To characterize the interactions between ricin and cells of the lung, mice were intranasally exposed to fluorescently-labeled ricin at a lethal dose of 2 LD₅₀. We monitored toxin-associated cells by examining the forward scatter (FSC) profile of pulmonary cells isolated 3 h after intoxication. Intoxication led to the labeling of a defined population of cells, mainly of hematopoietic (CD45⁺) origin (Figure 1A). Since fluorescent toxin identifies both internal and external cell-associated toxin, ricin binding was monitored by an additional technique, utilizing purified anti-ricin antibody, RAF5, for tagging the toxin. It should be noted that when this technique is employed, the detection of cell-associated ricin reflects a primary interaction between the antibody and the toxin, which occurred at the cell exterior. Indeed, monitoring toxin-associated cells by this technique revealed a distinctly different pattern, attesting to the binding of the toxin mainly to parenchymal (CD45⁻) cells (Figure 1B). Identification of ricin binding to CD45⁺ cells mostly by toxin fluorescence and not by antibodies indicates that in these cells, the bulk of cell-associated toxin has already been internalized. In contrast, ricin-bound CD45⁻ cells were mainly detectable by antibodies, indicating that in these cells, ricin was associated with the exterior surface of the cells. Toxin/cell interactions in CD45⁻ cells could not be observed by the direct fluorescent-toxin monitoring technique, suggesting that this technique is less sensitive than the antibody-tagging technique.

These findings prompted us to determine the kinetics of ricin binding to individual cell populations of the lung. Mice were intoxicated with fluorescently-labeled ricin, and lung cells isolated at different time points were then stained for anti-ricin antibodies and analyzed for ricin binding. In MΦs and DCs, as gated in Figure S1A, two temporally-distinct peaks were detected by the two techniques (Figure 2A,B). Whilst peak binding at 3–6 h after intoxication was detected by fluorescent toxin, labeling with antibodies identified a later peak, at 18 h after intoxication. The failure of the sensitive antibody-based technique to detect toxin-associated cells at the earlier time point indicated that the peak binding at this time point relates to cells that have already internalized the toxin. However, the ratio between the early and late peaks of toxin association was different in the two cell types. In the case of MΦs (Figure 2A), the greater amount of toxin was internalized at the early time point, while in the case of DCs (Figure 2B), most of the association of toxin with the cells occurred at the late time point, remaining bound to the cell exterior. Early binding of ricin by MΦs preceded that of DCs, while significant binding to MΦs was detected already 1 h after intoxication; binding to DCs was observed at 3 h after exposure (Figure S2A–C). The determination of the binding profile to B cells (Figure S1B) identified a single late peak at 18 h after intoxication by both techniques (Figure 2C), while toxin binding to neutrophils (Figure S2A) could not be detected at all (Figure S2D). Analysis of toxin interactions with parenchymal cells demonstrated that binding to epithelial cells (Figure S1C) is characterized by a single peak detected by anti-ricin antibodies at an early time point of 6 h (Figure 2D). In contrast, ricin binding to endothelial cells (Figure S1C) was discernable only at 24 h after intoxication and did not reach peak levels within the time frame of these experiments (Figure 2E). The spatial localization of the different cells types within the lungs could play a role in the bi-phasic shaping of the ricin binding profile. Following pulmonary intoxication, ricin first encounters the MΦs located in the alveolar lumen and the DCs protruding through the epithelial network, and only then, the toxin has access to the interstitium. However, the fact that a single cell type may display both early and late binding peaks, as in the case of MΦs and DCs, suggests that the composite binding patterns are also an outcome of different mechanisms of binding. The B-chain of ricin is a galactose-specific lectin that binds to glycoproteins or glycolipids at the cell surface [2,3]; yet, in addition to this canonical route, ricin can enter cells through a secondary route, via the mannose receptor that is present on MΦs [12]. This receptor binds to high mannose oligosaccharide chains present on both A and B subunits of the toxin and is more efficient in delivering ricin to the cytoplasm [13]. To evaluate the possible role of this alternative binding pathway, mannose receptor density was measured for different cell types of the lung (Figure 2F). In both MΦs and DCs, the presence of this highly efficient receptor could explain the early peak of ricin binding and the measured difference in receptor density between the two cell types, correlating well with the relative amounts of toxin that were associated with these cells (Figure 2A,B). Epithelial cells were found to express the mannose receptor at levels that did not fall from those of MΦs (Figure 2F) and, indeed, displayed a high early binding peak (Figure 2D), while B cells and endothelial cells, both of which were devoid of the mannose receptor (Figure 2F), did not display an early peak of binding (Figure 2C,E). Toxin association via galactose, which is less efficient than via the mannose receptor, could be responsible for the later peaks of binding observed in the different cell types.

Our findings show that ricin binds to cells of the lung in a differential manner, perhaps through different receptors. Upon entry to the cytosol, ricin inhibits protein synthesis by the irreversible inactivation of ribosomes, which in turn leads to cell death. Indeed, massive apoptotic changes were readily detected in the lung tissue at late time points after intoxication (24–72 h; Figure 3). In contrast, apoptosis could not be detected at early time points, suggesting that cell elimination early after intoxication, such as that displayed for MΦs and DCs, stems from necrotic death or pyroptosis.

To appreciate the correlation between binding kinetics and alterations in the lung cell populations, we performed a quantitative analysis of pulmonary cells following ricin intoxication. Evaluation of the alterations in MΦs allowed us to determine that a two-fold reduction occurred as early as 18 h after intoxication (Figure 4A) in line with our finding that ricin binding to this cell type occurred early (Figure 2A). Likewise, a two-fold reduction in the DCs population could be discerned already at 18 h after intoxication (Figure 4B). Profiling of the B cell and endothelial cell populations, both of which were found to bind toxin at late time points (Figure 2C,E), demonstrated that a two-fold reduction in these cells was reached at 48 h and 72 h after intoxication, respectively (Figure 4C,D). In the case of epithelial cells, although binding of toxin occurred early after intoxication (Figure 2D), a significant decrease in their number was evidenced only later: epithelial cell counts dropped from (3.3 ± 0.8) × 10⁶ down to (1.2 ± 0.5) × 10⁶ at 48 h after intoxication (Figure 4E). In spite of the cytotoxic effect of ricin and the marked elimination observed for various cell types, the overall number of cells comprising the lungs increased over time. Thus, at 6 h after intoxication, the number of lung cells (48.5 ± 6) × 10⁶ increased to (65.6 ± 6) × 10⁶, while 48 h after intoxication, numbers reached (95 ± 13) × 10⁶ cells (Figure 4F). This striking rise in cell number correlated with the concomitant recruitment of neutrophils to the lungs. Neutrophil counts (~2 × 10⁶ cells) were tripled 6 h after exposure to ricin, and at 48–72 h after intoxication, the size of this cell population was nearly equal to the number of cells comprising the entire lungs in naive mice (Figure 4F).

Lung alveolar epithelial cells are of two types, respiratory alveolar type I (ATI) and surfactant-secreting alveolar type II (ATII) cells [14], the latter being dispersed between the ATI cells. ATI and ATII cells can be differentiated by specific labeling of podoplanin (T1α) [15] and pro-surfactant C (pro-SPC) [16], respectively. FACS analysis demonstrated that ATII cells (CD45⁻, CD31⁻, pro-SPC⁺), but not ATI cells (CD45⁻, CD31⁻, T1α⁺) (Figure S3), were depleted from the intoxicated lung at 48 h after exposure (Figure 5A,B). This unexpected finding, that epithelial elimination in the intoxicated lung was restricted to ATII cells, was also manifested by immunohistochemical analysis (Figure 5C). TUNEL staining for apoptosis was prevalent in pro-SPC labeled cells, suggesting that apoptosis of ATII cells precedes their elimination (Figure 6A). It should be mentioned that at 24 h after ricin intoxication, but not at later time points, pro-SPC⁺ ring structures were detected in the damaged parenchyma (Figure 6B). Moreover, pro-SPC expression, undetected in the bronchial epithelia of naive lungs, was readily visualized in the bronchial epithelia of intoxicated lungs (Figure 6C). This incongruous de novo expression of the surfactant precursor presumably reflects an attempt to restore pulmonary functionality by means of synthesizing surfactant.

Animal experiments were performed in accordance with the Israeli law and approved by the Ethics Committee for animal experiments at Israel Institute for Biological Research. Treatment of animals was in accordance with regulations outlined in the USDA Animal Welfare Act and the conditions specified in the Guide for Care and Use of Laboratory Animals (National Institute of Health, 1996). Mice were female CD-1 (Charles River Laboratories Ltd., Margate, UK) weighing 27–32 g.

Ricin was purified as described [29] and conjugated (1 mg) with the Alexa Fluor^® 488 protein labeling kit (Molecular Probes, Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions. The cytotoxicity of labelled toxin (~5 dye/protein (mol/mol)) was determined in a cell-based assay developed in the past [31]. Briefly, labeled ricin was added to HEK-293 cell cultures, which secrete the enzyme acetylcholinesterase (AChE) in a constitutive manner. Secreted AChE was measured in the cell growth medium at 18 h post-exposure, and activity levels were compared to those measured for non-labelled ricin. This allowed us to determine a ~2-fold reduction in toxicity of labeled ricin. Intranasal intoxication with ricin at a dose of 2LD₅₀ (unlabeled and labeled ricin, 7 and 14 µg/kg, respectively, diluted in PBS) was performed after anesthetizing mice i.p. with ketamine (1.9 mg/mouse) and xylazine (0.19 mg/mouse). The total volume of toxin, 50 µL, was slowly applied (25 µL/nostril) using a gel-loading tip (×0.6 mm, USA Scientific, Orlando, FL, USA).

Ricin was coupled to an affinity column (AminoLink^® kit, Pierce Biotechnology, Waltham, MA, USA) according to the manufacturer’s instructions. Rabbit anti-ricin hyperimmune sera were loaded on to the column. Following incubation (1 h), the column was washed (PBS), and the specific antibody was eluted (50 mM glycine-HCl, pH 2.7, 10% ethylene glycol). The antibody concentration was 1.3 mg/mL, and K_D = 1 nM (Octet^red). For RAF5 staining, cells were incubated on ice for 15 min with the purified RAF5 antibody, then washed and stained for an additional 15 min with secondary donkey anti-rabbit allophycocyanin (APC)-labeled antibody (Jackson ImmunoResearch, West Grove, PA, USA). To avoid internalization of the stained ricin with RAF5 into the cells, during sample processing and staining, NaN₃ was added to the buffers (final concentration = 0.05%), and samples were kept on ice during the whole experiment.

Lungs harvested and cut into small pieces were digested (2 h, 37 °C) with 4 mg/mL collagenase D (Roche, Mannheim, Germany). Tissue was meshed through a 40-µm cell strainer, and RBCs were lysed. Cells were stained using antibodies coupled to: phycoerythrin (PE) (anti-Gr-1 (RB6-8C5), I-A/I-E (M5/114.15.2), CD19 (1D3), CD31 (PECAM-1, 390)); allophycocyanin (APC) (anti-CD11c (N418), CD326 (EpCAM, G8.8)); peridinin chlorophyll protein-Cy5.5 (PerCP-Cy5.5) (F4/80 (BM8)); biotinylated (anti-CD45 (30-F11)); Alexa Fluor 488 (anti-podoplanin (T1alpha, 8.1.1)) and purified polyclonal anti-pro-SPC (Millipore, Temecula, CA, USA) followed by donkey anti-rabbit IgG (H+L) coupled to APC (Jackson ImmunoResearch, West Grove, PA, USA). Biotinylated antibodies were followed by staining with PerCP-Cy5.5-labeled streptavidin. Cells were defined as: neutrophils (CD11c⁻, Gr-1^high); alveolar MΦs (autofluorescent, CD11c^high, Gr-1^int, F4/80⁺, MHCII^int); DCs (CD11c^int, Gr-1⁻, F4/80⁻, MHCII^high); B cells (CD19⁺); endothelial cells (CD45⁻, CD31⁺); and epithelial cells (CD45⁻, CD326⁺). For mannose receptor staining, cells were stained for biotinylated-anti-CD206 (C068c2). Unless otherwise indicated, the reagents were obtained from BioLegend (San Diego, CA, USA) and eBioscience (San Diego, CA, USA). Cells were analyzed on FACSCalibur (BD Biosciences, San Jose, CA, USA) using FlowJo software (version 7.1.2, Tree Star, Ashland, OR, USA). The number of each cell type was calculated by multiplying the percentage of a specific cell population (as gated and determined by FACS) by the total lung cell count.

Lungs perfused with PBS were harvested and immersed in neutral buffered formalin (4%, 2 weeks, room temperature) prior to embedding. Sections (5 µM) were mounted on glass slides and deparaffinized. Antigen retrieval was performed by incubation in Target Retrieval Solution (DAKO, 30 min, 95 °C). After blocking in 5% BSA in PBS, slides were incubated (overnight, 4 °C) with purified polyclonal anti-pro-SPC (Millipore, Temecula, CA, USA) and anti-podoplanin (T1alpha, 8.1.1, eBioscience, San Diego, CA, USA). Staining with primary antibodies was followed by Alexa Fluor 594-coupled donkey anti-rabbit and FITC-coupled goat anti-Armenian hamster antibodies (Molecular probes, Eugene, OR, USA), respectively. For nuclear staining, slides were mounted with Prolong^® Gold antifade reagent containing DAPI (Molecular probes, Eugene, OR, USA). For apoptotic staining [32], slides were incubated with 20 µg/mL proteinase K diluted in TE buffer (30 min, 37 °C) and then stained for TUNEL (Serotec, Oxford, UK) according to the manufacturer’s instructions. Analysis was performed using an LSM 710 confocal microscope (Carl Zeiss, Oberkochen, Germany) equipped with the following lasers: argon multiline (458/488/514 nm), diode 405 nm, DPSS 561 nm and helium-neon 633 nm.

Two-tailed t-tests were used to compare groups. Significance was set at p < 0.05. Analysis was performed using GraphPad Prism. Data are represented as the means ± SEM.