Ricin Crosses Polarized Human Intestinal Cells and Intestines of Ricin-Gavaged Mice without Evident Damage and Then Disseminates to Mouse Kidneys

Ricin is a potent toxin found in the beans of Ricinus communis and is often lethal for animals and humans when aerosolized or injected and causes significant morbidity and occasional death when ingested. Ricin has been proposed as a bioweapon because of its lethal properties, environmental stability, and accessibility. In oral intoxication, the process by which the toxin transits across intestinal mucosa is not completely understood. To address this question, we assessed the impact of ricin on the gastrointestinal tract and organs of mice after dissemination of toxin from the gut. We first showed that ricin adhered in a specific pattern to human small bowel intestinal sections, the site within the mouse gut in which a variable degree of damage has been reported by others. We then monitored the movement of ricin across polarized human HCT-8 intestinal monolayers grown in transwell inserts and in HCT-8 cell organoids. We observed that, in both systems, ricin trafficked through the cells without apparent damage until 24 hours post intoxication. We delivered a lethal dose of purified fluorescently-labeled ricin to mice by oral gavage and followed transit of the toxin from the gastrointestinal tracts to the internal organs by in vivo imaging of whole animals over time and ex vivo imaging of organs at various time points. In addition, we harvested organs from unlabeled ricin-gavaged mice and assessed them for the presence of ricin and for histological damage. Finally, we compared serum chemistry values from buffer-treated versus ricin-intoxicated animals. We conclude that ricin transverses human intestinal cells and mouse intestinal cells in situ prior to any indication of enterocyte damage and that ricin rapidly reaches the kidneys of intoxicated mice. We also propose that mice intoxicated orally with ricin likely die from distributive shock.


Introduction
The potent plant toxin ricin from the bean of the castor plant Ricinus communis is a 64 kDa bipartite protein comprised of disulfide bond-linked A and B subunits [1]. The enzymatic action of the A subunit is termination of protein synthesis by inactivation of ribosomes [2]. The B subunit binds to terminal galactose residues on glycolipids and glycoproteins, moieties so ubiquitous on cells that potential receptors for ricin may be found on every known cell type [1]. The fates of ricin following receptor-mediated endocytosis include transport back out of the cell, degradation following endosome-lysosome fusion, or retrograde transport to the Golgi apparatus. Only 5% of all internalized ricin reaches the Golgi apparatus [3], while the remainder follows the other two pathways.
The ubiquitous nature of the R. communis plant as a commercial source of castor oil, its cultivation worldwide, and the ease with which ricin is extracted from castor beans support the concern that homemade ricin weapons could readily be synthesized [4]. These attributes, coupled with the lethality of ricin, prompted the Centers for Disease Control and Prevention (CDC) to classify ricin as a Category B select agent. The amount of ricin required for toxicity by parenteral or inhalational routes is about 1,000-fold less than that required for oral intoxication [1]. Nevertheless, ingestion of castor beans causes significant morbidity and occasional mortality in humans [5]. Indeed, the lethal dose of ricin in humans following ingestion is estimated to range from 1 to 20 mg/kg [1]. Such variability in toxicity is likely dependent on several factors that include the type and germination state of the castor bean, when the bean was harvested, and patient factors such as weight and intestinal contents. In contrast to castor beans, purified ricin is colorless, odorless, and tasteless. These features, combined with the broad distribution of R. communis and the potential ease of generation of a large supply of ricin, are concerns that led us to investigate the consequences of oral ingestion of ricin. Our objective was to characterize the steps of intoxication following oral exposure to ricin.
In mice, reports of the 50% lethal dose (LD 50 ) of ricin after ingestion have varied from as low as 100µg/kg to about 10 mg/kg [6,7]. Smallshaw et al reported damage to the small intestines of mice only following exposure to excessive doses of toxin, i.e., approximately 10 times the LD 50 . Others have reported the need for large doses of ricin (≥ 2.5 mg/kg) to observe pathology in the duodenum of mice [7]. Together, these results suggest that toxin absorbed through the GI tract can result in death and that toxin escapes the GI tract of mice by a mechanism that does not damage the epithelium. Here we tested our theory that ricin can cross the intestinal epithelium without disrupting the single-cell barrier but with subsequent lethal effects. We found that ricin specifically bound human small intestinal sections on overlay, and that high doses of ricin transited human intestinal cells in transwell cultures and in a novel three-dimensional tissue culture model (organoid) without apparent damage to the cells early after intoxication. Furthermore, after orogastric administration of lethal doses of ricin to mice, we observed that dissemination of the toxin, as assessed by in vivo imaging of fluorescently labeled toxin, was apparent before histological changes to the small intestine were seen. We also showed that the kidney was the first internal organ targeted by ricin, and we obtained serum chemistry data from lethally-intoxicated mice that support a hypothesis that ricin-intoxicated mice die from distributive shock.

Ricin Purification
Ricin toxin was purified from whole castor beans (D. Landreth Seed Company, New Freedom PA). Shelled seed pulp was extracted with phosphate buffer (20 mM Na 2 HPO 4 , pH 8.0). All column chromatography steps were done on an AKTA FPLC instrument (GE Healthcare, Piscataway, NJ). The watersoluble portion of the extract was passed over sulfopropyl sepharose cation exchange resin (GE Healthcare) and bound ricin was eluted with 0.25 M NaCl. The eluate was brought to a pH of 7.4 by addition of NaOH and then was passed over a Q sepharose anion exchange resin column (GE Healthcare). The unbound fraction from the column contained ricin toxin with an estimated purity of > 95%. A Sephadex desalting column (GE Healthcare) was used to exchange the purified toxin into 1X PBS. The solution was concentrated in an Amicon Ultra centrifugal filter device (Millipore, Billerica, MA) with a 30 kDa nominal molecular weight limit. The final protein concentration was 20 mg/mL as determined by a bicinchoninic acid assay (BCA; Thermo, Fisher, Pittsburgh, PA). The specific activities of the toxin preparations used in this study ranged between 1x 10 7 to 4x10 7 50% cytotoxic doses (CD 50 )/ mg (see Vero cell cytotoxicity assay description below). Each toxin preparation was sterilized by passage through a 0.22 µm filter membrane (Millipore). Procedures for containment of ricin during purification and requirements for personal protective equipment, as approved by the U.S. Division of Select Agents and Toxins, were strictly followed by the researcher who prepared the toxin (AFF).

Vero Cell Cytotoxicity Assay
Vero cell line CCL-81 was purchased from the American Type Culture Collection (ATCC, Manassas, VA). Ricin toxicity was assessed on Vero cells in culture as previously described [8]. Briefly, cells were suspended in Eagle's minimal essential medium (EMEM) (Lonza, Walkersville, MD) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Life Technologies, Carlsbad, CA), penicillin (10 U/mL), gentamicin (100 µg/mL), and streptomycin (10 µg/mL), seeded onto 96well plates (Corning Inc, Corning, NY) at a density of 1x10 5 cells/mL, and grown for 24 h. Ricin was serially diluted 10-fold, and 100 µl of each dilution was overlaid on cells. Cells with ricin or medium alone were incubated for 48 h at 37° C, in a 5% CO 2 environment. Cells were then fixed in 10% formalin and stained with 0.13% crystal violet. The optical density at 630 nm was measured in each well of the stained plates with a spectrophotometric plate reader (BioTek EL800, BioTek U.S., Winooski, VT). The CD 50 of purified ricin for Vero cells was calculated by taking the inverse of the dilution at which 50% of the cells were killed by ricin.
Vero cell cytotoxicity was used to measure active ricin in fecal pellets of mice gavaged with varying doses of ricin. At 6, 12, 24, 48, 72, and 96 hours after gavage of mice, fecal samples were obtained by placing individual mice in empty cages and collecting pellets from the cages. Fecal pellets were resuspended in PBS (1: 9 w/v), then diluted 1:5 in 1X PBS, and applied to Vero cells in 100 µl aliquots. Homogenized stools from unintoxicated mice were used to determine background levels of Vero cell cytotoxicity.

Labeled Ricin and Free Label Control
Ricin was covalently labeled with the following fluorescent probes: Alexa Fluor ® 488 for application to polarized epithelial cells in transwell cultures; Alexa Fluor ® 633 for 8 and 16 hour in vivo imaging studies; or Alexa Fluor ® 700 for all other imaging studies and for treatment of organoids with labeled ricin. Ricin was labeled according to instructions provided by the manufacturer (Life Technologies, Grand Island, NY). Labeled toxin was separated from free fluorophore by size exclusion chromatography. The protein concentration of the labeled product was determined by the BCA assay. The in vitro specific activity (CD 50 /mg protein) of each batch of fluorescently-labeled ricin was assessed on Vero cells and found to be comparable to that of unlabeled toxin. Free label was eluted from the column and used as a control for in vivo imaging studies to account for the possibility that label might become dissociated from ricin and fluoresce in organs not affected by labeled ricin.

Polarized Epithelial Cell Growth and Intoxication
Twenty-four well transwell plates (Transwell ® -24) with permeable support membranes in each well were purchased from Corning Inc. (Corning, NY). Each permeable membrane was coated with type 1 collagen from rat tails (Sigma-Aldrich, St. Louis, MO) to promote cell polarization. The human embryonic small intestinal cell line Int-407 [9] and the human colonic epithelial cell line (HCT-8) [10] were purchased from the ATCC (Note that the Int-407 cell line supplied by ATCC is contaminated with HeLa cells per the catalog description). Cells from each of these lines were seeded onto the transwell membranes at a cell density of 1x10 6 cells/ml in RPMI-1640 medium (ATCC) supplemented with 10% heat-inactivated FBS, penicillin (10 U/mL), gentamicin (100 µg/mL), and streptomycin (10 µg/mL). Cells were maintained at 37° C in a 5% CO 2 environment until cell monolayers registered a transepithelial electrical resistance (TEER) above 2,000 ohms/cm 2 , a value consistent with cell polarization [11]. Monolayers of Int-407 cells did not achieve polarization and were not used in intoxication studies. Horseradish peroxidase (HRP) was added to the apical side (25 µg/well) of the transwells as an additional control for cell polarization as well as a control for paracellular diffusion [11]. Various concentrations of labeled ricin were then added to the apical chamber supernatant of the polarized HCT-8 cells. Every four hours post-intoxication, the TEER was measured and a small sample of medium was taken from the basolateral chamber in each well. These samples were tested by the Vero cytotoxicity assay for the presence of active ricin that had crossed the membrane. An enzymatic assay using TMB (3,3′,5,5′-tetramethylbenzidine) substrate (Bio-Rad, Hercules, CA) was used to detect HRP in the basolateral chamber. The reaction was stopped with 1M phosphoric acid, and absorbance was read at 405nm. We used 2-way ANOVA to analyze the transcytosis data. Additionally, at each time point, one transwell membrane was excised and examined with a Zeiss Pascal LSM confocal microscope (Carl Zeiss Microscopy, LLC, Thornwood, NY) to detect cell-associated ricin.

Organoid Model
We generated HCT-8 cell organoids and attempted to produce Int-407 cell organoids as previously described [12]. Briefly, hydrated small intestinal submucosa (SIS) scaffolding material (Cook Biotech, West Lafayette, IN) was cut into 3mm 2 sections under sterile conditions. Five scaffold sections were added to a rotating wall vessel [RWV (Synthecon, Houston, TX)] that contained 1x10 5 HCT-8 cells in 10 mL supplemented RPMI-1640 medium. These cultures were grown under microgravity conditions, by constant rotation of the vessels to allow free-fall suspension of the scaffold material in the center of the RWV, for 7 days at 37° C in a 5% CO2 environment. The medium in each RWV was changed every 2 days with care taken to leave sufficient fluid to cover the organoids during this exchange. Alexa Fluor ® 700-labeled ricin was then added to one vessel that contained five SIS pieces, while another vessel with five SIS pieces was left untreated. Toxin-treated and untreated organoids were harvested 1, 6, 12, 24, and 48 hours after treatment. Organoids were then fixed in 10% formalin, paraffin-embedded, and sectioned onto slides. Sections were stained with hematoxylin and eosin (H&E) to observe the organoid integrity and structure. Unstained sections were stained with DAPI and imaged with a confocal microscope equipped with a white light laser (Leica TCS-SP2, Leica Microsystems GmbH, Mannheim, Germany) to visualize Alexa Fluor ® 700-labeled ricin within the organoid samples.

Mice
Male CD-1 mice that weighed 10-12 grams were purchased from Charles River Laboratories (Wilmington, MA) and were used in all animal studies with one exception. In a preliminary LD 50 study in which ricin was administered by oral gavage, 6week-old male Swiss Webster mice (Charles River Laboratories) were used. All mice were housed in filter-top cages in an environmentally controlled room approved by the American Association for Accreditation of Laboratory Animal Care (AAALAC). Animals had access to food and water ad libitum unless stated otherwise. The protocol for these mouse studies was approved by the Institutional Animal Care and Use Committee (IACUC) of the Uniformed Services University.
To establish an LD 50 for ricin administered by intragastric gavage to mice, we followed a protocol similar to that reported by Smallshaw et al. [6]. Food was removed from the mice 20 hours prior to and for 4 hours after intoxication. Water was withheld 1 hour prior to and one hour after intoxication. Mice were then orally gavaged with PBS or 1 mg/kg, 5 mg/kg, 10 mg/kg, 25 mg/kg, or 50 mg/kg of ricin in PBS with a sterile 20 gauge disposable plastic feeding tube (0.9 mm x 30 mm; Solomon Scientific, San Antonio, TX). Morbidity and mortality were monitored over 7 days. Morbidity was defined as ruffled fur, lethargy, hunched posture, impaired ambulation that prevented the animals from reaching food and water, >25% weight loss, difficult or labored breathing, and the inability to remain upright. Mice that exhibited two or more symptoms were humanely euthanized by isoflurane overdose followed by cervical dislocation. We conducted probit analysis to determine the LD 50 of ricin administered orally. The value was 9.4 mg/kg with 95% confidence intervals of 7.5 to 11.5 mg/kg.

In Vivo Imaging
Mice were maintained for one week on the AIN-93M Purified Diet (Harlan, Madison, WI), a normal nutritive mouse food with reduced autofluorescent components [13]. The abdomens of isoflurane-anesthetized mice were shaved with an electric razor to reduce autofluorescence from fur during in vivo imaging. Food was withheld from the mice for 20 hours prior to and 4 hours after intoxication. Water was withheld for 1 hour prior to and 1 hour post-intoxication [6]. Typically, the equivalent of 1 LD 50 (9.4 mg/kg) of Alexa Fluor ® 700-labeled ricin in a volume of 0.3 mL PBS was intragastrically administered to each mouse by gavage. For some studies, Alexa Fluor ® 633-labeled ricin was used. At various times after ricin administration, groups of 5 mice were anesthetized by inhalation of isoflurane (3-4%) in oxygen and imaged at 690 nm in the Carestream MultiSpectral FX Pro in vivo imaging system (Bruker BioSpin, Woodbridge, CT). Immediately after fluorescent imaging, mice were x-rayed on all four sides to localize the fluorescent signal. Three mice at each time point were euthanized and their organs (stomach, small intestine, large intestine, cecum, liver, spleen, kidneys, heart, and lungs) were excised. Fluorescent images of the excised organs were recorded. The imaged organs were then preserved in 10% formalin and processed for microscopy.

Detection of Ricin in Mouse and Human Tissue by Immunofluorescence
Paraffin-embedded organs from intoxicated and buffertreated mice were cut into 5 µm sections that were then mounted onto glass slides. Organ sections on slides were deparaffinized and rehydrated with an ethanol gradient (100%-70%) followed by incubation of the slides in water. Sections were blocked with 3% bovine serum albumin (BSA) in PBS (BSA-PBS), immunostained with a 1:50 dilution of rabbit polyclonal antiserum against the ricin A subunit (BEI Resources, Manassas, VA) in PBS, followed by a 1:500 dilution of goat anti-rabbit IgG Alexa Fluor ® 488 (Life Technologies) in 1x PBS. The sections were then counterstained with 0.01% Evans Blue dye.
Slides that contained paraffin-embedded slices of normal human small intestine were obtained commercially (ProSci, Inc, Poway, CA). The sections were deparaffinized and rehydrated with an ethanol gradient (100%-70%), then incubated in water. Tissue sections were blocked with 3% BSA in PBS, then purified ricin (5 µg/ml) in BSA-PBS or ricin (5 µg/ml) preincubated with 0.1 M lactose for 30 minutes was applied for one hour at room temperature. Unbound ricin was washed from the slide, and the sections were fixed with buffered 10% formalin. The tissue was stained by immunofluorescence for ricin and counterstained with Evans Blue dye as above. Fluorescent images of both mouse and human tissues were obtained with a Zeiss Pascal LSM confocal microscope (Carl Zeiss Microscopy, LLC).

Ricin Can Cross Human Intestinal Epithelial Cells in Polarized Monolayers or Organoids without Apparent Damage
Two groups of investigators have previously reported histological changes in the small intestines of mice after animals were gavaged with ricin at doses considerably above the measured LD 50 [6] or somewhat below the estimated LD 50 [7]. We wanted to determine if such findings in mice might serve as a surrogate for observations in humans. Therefore, we first sought to assess whether ricin could translocate across intact human intestinal epithelium and to ask if any visible changes occurred. We addressed these question by culture of human intestinal cells in plates that contained transwell inserts to generate polarized monolayers [9]. Although others have shown ricin transcytosis through polarized intestinal epithelial cells [14,15], we deemed it important to repeat the experiments with the human colonic HCT-8 cell line since it has been characterized in both the transwell assay as well as the organoid model that is described below [11,12]. We also attempted to generate polarized monolayers from human small intestinal Int-407 cells, but we were unable to observe polarization in the transwell format or to generate an organoid with Int 407 cells.
HCT-8 cell monolayers in transwell chambers reached TEER levels above 2,000 ohms/cm 2 after 8-10 days in culture, a value consistent with that reported by Hurley et al. [11] We then used these polarized monolayers to assess whether active ricin could transcytose across an intact gut epithelial cell layer. Alexa Fluor ® 488-labeled ricin or ricin mixed with HRP was added to the apical chamber that contained the polarized HCT-8 cells, and every four hours the TEER across the cell monolayer was monitored. The TEER remained constant for up to12 hours after intoxication and began to decrease approximately 16 hours post-intoxication ( Figure 1A), although the values for the intoxicated monolayers remained above the polarization threshold of 2,000 ohms/cm 2 . Although our studies indicated the monolayer was intact at 16 hours, we did not assess whether inhibition of protein synthesis occurred at that time point, as might be predicted from in vitro studies published by Mantis et al [8]. By 24 hours, resistance was lost in ricintreated wells but was maintained in media-treated wells. We also removed samples from the media in the basolateral chambers at various time points and tested these aliquots for cytotoxic activity on Vero cells ( Figure 1B). By eight hours postintoxication, active toxin was detectable in the basolateral chamber, and toxin activity in that medium increased for up to16 hours post-exposure. However, HRP added with ricin was not detected at levels above those of HRP alone in the basolateral chamber until 16 hours (Table 1), a time point that correlated with a dramatic drop in TEER measurements ( Figure  1A). As HRP does not permeate a polarized monolayer [11], we took our findings as evidence that the monolayer was intact at 8 hours when ricin was detected basolaterally. Transwell membranes were removed at each time point to stain for bound ricin. Like polarized cells exposed to medium alone ( Figure  2A), no cell-associated ricin was evident 4 hours after ricin intoxication ( Figure 2B). However, we cannot rule out the possibility that some ricin was bound but was below the limit of detection by direct fluorescence. At 8 and 16 hours postintoxication, a time-dependent increase in toxin bound to cells on the membrane was observed; specifically, ricin appeared to bind selectively to small regions of cells, and the size of these regions expanded over time ( Figure 2C, D). In addition to the HCT-8 transwell cultures, we used a previously described three-dimensional (3-D) tissue culture model in which cells were grown in a RWV to provide a low sheer and low gravity environment that is considered to mimic the milieu in the intestine [16]. We added collagen-rich acellular scaffold material to provide a core on which multicellular tissue masses called organoids assembled. Prior studies in our laboratory showed that HCT-8 organoids are better differentiated, form tight junctions, and express surface markers and enzymes more characteristic of native tissue than do HCT-8 cells grown in conventional 2-D tissue culture [12]. After 7 days of incubation, we added Alexa Fluor ® 700-labeled ricin directly to RWVs in which organoids were growing. At 1, 6, 12, 24, and 48 hours, a ricin-intoxicated organoid was removed from its RWV, as was a corresponding control organoid from an unintoxicated RWV.
After one hour of ricin exposure, an H&E-stained organoid was intact and the cells appeared healthy ( Figure 3A). Alexa Fluor ® 700-labeled ricin was only detected in one region on the tissue by confocal microscopy ( Figure 3B). After six hours, the exterior cell layer of the H&E-stained organoid appeared to be intact; however, there was mild damage beneath the first few cell layers ( Figure 3C). Additionally, the labeled ricin appeared to localize beneath intact cells and may have been associated with the scaffolding material ( Figure 3D). Only a small quantity of labeled ricin appeared to be cell-associated. A similar phenotype was observed at 12 hours (data not shown). By 24 hours and, more prominently, 48 hours after intoxication, damage to ricin-exposed organoids was evident ( Figure 3E). By 24 hours post-intoxication, many of the HCT-8 cells were rounded and appeared to be detaching from the scaffold ( Figure 3E); by 48 hours, most cells were rounded and had detached from the scaffold (data not shown). Labeled ricin was associated with the remaining cells in the organoid at 48 hours (data not shown). These findings suggest that ricin moved across the first layer of cells in the organoid, and in some cases multiple cell layers, before any damage became evident, an observation consistent with our findings from the 2-D transwell experiments. Damage to the outer layer of the organoid did not occur until 24 hours after intoxication, and the damage was amplified by 48 hours. This time-dependent increase in damage probably occurred because the toxin remained in the media that supplemented the organoids in the RWV.

Movement of Ricin in vivo After Gavage of Mice
We attempted to track the movement of a lethal dose of ricin over time in mice that had been intoxicated by oral gavage. One of our goals was to ascertain whether damage to the intestine occurred, and, if so, was it before or after the toxin migrated to distal organs. We also sought to identify the organs that were targeted by toxin dissemination. To address these objectives, we first determined the LD 50 of our ricin preparation when administered to mice by oral gavage. For these in vivo studies, we followed the protocol of Smallshaw et al. [6], i.e. we fasted mice for 20 hours before and for 4 hours after oral intoxication. We determined the intragastric (i.g.) LD 50 of our purified ricin preparation in Swiss Webster and CD-1 mice, both of which are outbred strains, and found that the LD 50 was 9.4mg/kg for both strains. This LD 50 value is similar to that estimated by Yoder et al. for BALB/c mice [7] but is considerably higher than the 10 µg/kg reported by Smallshaw et al. for Swiss Webster mice [6]. In subsequent experiments, we used CD-1 mice.
We administered 9.4 mg/kg (1 LD 50 ) of Alexa Fluor ® 700-or Alexa Fluor ® 633-labeled ricin i.g. and captured fluorescent images of live mice at 0.5, 3, 8, 16, 24, 48, 72 and 96 hours ( Figure 4). The image from 0.5 hours is not shown because the signal was saturated. Fluorescence was clearly evident in a region that corresponds to the gastrointestinal tract, as determined from x-ray imaging conducted immediately after fluorescent imaging, for 48 hours and perhaps 72 hours as well. The intensity of the fluorescent signal diminished over time. We also administered the free Alexa Fluor ® 700 label as a control. Whole body imaging showed that the free label signal was very strong at 0.5 hour but was not apparent after 24  hours (data not shown); this finding supports our conclusion that the prolonged signal seen after administration of labeled ricin by gavage was specific rather than due to retained free label.
To associate the labeled ricin signal with specific organs, we euthanized three mice at each time point (unless noted otherwise) for ex vivo imaging. We then examined fluorescence intensities of the following excised organs from each mouse: stomach, small intestine, large intestine, cecum, kidneys, liver, spleen, heart and lungs [24 hour time point only for heart and lungs, (see Figure 5)]. The 0.5 hour images are not shown due to signal saturation. As expected based on whole mouse imaging, ricin was evident in the G.I. tract at 0.5 hour (data not shown) and persisted in the G.I. tract for 96 hours. The persistence of active ricin in the G.I. tract for up to 96 hours was confirmed by Vero cell assay of fecal pellets obtained at that time point (data not shown).
Fluorescence was first detected in the kidneys and liver 16 hours after intoxication ( Figure 5A) with the exception of one mouse that had signal in the kidneys after 5 hours (data not shown). Specific fluorescence signals from the kidneys and liver remained evident for 96 hours post-intoxication with labeled ricin (Figure 5A-5E, numbers 5 and 6, respectively, in each panel). For analysis of ricin dissemination 8 and 16 hours after intoxication, we used Alexa Fluor ® 633-labeled ricin to permit visualization of organ-associated toxin with a confocal microscope. Background fluorescence was noted in the PBS controls imaged at both wavelengths (620 and 690 nm) for both the stomach and liver samples at all time points (examples, Figure 5F numbers 1 and 6 and Figure 5G number 6). However, either the area of the signal over the organ or the intensity of the signal was consistently higher in organs from animals intoxicated with Alexa Fluor ® 700-labeled ricin than in organs of control animals administered PBS and imaged at 690 nm. All other organs distal to the G.I. tract were negative for detectable labeled ricin as assessed by ex vivo imaging. Organs removed from mice that were administered free label were also negative for fluorescence after 24 hours ( Figure 5H), an observation that suggests that ricin, and not just free label, moved to the kidneys as shown in Figure 5A-E.
Immunofluorescence (IF) was used to screen organs for the presence of ricin at each time point. We focused on those organs that had the greatest signal intensity above background by in vivo imaging, i.e. stomach and G.I. tract, kidneys, and liver. Although a strong fluorescent signal was observed over the stomach region of live mice at 24 hours, ricin was only sporadically detected in association with the lining of the stomach (data not shown). Similarly, very little ricin was seen bound to the epithelia in the cecum and large intestines of mice (data not shown). In contrast, ricin was clearly associated with the epithelium and the lamina propria of small intestinal tissue at 24 hours (compare Figure 6A and with PBS control in Figure 6B). The PBS control had some auto fluorescence likely due to food residue; however, the fluorescence was much less than that seen in sections from ricin intoxicated mice. For comparison, we asked whether ricin bound to normal human small intestinal sections in a similar manner. We overlaid sections of normal human small intestine with ricin ( Figure 6C) or PBS (Figure 6D), and we found that ricin bound the villi and goblet cells in human small intestinal sections as well ( Figure  6C). Background fluorescence in the absence of ricin was negligible ( Figure 6D). Note that our findings with ricin binding to normal human small intestinal tissue corroborates those of Mantis et al [17]. We further demonstrated that the binding of ricin to the small intestine sections was specific in that it was ablated by pretreatment of the toxin with lactose (6E), a sugar with known affinity to the ricin binding domain [18]. Ricin did not appear to bind to human colon sections (data not shown), a finding consistent with our failure to note binding of ricin to mouse colon sections (data not shown).
We observed an IF signal for ricin within kidney tubules 24 hours after intoxication ( Figure 7A). No fluorescent signal was observed in the kidneys of PBS-treated mice not exposed to ricin ( Figure 7B). Although in vivo imaging showed fluorescence in liver tissue after exposure to labeled ricin, we could not confirm microscopically by IF that ricin was in the liver due to high tissue auto-fluorescence. The remaining excised organs (i.e., the lungs, brain, spleen, and heart) were either clearly negative for ricin (control organs from PBStreated mice were also negative) or the results were inconclusive by IF. A summary of the number of animals subjected to whole body imaging at various time points postintoxication as well as the numbers of organs from these mice that were tested for the presence of ricin by IF is given in Table  2.

Histological Evaluation of Organs from Mice Gavaged with Ricin
Mice were orally gavaged with 10 mg/kg labeled ricin, unlabeled ricin or PBS. At 8,16,24,48,72, and 96 hours post intoxication organs were excised, fixed, sectioned, and stained with H&E. These stained tissue sections were obtained from, on average, 3 animals per time point from 8 to 96 hours after intoxication (See Table 3 for summary of details). Sections were assessed for damage in a blinded manner by a veterinary pathologist (author MAS). For small intestine sections, histological changes were not noted at 16 hours ( Figure 8A), but were present from 24 to 96 hours after ricin intoxication, as indicated by swollen/edematous areas below a mostly intact epithelium with the lamina propria dissociated from the epithelial layer (Figure 8 B, C). These changes were not observed in the PBS control section ( Figure 8D). In addition, kidneys from 2/5 mice that were sacrificed 24 hours postintoxication and one mouse that was sacrificed at 48 hours due to severe morbidity exhibited mild damage, i.e several vacuolated tubular epithelial cells when compared to PBS controls ( Figure 9A, B respectively). We conclude that no consistent damage to organs was evident that could explain why mice died following gavage with ricin at doses at or above the LD 50 .

Blood Chemistry Analyses of Sera from Mice Gavaged with Ricin
Since ricin did not appear to cause significant histological damage to any internal organs, we sought to determine the cause of death by analyzing blood chemistries from orally intoxicated CD-1 mice. Nine mice were gavaged with 10 LD 50 (94mg/kg) of ricin or with PBS, and sera were obtained 30 hours later. We were unable to obtain blood samples from two ricin-intoxicated mice due to significant morbidity. Blood chemistry analyses showed elevated levels of creatine phosphokinase (CPK), a muscle and cardiac enzyme; and alanine transaminase (ALT), aspartate aminotransferase (AST) and gamma-glutamyl transpeptidase (GGT), all of which are indicators of liver function abnormalities. The CPK level was 244% higher in ricin-treated mice than in those mice that received PBS. Additionally ALT, AST, and GGT were elevated 73%, 58%, and 100% respectively (Table 4). We also noted  decreased levels of CO 2 compared to mice that received PBS. Enzymes specific to kidney function [blood urea nitrogen (BUN) and creatinine] were not elevated despite our detection of bound ricin in the tubules. Perhaps due to outlier values in both the ricin-treated and PBS-treated groups, the data from ricinintoxicated mice were not significantly different from PBStreated mice. However, the trends observed were consistent with the values others have reported both in mouse and human cases of ricin intoxication [19,20]. Based on serum chemistry analysis that indicated a release of enzymes specific to the   liver and perhaps the heart (CPK elevation), the vacuolation of some kidney tubules (a finding that is suggestive of hypoxia), and the lack of significant damage to organs, we hypothesize that mice may have died from distributive shock. Shock has previously been hypothesized to be the cause of death in several cases in which humans ingested castor beans [5].

Discussion
Several major results were derived from this examination of ricin transport across intestinal sections from humans (in vitro) and mice (in vivo). First, we confirmed the findings of Mantis et al [17] that ricin can bind to human small intestinal sections on overlay and further demonstrated the specificity of this binding by blocking with lactose. Second, we showed that early after ricin treatment of human colonic cells in polarized monolayers or in organoids, the toxin transited across epithelial cells without apparent cellular damage. Third, and similarly, we discovered that after mice ingested ricin, the toxin traversed through the single layer of epithelial cells in the small intestine without apparent histological damage for up to 16 hours. Fourth, we found on imaging of organs from mice gavaged with labeled ricin that the first distal organ in mice to be targeted by ricin was the kidney. Fifth, blood chemistry analyses of mice gavaged with 10 LD 50 of ricin suggested cardiac and liver damage, possibly as a result of lack of oxygen to those tissues, without apparent kidney malfunctions. Consistent with these findings was the observation of decreased CO 2 levels, potentially due to lactic acidosis, in the blood of ricin-treated mice versus controls. Overall, we concluded that ricin gains access to the circulation without damaging the intestines (small intestinal tissue changes were not evident until after labeled toxin was seen in the kidneys by ex vivo organ imaging), moves through the blood to target the kidneys and liver (IF data analyses on the liver were confounded by high autofluorescence of that organ), and appears to cause lethal vascular collapse or distributive shock as suggested by blood chemistry values and little or no observable histological damage to tissues. Similar to our work, ricin has been found in the liver [21,22] and kidney [23] by others. Additionally, mice that received ricin i.p. had similar blood chemistry data in which the liver enzymes were elevated but the kidney levels were normal [19]. Further analysis revealed both hepatotoxicity and nephrotoxicity in that study [19].
The absence of apparent damage to the small intestine of ricin-gavaged mice before 24 hours of intoxication is consistent with reports that suggest that only a small fraction of internalized ricin follows the retrograde transport pathway to exert its toxic effects on ribosomes and that the remainder is recycled or degraded [3]. Our veterinary pathologist noted that the histological alterations observed in the small intestine 24 hours and later were similar to those reported for toxin-treated mice by Yoder et al. [7]. Nevertheless, we saw no significant breaches of the small intestinal epithelial layer in toxin-treated tissues. Furthermore, our 2-D transwell and 3-D organoid data, which were generated with colonic rather than small intestinal epithelial cells because we could not generate organoids with Int-407 cells, support the hypothesis that ricin travels across the small intestinal epithelium without damage early in intoxication. We do note that, when the toxin remained in the in vitro models, damage did occur.
We have no explanation as to why our orogastric ricin LD 50 value of 9.4 mg/kg was so much higher than that reported by Smallshaw et al. (10µg/kg). We followed the pre-and postintoxication fasting protocol that they described. In addition, our ricin preparations were of comparable toxicity to the ricin deposited by Dr. Vitteta into the BEI Resources Repository; when tested in parallel, the Vero cell CD 50 /mg protein were similar (on the order of 5 x10 7 ) for both lots of ricin.
Our model of ricin oral intoxication is as follows. By methods unknown, ricin enters the bloodstream within hours after oral gavage and circulates throughout the body [21,22,24]. Due to a lack of consistent observable damage in any organ except the small intestine at 24 hours of intoxication and beyond, we predict that ricin causes the mouse to go into distributive shock. Distributive shock causes a loss of peripheral vascular resistance, similar to septic or anaphylactic shock. The loss of vascular resistance and the idea that ricin might attack the vascular system are consistent with findings by Baluna and Vitetta, who observed vascular leakage in mice given the ricin A subunit linked to a therapeutic (RTA-IT) [25]. The loss of vascular resistance after intoxication causes an increased cardiac effort to pump blood to various organs, which, in turn, causes a significant release of CPK. Our model would further suggest that, due to lack of sufficient blood to the liver, a state known as "shock liver" develops whereby there is a nonspecific release of liver enzymes, potentially leading to damage within the liver. Consistent with our work, He et al. identified ricin in the liver by immunoPCR but did not observe any lesions [24]. Finally, our model would indicate that the developing state of ischemia results in a decrease in CO 2 levels, leading to lactic acidosis and possibly vacuolated tubules in the kidney. Mice then succumb to intoxication. Our model is consistent with what has been observed in cases of accidental ingestion of castor beans in humans [20].
These results may suggest methods to optimize intervention strategies for ricin-intoxicated individuals. The standard method for treatment of patients who have ingested ricin or castor beans is to administer activated charcoal to absorb the toxin in the gut. Because we saw prolonged retention of ricin in the G.I. tracts of toxin-gavaged mice, and at least some of the toxin remained active as evidenced by the finding of toxicity in fecal pellets, we believe that this adsorption approach is both rational and important for prevention of further uptake of toxin from the gut. As more specific therapy, the oral administration of lactose to block absorption of ricin that remains in the gut has also been suggested [26]. Additionally, ricin could potentially be neutralized systemically through early passive administration of neutralizing antibodies. Other strategies to   cells or organs via liposomes or through conjugation of antibody to tissue-specific ligands or antibodies. Active immunization to evoke ricin-neutralizing antibodies that would protect targeted and particularly vulnerable populations against ricin intoxication by any route would be optimal.