved by the Institutional Animal Care and Use Committee of the Uniformed Services University of the Health Sciences and were conducted in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals (33). Female BALB/c mice, 5 to 6 weeks old, were obtained from Charles River Laboratories (Wilmington, MA) and used for all experiments. Food and water were removed for 18 and 2 h, respectively, prior to all i.g. intoxication studies. Mice were gavaged with 0.2 ml of toxin-PBS dilutions or PBS as a control for oral intoxication or injected with 0.1 ml of a Stx-PBS dilution for all i.p. intoxication studies. Mice were weighed daily and monitored for morbidity and mortality for 2 weeks postintoxication. When the starting weights of mice were similar among groups, we graphed the mean weights of animals in each group on that day. However, when the initial mean weights of the mice in groups differed by more than 5%, we plotted the percent weight change per day. Two-way analysis of variance (ANOVA) was used to determine a significant difference in weight change among intoxication groups. The Stx2a i.g. intoxication LD₅₀ and 95% confidence interval (CI) values were determined by probit regression analysis with log transformation of the values.

Serum biochemistry and histopathology.

Mice received Stx2a diluted in PBS or PBS alone as a control through i.p. or i.g. administration. Three or 4 days postintoxication, mice were anesthetized with isoflurane (VetEquip Incorporated, Pleasanton, CA), and blood was collected via cardiac puncture. After the blood clotted, the serum was separated by centrifugation at 6,000 × g for 10 min and sent to BioReliance (Rockville, MD). Serum samples were analyzed by an automated clinical chemistry analyzer (Cobas 600 series) and reported through a laboratory information management system (LIMS). For the initial studies, a full serum panel was done. However, only the kidney-specific markers showed differences from the expected normal range (http://www.criver.com/files/pdfs/rms/balbc/rm_rm_r_balb-c_mouse_clinical_pathology_data.aspx), so all subsequent serum analyses were limited to the renal panel: blood urea nitrogen (BUN), creatinine, sodium (Na), potassium (K), chloride (Cl), and albumin. One-way ANOVA was used to determine statistical significance between the experimental and control groups.

After blood collection, the mice were euthanized and necropsied. For some studies, kidney, small intestine, cecum, large intestine, and/or liver were collected, fixed in formalin, and sent to Histoserv (Germantown, MD) to be embedded in paraffin and sectioned. Sections from all organs were stained with hematoxylin and eosin (H&E). The semiquantitative severity modifier of a morphological diagnosis, such as mild, moderate, or severe, was used to describe lesions. Additionally, serial sections from some of the kidney samples analyzed with H&E were stained with periodic acid-Schiff (PAS) stain. PAS stains polysaccharides, including basement membranes and brush boarders of certain epithelial cells. Therefore, proximal tubules, which have a thick brush border, can be distinguished with the PAS stain from distal tubules that lack a prominent brush border. Slides were read by a veterinary pathologist who was blind to the study group identifications.

Immunofluorescence.

Slides with unstained kidney or intestinal sections were deparaffinized by incubation in Histoclear (National Diagnostics, Atlanta, GA) twice for 3 min each time. The tissue was then rehydrated in a graded ethanol (ETOH) series as follows: three incubations for 3 min in 100% ETOH, followed by 3 min in 95% ETOH, 3 min in 90% ETOH, and finally 3 min in 70% ETOH. The slides were rinsed in deionized water and heated in 1× citrate, pH 6, antigen retrieval buffer (Ag Plus buffer; Novagen) at 95°C for 10 min. The slides were then rinsed in deionized water and blocked overnight at 4°C in 1% goat serum diluted in antibody (Ab) diluent reagent solution (Invitrogen). Slides were incubated with primary Ab (polyclonal rabbit anti-Stx2a [34] diluted 1:500 in Ab diluent solution) for 1 h at room temperature. Slides were washed 3 times with PBS before the secondary Ab, goat anti-rabbit-Alexa-fluor 488 (Invitrogen) diluted 1:500 in Ab diluent, was applied for 1 h at room temperature. Finally, slides were washed 3 times with PBS and the coverslip was mounted with Vectashield (Vector Laboratories, Inc., Burlingame, CA), which contained the counterstain, 4′,6-diamidino-2-phenylindole (DAPI). Slides were observed on an Olympus BX60 microscope with a BX-FLA fluorescence attachment.

Passive Ab transfer.

MAb 11E10 (BEI Resources/Hycult), directed against the Stx2a A subunit (22, 35), was used for all passive Ab transfer studies. TFTB1 (BEI Resources), an irrelevant Ab which recognizes the ricin B subunit (36), was included as an IgG2a isotype control. Additionally, each experiment included a group that received Stx2a only as a positive control for toxin activity. A dose of 7.5 μg Stx2a was used for all Ab protection and rescue experiments because that amount of toxin is invariably lethal in the oral intoxication model. Both MAbs and Stx2a were diluted in PBS for all experiments.

Ab protection.

For the preliminary study, mice received 2 μg of 11E10 or TFTβ1 intravenously (i.v.) by tail vein 24 h prior to i.g. intoxication with 7.5 μg Stx2a. In all subsequent experiments, mice received 2, 4, or 40 μg of Ab i.v. 1 h before intoxication.

Statistical analyses of antibody protection data.

The percent weight change values were first transformed into log ratios {log [(Y/100) + 1], where Y is weight change, to make the data more normally distributed with more homogenous variance. Next, two-way repeated-measures ANOVA and Tukey's post hoc tests were used to determine statistical significance of weight loss among Ab dose groups. Since repeated-measures ANOVA is an appropriate tool only for evaluation of complete data sets, analyses started on day 2. Although there was no significant group effect (P = 0.06), there was a significant interaction effect (P = 0.0002), a finding that indicates that the difference among groups varies over time. Tukey's post hoc tests were used to determine statistical significance on individual days.

Ab rescue.

MAb 11E10 or TFTβ1 (4 μg) was delivered i.v. at 6, 24, 48, or 72 h after Stx2a i.g. intoxication. An Ab protection group, which received 4 μg 1 h prior to intoxication, was included as a positive Ab control.

RESULTS

Stx2a but not Stx1a is lethal for mice after oral intoxication.

We determined the LD₅₀ of Stx2a in mice by the oral intoxication route. Groups of mice were gavaged with Stx2a at concentrations from 0.25 μg through 15 μg toxin. Significant weight loss and death were observed at Stx2a concentrations of 2 μg or greater (Fig. 1A and B). The LD₅₀ for Stx2a after oral intoxication was calculated to be 2.9 μg (CI, 1.4 to 5.0 μg), 1,000× greater than the i.p. Stx2a LD₅₀ of approximately 2 ng (see reference 37), which was confirmed with our toxin lot (not shown). The mean time to death (MTD) was 4.9 days. In contrast, no morbidity or mortality was observed after i.g. intoxication of up to 157 μg Stx1a (data not shown).

FIG 1.

Morbidity and mortality after i.g. intoxication by Stx2a in mice. (A) Mean weights of mice over the course of the experiment for each Stx2a concentration tested in the oral-intoxication model. The sample size was 5 for all doses except 1 and 10 μg, for which the sample size was 3. Error bars indicate standard errors of the means. There was a significant dose-over-time interaction, such that mice that received 2 μg or more of Stx2a experienced significant weight loss compared to animals that received 0.25 or 1 μg of toxin (P < 0.05). The weight gain in the no-toxin group (PBS only) was statistically similar to that of the 0.25-μg Stx2a group (P < 0.05), so for simplicity, only the 0.25-μg-Stx2a weight data were plotted. (B) Mouse survival percentage over time at each Stx2a concentration. The PBS-only control group had 100% survival (data not shown).

Kidney damage and electrolyte imbalances occur in mice intoxicated i.p. or i.g. with Stx2a.

After we determined the oral Stx2a LD₅₀, we compared the serum biochemistry values after i.p. and i.g. intoxication. Mice were challenged with 6 times the i.p. or i.g. Stx2a LD₅₀ (13.2 ng or 17 μg, respectively) or PBS as a control. Serum biochemistry markers from all groups were evaluated 3 days postintoxication. The serum biochemistry markers from the i.p. and i.g. PBS control groups were within the normal range and statistically indistinguishable (only i.g. data are shown in Table 1, first row). Regardless of the route of intoxication and except with albumin, the serum biochemical values related to kidney function from Stx2a-intoxicated mice were significantly different than those from the PBS control group values (Table 1). Specifically, the BUN and creatinine levels rose significantly after intoxication, and there was an imbalance of the electrolytes Na, Cl, and K.

TABLE 1.

Mean renal panel serum biochemistry values

Agent and route	No. of subjects	Dose	Time of blood collectiona	Level (range)f
				BUN (mg/dl)	Creatinine (mg/dl)	Na (mmol/liter)	K (mmol/liter)	Cl (mmol/liter)	Albumin (g/dl)
PBS i.g.b	8	0.2 ml	Day 3 p.i.	33.9 (25–40)	0.24 (0.2–0.28)	135.5 (132–137)	19.9 (13.7–24.5)	105.3 (102–108)	3.3 (3.2–3.5)
Stx2a i.p.c	4–9	13.2 ng	Day 3 p.i.	193.9 (81–246)	0.95 (0.67–1.2)	119.5 (99–128)	30.2 (23.2–45)	84.8 (67.3–91.1)	4.1 (3.9–4.2)
Stxa2 i.g.c	8–9	17 μg	Day 3 p.i.	116.6 (82–245)	0.61 (0.44–1.1)	105.9 (76–121)	37.3 (25.6–67.3)	77.2 (68.9–88.8)	3.7 (1.9–4.3)
Stx2a i.g.	2	2 μg	Wt loss plateau	114.5	0.50	119.0	37.3	87.1	3.7
Stx2a i.g.	4	2 μg	After 1 day of recovery	35.8 (34–38)	0.20 (0.14–0.24)	135.8 (134–139)	25.3 (23.3–27.2)	106.6 (106–109)	3.3 (3.2–3.4)
Stx2a i.g.	2	7.5 μg	Day 4 p.i.	275	1.62	NV	NV	NV	NV
+TFTB1d	5	7.5 μg	Day 4 p.i.	213.2e (150–269)	1.23e (0.99–1.41)	NV	NV	NV	NV
+11E1d	4	7.5 μg	Day 4 p.i.	30.5e (25–40)	0.24e (0.2–0.28)	NV	NV	NV	NV

^aBlood collection times were dictated by parameters for each study.

^bSerum biochemistry values from i.p. and i.g. PBS control mice were not statistically different (P < 0.05).

^cThe serum chemistry values were significantly different (P < 0.01) from PBS control values except for the K in the Stx2a i.p. group and the albumin values in both groups.

^dFour micrograms of Ab.

^eP < 0.0001, TFTB1 versus 11E10.

^fNV, no value returned.

Renal tubular damage is seen in kidney sections of mice orally intoxicated with Stx2a.

We next examined intestinal, liver, and renal tissues for damage after i.g. Stx2a intoxication. Four days after i.g. intoxication with 7.5 μg Stx2a (∼2.5 times the LD₅₀) mice were euthanatized and their intestinal tracts and kidneys removed. For comparison, we also collected the kidneys from mice intoxicated i.p. with 13.2 ng or i.g. with 17 μg Stx2a (6 times the LD₅₀) on day 3 (the intestinal tract does not exhibit damage after i.p. intoxication [37]). The earlier time point (day 3 as opposed to day 4) of kidney collection was necessitated by the earlier MTD and the increased Stx2a i.g. dose for i.p. intoxicated mice. No lesions were noted in the small intestines, ceca, or large intestines of mice orally intoxicated with Stx2a (see Fig. S1A to C in the supplemental material). We next reviewed kidney sections stained with H&E from mice given PBS or Stx2a i.g. No lesions were observed in renal sections from PBS control mice (Fig. 2A). In contrast, we found diffuse tubular dilation and basophilia in the renal cortexes of mice intoxicated i.g. with 7.5 μg Stx2a (Fig. 2B) or 6 times the LD₅₀ of Stx2a regardless of the route of toxin administration (not shown). We then used the PAS stain to examine kidney sections. As expected, no renal lesions were detected in the kidneys of either i.p. (not shown) or i.g. intoxicated PBS control mice (Fig. 2C). However, we found minimal-to-moderate multifocal acute tubular necrosis (ATN) of distal tubules, characterized by tubules lined with degenerating, necrotic, or sloughed epithelial cells, independent of dose or route of intoxication (result after administration of 2.5 times the LD₅₀ is shown in Fig. 2D).

FIG 2.

H&E (A and B)- and PAS (C to F)-stained kidney sections from PBS-treated or Stx2a-intoxicated mice. At a ×40 magnification, no lesions were noted in the i.g. PBS control section (A), while tubular dilation (▲) was evident in the cortexes of mice i.g. intoxicated with 7.5 μg Stx2a (B). The PAS stain was used to analyze kidney sections at a ×100 magnification from PBS control (C) and Stx2a-intoxicated (D to F) mice. (C) No histopathology was observed in kidney sections from PBS control mice; as expected, the proximal tubules (+) contained a prominent pink-staining brush border, and the distal tubules (∧) did not. Kidneys from Stx2a-intoxicated mice (2.5 times the LD₅₀) (D) or mice intoxicated with a sub-LD₅₀ dose when necropsied at a weight loss plateau (E) exhibited minimal-to-moderate, multifocal ATN (*) of distal tubules; tubules were lined with fragmented cells or contained cellular debris within the lumina. Adjacent proximal tubules were not affected. Mice intoxicated with 2 μg and allowed to recover for 1 day after weight loss peaked (F) had limited tubular necrosis and regeneration (■) characterized by epithelial cells with basophilic cytoplasm, large vesicular nuclei, and increased mitotic figures (→) present in multiple distal tubules. PBS i.g., n = 13; Stx2a i.g. at 2.5× the LD₅₀, n = 10; Stx2a i.g. at 2 μg, weight loss peak, n = 4; Stx2a i.g. at 2 μg with 1 day of recovery, n = 6; PBS i.p., n = 13 (data not shown); Stx2a i.g. at 5× LD₅₀, n = 10 (data not shown).

Stx2a is present in the kidneys of intoxicated mice.

We next asked if we could detect Stx2a in the kidneys of mice after i.p. or i.g. intoxication. Only a minimal green fluorescent background was observed in kidney sections from mice given PBS i.p. (Fig. 3A) or orally (not shown). However, we were able to confirm the findings of Rutjes et al. that Stx2a can be found in kidney sections after parenteral (in our case, i.p.; in their studies, i.v.) intoxication (38) (Fig. 3B). Moreover, we also detected Stx2a in kidney sections from i.g. intoxicated mice (17 μg) (Fig. 3C). The Stx2a-positive cells appeared to be tubule epithelial cells for both i.p. and i.g. intoxicated mice, as the Stx2a staining pattern in the tubules coincided with the histopathological location of renal damage in the intoxicated mice (Fig. 2D).

FIG 3.

Staining to detect Stx2a in kidney sections from PBS-treated or intoxicated mice. Immunofluorescence of kidney sections treated with rabbit polyclonal anti-Stx2a combined with goat anti-rabbit conjugated to Alexa 488. DAPI stained the cell nuclei. (A) i.g. PBS control sections were negative for Stx2a. Stx2a-positive cells adjacent to tubule lumens (bright-green fluorescence) were detected in kidney sections from mice i.p. intoxicated with 13.2 ng (B) and i.g. intoxicated with 17 μg (C) of Stx2a. Arrows point to the lumen of a tubule. Magnification, ×400.

Morbidity in Stx2a-intoxicated mice correlates with kidney function.

Because we observed significant weight loss followed by rapid weight gain in the majority of mice intoxicated i.g. with 2 μg of Stx2a, we speculated that mice given 2 μg of Stx2a i.g. might exhibit kidney damage even though they typically recovered from that toxin dose (Fig. 1A and B). In a subsequent study, we found that mice intoxicated with 2 μg of Stx2a, a sub-LD₅₀ dose, and sacrificed when weight loss began to plateau (day 8 or 9 in this experiment) exhibited serum biochemistry profiles and kidney histopathology similar to those of mice intoxicated with 6 times the LD₅₀ (Table 1, 4th row; Fig. 2E). Stx2a was also detected by immunofluorescence in the kidneys of those same mice (not shown). In mice given 2 μg of Stx2a and allowed to recover 1 day after the weight loss plateau, serum biochemistry values returned to control levels and kidney histology improved (the ATN was less severe and tubules contained regenerating epithelial cells) (Table 1, 5th row; Fig. 2F). Three days after the rebound in weight, no kidney lesions were noted (data not shown).

Anti-Stx2a MAb fully protects mice when given before intoxication.

We next tested whether i.g. intoxicated mice could be protected from morbidity and/or mortality with a MAb to Stx2a. In a pilot study, mice received 2 μg of 11E10 (anti-Stx2a) or TFTB1 (isotype-matched control) i.v. 24 h prior to i.g. intoxication with 7.5 μg Stx2a. A positive-control group received only Stx2a i.g. (In this study and all subsequent experiments, MAbs and Stx2a were diluted in PBS.) In this preliminary study, complete mortality was observed in the Stx2a-only and Stx2a-plus-TFTB1 groups, while MAb 11E10 prevented mortality though not morbidity (Fig. 4A, 1st row; Fig. 4B). Protected mice experienced a cumulative weight loss of 1.9 g or 11% that peaked on day 7 postintoxication compared to their starting weight before they recovered and exhibited positive weight gain by day 14 (data not shown). For our next study, in the hopes of limiting morbidity in the toxin-treated mice, we altered the time of administration of 11E10 from −24 h to −1 h so that higher levels of MAb would be present during the intoxication window. We found that although the mice were protected from lethality, as expected, they still experienced weight loss before recovery (Fig. 4A, 2nd row, and B). Therefore, in an attempt to reduce morbidity in the subsequent protection experiment, the amount of MAb administered was doubled. Mice that received 4 μg 11E10 were completely protected (Fig. 4A, 3rd row), and in addition, their weight loss was reduced, although not to a statistically significant level (P = 0.055) (Fig. 4B). In a final attempt to further reduce or prevent Stx2a-mediated morbidity, 10-fold-more 11E10 was administered in the next study. Although 11E10 again protected (Fig. 4A, 4th row), the level of weight loss in the intoxicated, treated animals was similar to that in the previous protection experiment in which mice received 4 μg MAb; i.e., there was no statistical difference in weight loss in mice treated with 4 or 40 μg MAb (P < 0.064) (Fig. 4B).

FIG 4.

MAb 11E10 prevented mortality and limited morbidity due to Stx2a i.g. intoxication. (A) Mortality results from four independent Ab studies. Two, 4, or 40 μg 11E10 protected mice from Stx2a i.g. intoxication, while TFTB1, the irrelevant MAb, had no therapeutic effect. Ab Rx, antibody administration; ND, not done. (B) Percent weight changes from day 0 to day 14 of antibody-treated, Stx2a-intoxicated, and PBS control groups. Each group received Ab 1 h prior to intoxication. The PBS control group experienced positive weight gain over the course of the experiment, with no effect from 40 μg 11E10. Error bars indicate standard deviations. There was a significant interaction effect (P = 0.0002), which indicates that the percentages of weight change vary over time between Ab treatment groups. The difference in percent weight change was significant between the groups given 2 or 4 μg 11E10 on days 3, 7, and 9 and between the 2-μg and 40-μg groups on days 5 to 8. There was no difference between the 4- and 40-μg groups.

We next compared mouse weight changes after administration of 4 μg MAb in the i.p. and i.g. intoxication models to evaluate whether MAb protection from morbidity might be greater in the i.p. than the i.g. model. As was done previously, an Stx2a-only group served as a positive control for mortality, and MAb TFTB1 was included as an isotype-matched control for 11E10. One hour after injection of treatment groups with 4 μg MAb, mice received 2.5 times the i.p. or i.g. LD₅₀, 5.7 ng or 7.5 μg, respectively. All mice in the Stx2a-only positive-control groups succumbed to intoxication, but the MTD was longer in i.g. intoxicated animals (5 days) than that of mice given toxin i.p. (3.4 days). As expected, TFTB1 did not alter morbidity or survival in either the i.p. or i.g. intoxicated groups (data not shown). In contrast, all mice given11E10 survived and had positive weight gain by day 14 postintoxication. However, weight loss began earlier in the 11E10-treated animals intoxicated i.p., and those mice lost significantly more weight prior to recovery than MAb-treated mice given toxin i.g. (P < 0.05) (Fig. 5).

FIG 5.

Average weight over time in Stx2a-intoxicated mice given 11E10 (open symbols) or no treatment (filled symbols) 1 h before toxin administration. Mice were intoxicated with 2.5 times the Stx2a LD₅₀ i.p. (circles) or i.g. (squares). Although MAb 11E10 was protective in both intoxication models, i.p. intoxicated mice exhibited significantly more morbidity than i.g. intoxicated mice (P < 0.05). Additionally, there was a significant difference in weight between the Stx2a i.p. and i.g. 11E10-protected groups on days 5, 6, 8, and 11.

Levels of renal serum markers and the degrees of histopathology in the kidneys of Stx2a-intoxicated and MAb-protected mice are related.

We analyzed sera from i.g. intoxicated mice treated with nothing or MAb TFTB1 or 11E10 (MAbs were given 1 h prior to intoxication) to determine if there were alterations in the values for the renal damage markers BUN and creatinine. Compared to animals given 11E10 and then i.g. Stx2a, mice intoxicated with Stx2a only or administered TFTB1 and then intoxicated with Stx2a exhibited a significant increase in BUN and creatinine levels (P < 0.0001) (Table 1, 6th and 8th rows). However, the renal biochemistries from mice protected with 11E10 were indistinguishable from those of the PBS controls from previous experiments (P > 0.05) (Table 1, 1st row).

Finally, the effect of MAb treatment on kidney histopathology was evaluated. Moderate ATN was detected, as expected, in kidneys from the Stx2a-only control group (similar to the pathology previously described for Fig. 2D) and from mice treated with TFTB1 (Fig. 6A). The pathologist did not note any difference in kidney lesions between Stx2a-only- and TFTB1-treated, Stx2a-intoxicated mice. ATN was also observed in mice given 11E10 and then i.g. Stx2a, although the lesions were minimal rather than moderate, and there was increased evidence of regeneration of distal tubules (Fig. 6B).

FIG 6.

PAS-stained kidney sections from the MAb protection/Stx2a i.g. intoxication study. (A) Moderate ATN (⧫) of distal tubules was observed in the kidneys of mice in Stx2a-plus-TFTB1 groups. (B) Minimal ATN (•) and increased regeneration of distal tubules were seen in Stx2a- and 11E10-treated mice.

MAb can rescue Stx2a-intoxicated mice.

We tested the capacity of 11E10 to rescue mice after oral intoxication with Stx2a. MAb 11E10 or TFTB1 was administered i.v. at 6, 24, 48, or 72 h after i.g. intoxication with approximately 2.5 times the LD₅₀ (7.5 μg) of Stx2a. None of the mice treated with TFTB1 or in the Stx2a-only group survived toxin challenge. Additionally, all TFTB1-treated and Stx2a-only groups had similar MTDs, a finding that indicates that the addition of fluids alone had no therapeutic effect (since the TFTB1 was given in PBS) (data not shown). Mice that received 11E10 1 h prior (a control) or 6 h after intoxication were completely protected from Stx2a. In contrast, only one mouse was rescued when 11E10 was given at 24 h postintoxication, and no mice were rescued if the MAb was administered at 48 or 72 h after toxin administration (Table 2). Both the −1-h and +6-h treatment groups also exhibited similar patterns of weight loss throughout the experiment, and both groups displayed positive weight gain by day 14 (data not shown).

TABLE 2.

Rescue by 11E10 of Stx2a i.g. intoxicated mice

Time of 11E10 treatment relative to i.g. Stx2a (h)	No. of dead mice/total	MTD (days)
−1	0/8
+6	0/10
+24	9/10	4.4
+48	10/10	4.4
+72	8/8	4.3
NAa (Stx2a only)	8/8	4.4

^aNA, not applicable.

DISCUSSION

That Stx (specific toxin type[s] unknown) alone can recapitulate some of the symptoms of STEC disease was first reported by Pai et al. (24). These investigators found that infant New Zealand White rabbits orally intoxicated with Stx develop diarrhea and succumb to intoxication in a manner analogous to that of rabbits challenged with E. coli O157:H7 (24). Similarly, Ritchie et al. showed that rabbits orally intoxicated with Stx2a develop inflammation comparable to that of rabbits infected with an Stx2a-producing STEC strain (25). Additionally, Rasooly et al. recently reported that oral intoxication of mice with high doses of Stx2a was lethal (26). We extended such observations with the determination that the LD₅₀ for Stx2a in BALB/c mice by the oral route is 2.9 μg and that both the i.p. and i.g. Stx2a intoxication routes lead to similar systemic pathologies in those animals. Additionally, our data illustrated that renal damage similar to that of mice inoculated with the LD₅₀ occurred even in mice intoxicated with a sub-LD₅₀ Stx2a dose and that survival in those animals correlated with a return to normal serum chemistry values and tubule regeneration. Finally, we demonstrated protection and rescue of mice from Stx2a oral intoxication through passive transfer of MAb 11E10.

We found that Stx1a was not lethal by the intragastric route, in contrast to the oral lethality of Stx2a. We speculate that if, as with Stx2a, the i.g. LD₅₀ for Stx1a is 1,000 times greater than the i.p. LD₅₀ (as it was for Stx2a), more than 500 μg Stx1a per mouse would be necessary to result in lethality. However, it is not practical to purify such large quantities of toxin to test the hypothesis that Stx1 is potentially lethal by the oral route. Our finding that Stx1a is not lethal by the oral route as far as we can test may not be surprising, since STEC strains that produce only Stx1a are not lethal in mouse models of oral infection (39).

We did not observe lesions in any section of the intestinal tracts from Stx2a i.g. intoxicated mice, a finding that suggests that the toxin reached systemic circulation without injury to the intestinal epithelial cell lining. We also examined the liver for damage because we thought that that organ had the potential to absorb a large concentration of Stx2a as the toxin traveled through the hepatic portal system. However, we did not observe lesions in the livers of orally intoxicated mice, and we did not detect alterations in the hepatic serum biochemistry values (data not shown). Therefore, we concluded that the liver is not a target site for Stx2a in mice, a contention supported by the observation that Gb3 concentrations are low in that organ (40).

We found that once sufficient Stx2a entered systemic circulation, regardless of the route of intoxication (i.p. or i.g.), similar elevations in renal serum biochemistry values occurred. The elevated BUN and creatinine values suggest acute renal failure, which is characterized by a reduced glomerular filtration rate (GFR). Although the pathophysiologic mechanism of the GFR is not fully understood, it likely involves reduction in blood flow to or within the kidney. The apparently reduced GFR is not related to problems with the glomerulus in these mice. In addition, the significant reduction in sodium and chloride levels due to Stx2a exposure indicate renal tubular malfunction (41, 42). We do not believe that the changes in serum biochemistry values after Stx2a intoxication are a result of hemoconcentration due to dehydration because the experimental serum albumin levels were equivalent to albumin levels from control mice.

Our immunofluorescence data suggest that once Stx2a reached the kidney, it targeted tubular epithelial cells because the toxin staining was in the cells adjacent to the renal tubule lumen. The toxin molecule likely killed those cells, events that would lead to ATN. ATN causes loss of renal function, which, in turn, can result in mortality. ATN is the same pathology that occurs in STEC mouse infection models. Overall, these results demonstrate that oral intoxication with Stx2a alone recapitulates disease noted after infection of mice with STEC in various models (21).

The specificity of the PAS stain revealed approximately equivalent proportions of distal and proximal tubules in both control and experimental mice (data not shown). The PAS stain further indicated that the Stx2a-mediated lesions were in distal rather than proximal tubules (as we had previously believed [37]), because of the absence of a brush boarder in the affected tubules. The hypothesis that distal rather than proximal tubules are more affected by Stx2a is supported by a study that showed that distal tubules have a higher concentration of Gb3 receptors on the cell surface than do proximal tubules (43). We were surprised to find that the renal pathology and serum biochemistry values due to a sub-LD₅₀ oral Stx2a dose were similar to those found after i.g. delivery of 6 times the LD₅₀ of toxin if the mice were necropsied when weight loss reached its nadir. In contrast, once the sub-LD₅₀-intoxicated animals began to gain weight, they had serum biochemistry values similar to those of control mice and exhibited less severe kidney histopathology. We believe that although the types of histopathology observed in the sub-LD₅₀ and 6-fold-LD₅₀ Stx2a groups were similar when assessed at the point of lowest weight for the sub-LD₅₀ group, the magnitude of damage was greater in the higher-dose group. We further speculate that those mice that received a sub-LD₅₀ dose survived if they were able to regenerate tubular epithelial cells (observed in Fig. 2F) and restore a minimum threshold of kidney function. This study also demonstrated the importance of timing of sample collection after intoxication; we note that damage can occur yet go undetected because of subsequent repair if samples are taken too long after toxin insult.

We protected all mice from 2.5 times the LD₅₀ of Stx2a with a single passive transfer of 4 μg 11E10. Although there was a reduction in weight loss at the 4-μg compared to the 2-μg 11E10 dose, a 10-fold increase to 40 μg of the MAb did not result in a further decrease in morbidity. Thus, there appears to be a threshold for levels of antibody that can decrease morbidity in intoxicated animals. We also found that mice treated with 11E10 and then given Stx2a had normal kidney function 4 days postintoxication, as indicated by serum BUN and creatinine levels. Although minimal renal ATN was observed in the MAb-treated Stx2a-intoxicated mice, there were increased instances of tubule regeneration compared to that in animals that received Stx2a only. Our study suggests that 11E10 pretreatment reduced kidney damage in Stx2a-intoxicated mice. Finally, since we observed no protective effect by irrelevant MAb TFTB1, an IgG2a isotype control, we conclude that fluid alone was not protective.

We also observed that in mice treated with 11E10, greater morbidity occurred after i.p. than i.g. Stx2a intoxication. The i.p.-intoxicated animals exhibited a shorter time to death of approximately 1.5 days than mice given toxin i.g. The reduced MTD in i.p. intoxicated mice suggests that Stx2a reached the target site of the kidneys earlier when it was delivered to the peritoneum than when it was delivered to the gastrointestinal tract. Although ATN was probably present at increased levels in i.p. compared to i.g. intoxicated mice, kidney function was not lost and the mice were able to recover.

We demonstrated rescue of mice with 4 μg of 11E10 at 6 but not 24 h after oral intoxication with Stx2a. A recent study in which mice were orally intoxicated with botulinum neurotoxin serotype A (BotA) and then treated by passive Ab transfer demonstrated that differences of just 1 to 2 h in antibody administration time points resulted in the loss of the capacity to rescue intoxicated animals (44). We speculate, therefore, that as with BotA, once Stx2a is sequestered in target tissues (the kidneys), passive Ab transfer is not able to rescue intoxicated animals. However, the rescue time frame may be increased if a larger MAb dose is provided or if additional doses are delivered. In support of the latter hypothesis, another group was able to rescue baboons 24 h after intravenous intoxication with Stx2a by providing a therapeutic dose of TVP, an acetylated tetravalent peptide, daily until day 4 (45). We believe that passive Ab transfer is a viable therapeutic option for STEC infections. Furthermore, for someone infected with STEC, the rescue time frame is likely extended compared to that of the oral intoxication model, as Stx would be delivered at a continual low dose rather than in a single large bolus. Previous research in our lab demonstrated that low levels of active Stx2a could be found in the feces of mice with an intact commensal flora up to 72 h postinfection with an Stx2a-producing O157:H7 strain (46).

Finally, we found that an oral Stx2a dose of 2 μg (∼0.1 mg/kg of body weight) was sufficient to cause weight loss and renal injury but not death in mice. If mice and humans are equivalently susceptible to Stx, then we calculate that 1.8 mg Stx2a would be sufficient to cause some disease in an 18-kg child. However, we believe that humans are more susceptible to Stx than are mice because, unlike mice, people have Gb3 in their glomeruli in addition to in their tubules (37, 47, 48). Therefore, we believe that the oral toxicity of Stx2a is potentially relevant as a public health issue. There is a possibility that preformed Stx occurs in contaminated food, which might contribute to morbidity (49–51) and possibly enhance colonization by the organism, as Stx2 has been shown to increase the adherence of O157:H7 to tissue culture cells and in animal models (52, 53).