3. DiscussionRicin toxin has the potential to be used as a weapon of bioterrorism but may also be a therapeutic agent in immunotoxins. Ricin binds promiscuously to cell-surface glycans found on all cells. Therefore, the clinical syndrome associated with ricin toxicosis is highly dependent on the route of exposure, although one commonality is the induction of a rapid and profound inflammatory response [16,28,29,30,31]. We were surprised to observe that hypoglycemia was the sole change in clinical chemistry analyses following parenteral administration of ricin [21]. To understand the mechanism(s) that result in this fatal hypoglycemia, as well as the response to counter this phenomenon, we have examined the systemic, hepatic, and pancreatic regulators of glucose metabolism in mice injected with ricin toxin. Because the observed hypoglycemia did not develop in mice until 6–8 h post injection (Figure 1), we postulate that changes prior to this are primarily in response to ricin, whereas alterations after that time reflect both ricin toxicity and the homeostatic response to the developing hypoglycemia. Early and dramatic responses occurred within the first 2 h post ricin, including alterations of specific cytokines (Figure 3), as well as suppression of hepatic G6Pase expression (Figure 10).We initially suspected that increased production of insulin or a failure to produce glucagon might be the basis of hypoglycemia. However, we found that levels of insulin did not increase until a late stage, when hypoglycemia was already established (Figure 3 and Figure 4). Paradoxically we observed an early and impressive increase in the size of pancreatic β cells (Figure 8). We also found evidence of insulin resistance through a glucose tolerance test (Figure 5) and by directly measuring insulin signaling (Figure 6). However, most importantly, we have shown the ability of ricin to induce hypoglycemia despite STZ-induced β-cell destruction (Figure 7), suggesting that insulin plays little role, if any, in the development of ricin-induced hypoglycemia. The β-cell hypertrophy likely indicates that the cells are responding to an unidentified signal, perhaps resulting from the developing insulin resistance. The insulin resistance that developed in concordance with developing hypoglycemia may represent a downstream defense response, preventing any further lowering of the blood sugar by insulin. Surprisingly, there was no evidence that the pancreas responded to the developing hypoglycemia by producing glucagon (Figure 3, Figure 4 and Figure 9), and in fact glucagon mRNA levels were lower when the hypoglycemia was at its nadir.Gene expression analyses have identified several alterations that could contribute to the development of hypoglycemia. The earliest and most consistent change is a ~10–30× drop in mRNA encoding hepatic G6Pase, accompanied by a decrease in immunoreactive G6Pase protein and an increase in hepatic glycogen stores. Unexpectedly, there is a decrease in hepatic glucose-6-phosphate, suggesting that in the absence of G6Pase activity, the substrate is shunted to other pathways rather than accumulating, possibly to the pentose phosphate pathway to generate NADPH to cope with oxidative stress. Our metabolomic analysis of ricin-treated mice, published in a companion manuscript, support this contention [48]. The decrease in hepatic glucokinase expression that occurs at a late stage also contributes to the decreased level of glucose-6-phosphate. The removal of the phosphate by G6Pase is the final step in the release of free glucose into the circulation from hepatic glycogen stores. A defect in the genes encoding G6Pase leads to GSD-I, which is characterized by hypoglycemia, hepatomegaly, hyperlipidemia, hyperuricemia, and lactic acidosis [36,37,38]. G6Pase−/− mice have similar biochemical findings to human patients with GSD-I, except for the absence of lactic acidosis [49], which is comparable to what we have observed. G6Pase hydrolyzes glucose-6-phosphate so that it may enter the bloodstream as glucose. A deficit results in a decrease in the amount of glucose available to move from stores in the liver to other organs in times of low blood glucose. Because a decrease in G6Pase expression is the earliest and most profound alteration in gene expression that we observed, and the resulting syndrome resembles the effect of GSD-I and G6Pase knockout in mice, we believe it to be a primary driver of hypoglycemia following ricin administration.Other gene expression alterations having a potential impact on glycemic control were also observed. Hepatic glucokinase was markedly suppressed at later time points, whereas pancreatic glucokinase expression was somewhat increased earlier, before it leveled off. Enlarged islets, resulting from β-cell proliferation, and neonatal hypoglycemia have been reported in a patient with an activating mutation in glucokinase [46], however we did not observe evidence of β-cell proliferation. It has been proposed that hepatic glucokinase plays a central role in glucose homeostasis by functioning as a glucose sensor that is induced by a high insulin to glucagon ratio, such as that seen after feeding [50]. Thus, the decrease in insulin signaling that we observed in the liver (Figure 6) could account for the downregulation of glucokinase expression. GLUT2 expression was significantly diminished with time in both the pancreas and liver. GLUT2 is bidirectional and critical for the movement of glucose across cell membranes [51]. Decreased expression contributes to hypoglycemia due to a lack of glucose distribution from the liver to other tissues. This is observed in patients with GSD-XI, Fanconi–Bickel syndrome, which is characterized by a mutation in the gene encoding for GLUT2 rendering it non-functional [47]. In our study, glycogenolysis and gluconeogenesis were observed to be downregulated in the liver. Higher concentrations of glycogen at the same timepoint that the development of hypoglycemia was observed is indicative of a lack of glycogenolysis. Furthermore, the decrease in the expression of phosphoenolpyruvate carboxykinase in the liver (Figure 10), an enzyme involved in the first bypass of gluconeogenesis and catalyzing the conversion of oxaloacetate to phosphoenolpyruvate, results in a decrease of gluconeogenesis.In addition to alterations in glucose metabolism, we observed a marked cytokine response that resembles a cytokine storm in many, but not all, regards [52]. We observed a rapid and profound depletion of mononuclear leukocytes (Figure 2F) which may serve to initiate the cytokine storm. Lymphoid depletion has previously been observed following in vivo ricin exposure [17,26]. The pro-inflammatory cytokine TNF-α rises rapidly, but other cytokines associated with the development of a cytokine storm rise late, if at all: IL-1, IFN-γ, IL-17, IL-6. In addition to proinflammatory cytokines, plasma levels of chemokines, soluble cytokine receptors, and even anti-inflammatory cytokine levels rise. The anti-inflammatory and immunosuppressive cytokines, as well as soluble cytokine receptors, may serve as a homeostatic response to dampen the inflammation. Most notably IL-10, which can both promote and suppress inflammation [40], rises within an hour, but then falls. A similar relationship between TNF-α and IL-10 has been observed in other inflammatory syndromes [41,42,43].A key question is whether the inflammatory and hypoglycemic responses are causally related. The relationship of chronic inflammation to type 2 diabetes and insulin resistance is well established [53,54]. Increased concentrations of TNF-α have been shown to induce insulin resistance by increasing phosphorylation of the insulin signaling pathway in animal and cell models [55], although the administration of anti-TNF-α mAbs had little effect in patients with type 2 diabetes. We observed increased TNF-α concentrations and insulin resistance (Figure 5), decreased phosphorylation along the signaling pathway (Figure 6), and of course hypoglycemia rather than hyperglycemia. These data suggest that the effects of TNF-α on glycemic control may differ between acute and chronic inflammatory processes. Reports that TNF-α administration can induce hypoglycemia and that it decreases G6Pase expression by activation of NF-κB provide a link between ricin-induced inflammatory response and hypoglycemia [33,34,35]. Because the Kupffer cells and liver sinusoidal endothelial cells are especially sensitive to ricin’s effects in vivo [56], local synthesis of TNF-α, either directly by these cells or in response to danger signals released upon cellular damage [57,58], may make the hepatocytes especially susceptible to TNF-α effects, including G6Pase downregulation. A protective mechanism that prevents the development of hypoglycemia by countering oxidative inhibition of liver GTPase has been described in an experimental model of sepsis [59]. This mechanism does not appear to offer such protection following ricin injection, although other aspects of ricin-induced hypoglycemia resemble the hypoglycemia observed in sepsis, which has traditionally been considered to result from depletion of hepatic glycogen stores [60].A plausible mechanism for the development of hypoglycemia following ricin administration is that the toxin induces a rapid and profound inflammatory response including the rapid induction of TNF-α. This in turn downregulates the expression of G6Pase, preventing hepatic gluconeogenesis leading eventually to a syndrome similar to GSD-I, including the development of systemic hypoglycemia. Whether this path would occur in humans is not known. Inflammatory responses in mice and humans are known to broadly differ [61]. At least one clinical report of ricin toxicosis describes hypoglycemia, but does not indicate whether specific measures were taken to control the low blood glucose in this patient [23]. Other clinical reports do not indicate blood glucose, and direct follow-up with the authors failed to elucidate the question. We did not observe hypoglycemia in macaques exposed to aerosolized ricin [16].It is well established that apoptotic cell-death plays a major role in ricin induced cytotoxicity. In this manuscript, we have focused on evaluating the effects of ricin on inflammation and glucose homeostasis and have not examined apoptosis. The time course of the inflammatory response and hypoglycemia is considerably more rapid than that of apoptosis. We have observed that when ricin is directly applied to cells, blebbing and other morphological changes are observed within minutes, but that it takes 12–18 h for apoptosis to begin [20]. Here, we report changes in the level of TNF- α and G6Pase occurring within 1 h, and alterations in glucose homeostasis begin 4–6 h following ricin administration. Thus, the kinetics suggest that apoptosis is not the primary driver of the observed hypoglycemia. When the fatal drop in blood sugar was prevented by streptozotocin-induced diabetes, tissue damage and cell death led to fatal outcomes 2–3 days later, demonstrating the difference in the kinetics of hypoglycemia and apoptosis in these mice.In this report, we have evaluated gene expression in the pancreas and liver using mRNA RT-PCR, rather than by measuring protein levels. Messenger RNA levels are more responsive to changes in stimuli, and studying the mRNA expression of multiple genes is experimentally more direct than protein analyses. In some key cases we measured both mRNA and protein. For insulin and glucagon, plasma levels of protein and pancreatic mRNA were measured (Figure 3, Figure 4 and Figure 9), while G6Pase hepatic levels were measured for both (Figure 10). For G6Pase, the mRNA levels dropped more rapidly and severely than the level of enzyme. One concern about measuring mRNA levels in ricin-treated cells is that ricin affects both transcription and translation [8] and thus the quality or the quantity of the mRNA may suffer, affecting the PCR results. We have observed in cell lines (S.H. Pincus, et al., unpublished data) and in the mice reported here that ricin treatment affects neither RNA quantity nor quality (as measured by RIN) until cell death is observed with vital stains. Moreover, all the RT-PCR results are normalized to the housekeeping gene GAPDH. The normalization would account for differences in mRNA resulting from the overall health of the cells.

The in vivo manifestations of ricin toxicity are highly dependent upon the route of administration because of ricin’s promiscuous cell-binding abilities. Nevertheless, there are important commonalities to the response, notably apoptosis and inflammation. That parenteral administration of ricin would lead to a burst of pro-inflammatory cytokines within hours was predictable. That this would lead to fatal hypoglycemia was not.

To further understand the consequences of these changes, we have performed metabolite profiling analyses of the livers of ricin treated-mice and as a control, food-deprived animals. The results show that each treatment induces a metabolically distinct hypoglycemic state [48]. Understanding the metabolic causes, consequences, and response to hypoglycemia has direct relevance for the management of ricin intoxications, diabetes mellitus, and other metabolic syndromes. In the future we propose to extend these analyses to include insulin-induced hypoglycemia, because of the important adverse consequences of hypoglycemia on the long-term management of diabetes mellitus [62,63,64]. 4. Materials and Methods 4.1. MiceOutbred Swiss (Cr:NIH(S), Charles River Laboratories, Wilmington, MA, USA) mice were studied, including 164 females and 25 males. Outbred mice were chosen to avoid strain-specific genetic effects. The female to male imbalance resulted from different housing requirements. Outbred mice were between 2 and 5 months of age and weighed between 17–40 g. The mice were treated as follows: fasted (12 mice), injected with saline (34 mice) or ricin 30 µg/kg (124 mice), 15 µg/kg (1 mouse), 7.5 µg/kg (6 mice), 2.5µg/kg (6 mice), ricin A-chain 60 µg/kg (6 mice), or ricin B-chain 60 µg/kg (6 mice). Injections were ip (177 mice) or iv (12 mice). NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG, Jackson Labs, Bar Harbor, ME, USA) are defective in both adaptive and innate immunity, as well as in cytokine signaling. A total of 14 NSG mice were used: 8 females that were 10 months old, 3 females and 3 males that were 3–4 months old. The mice were 18–30 g and were injected with ricin 30 µg/kg. Fifty-six inbred BALB/c mice (colony breeders originated from Jackson Laboratory, and the colony was maintained at Montana State University’s Animal Resource Center), half male and half female, were studied. Inbred mice limit variability and the strain specific inflammatory response is well documented for BALB/c mice [65,66]. All of the mice were 4–11 weeks old, weighing 20–30 g when studied. BALB/c mice were injected with saline (16 mice) or ricin 30 µg/kg (40 mice). All strains of mice were monitored and euthanized at specific time points.Mice were housed in Tecniplast caging West Chester, PA, USA and individually ventilated with HEPA filters and aspen wood chips (Sanichips, P.J. Murphy’s, Montville, NJ, USA). For enrichment, the mice were given a red mouse house (Bioserve, Flemington, NJ, USA), a nestlet (Ancare, Bellmore, NY, USA, and EnviroDry (Shephard Specialty Papers, Amherst, MA, USA). All cage components were autoclaved prior to use. The mice were fed with irradiated PicoLab Rodent Diet 20 (PMI Nutritional, North Arden Hills, MN, USA). The animals had access to water and feed for the entire experiment, unless otherwise noted. The ambient air temperature was 70–72 °F with a minimum of 30% humidity. The light was on from 5 a.m. to 7 p.m. and off for the remainder of the day for a 14/10 cycle. The female mice were housed 4–5 mice per cage and the males 1–4, depending upon age and aggressiveness.Previous experiments demonstrated the degree of experimental variability [21]. Based on these results, power analysis was performed to determine the group size. It was determined that 6–8 animals per group provides a >90% probability of obtaining statistically significant differences (p

Animals were bled > 1 d prior to ricin injection into a heparinized tube or a blood glucose test strip. The mice were bled from the saphenous, facial, or submandibular vein, at the discretion of the animal technician. The mice were injected ip, unless otherwise noted, with 10 µL of volume per gram body weight, via a 27 ga 3/8″ needle. The animals were monitored by AALAS-certified laboratory animal technicians every 20–30 min for the initial 2 h post injection, and every 2 h to 14 h thereafter. Any mice demonstrating signs of distress, pain or suffering were euthanized. The specific criteria for early euthanasia were loss of appetite, markedly diminished activity levels, lethargy, neurologic signs, unexpected behaviors, mottled fur, hunching, difficulty breathing, respiratory sounds, or blood glucose < 25 mg/dL.

At the time of sacrifice, a terminal bleed was performed by intracardiac puncture while under anesthesia. Animals were anesthetized using isoflurane by inhalation, dosing was accomplished by adjusting the position of the nose cone, and anesthesia was confirmed by loss of toe withdrawal reflex. Isoflurane was administered for the duration of the intracardiac bleed and continued until euthanasia. Euthanasia was performed by a cervical dislocation and thoracic incision. The liver, skeletal muscle, pancreas, and plasma were then removed from the mice.

4.2. Abs, Reagents, and Test KitsRicin holotoxin (5 mg/mL, Ricin Agglutinin II (RCA 60)*, L-1090), ricin A chain (L-1190), and ricin B chain (L-1290) were obtained from Vector Laboratories. The ricin holotoxin (5 mg/mL) was in a saline solution containing 10 mM sodium phosphate, 0.15 M NaCl, pH 7.8, 0.08% sodium azide, 5 mM galactose, then diluted further in PBS prior to use. STZ (Sigma, St. Louis, MO USA) was dissolved immediately prior to use in citrate buffer at pH 4.5 to 0.1 M or 0.02 M. Fluorescent Abs to murine cell surface markers were obtained from eBiosciences (San Diego, CA, USA): PE-Cy7-anti-CD45R/B220, PE-hamster anti-CD3e, and APC-rat anti-CD14 (clone rmC5-3). Similarly labeled isotype matched controls were also obtained from eBiosciences. Abs to detect phosphorylated and unphosphorylated members of the insulin signaling pathway were purchased from Santa Cruz Biotechnology: phospho-IRS-1, IRS-1, phospho-AKT, AKT, phospho-GSK-3β and GSK-3 β. Anti-ricin mAbs used in antigen-capture ELISA were produced from hybridoma cell lines, RAC18 (anti-A chain) and RBC11 (anti-B chain), created in our laboratory [18]. Cells were grown in RPMI1640 (Gibco, Billings, MT, USA) and 5% Ig-depleted FCS (HyClone, Logan, UT, USA) and purified on Protein A agarose (Sigma). RAC18 was conjugated to biotin with an EZ-Link Sulfo-NHS-LC-kiut (Thermo-Fisher). Streptavidin conjugated to horse radish peroxidase was obtained from GE Healthcare (Chicago, IL, USA). Abs for the detection of insulin and glucagon by immunostaining were rabbit Abs (Cell Signaling Technology, Beverly, MA, USA, mAb anti-insulin, polyclonal Ab anti-glucagon) and were detected with alkaline phosphatase-conjugated goat anti-rabbit IgG (Thermo-Fisher, Waltham, MA, USA). Binding was visualized using aminoethyl carbazole (Sigma). The kits used to detect specific biochemistry include: the glucagon-HS ELISA kit (Cosmo Bio USA, Carlsbad, CA, USA), L-lactate assay kit I (Eton Bioscience, San Diego, CA, USA), a colorimetric enzyme assay based on the reduction of tetrazolium salt in a NADH-coupled enzyme reaction to formazan, ketone body assay kit (Sigma), which measures 3-hydroxybutyric acid and acetoacetic acid using an enzymatic assay based on 3-hydroxybutyrate dehydrogenase catalyzed reactions, mouse ultrasensitive insulin ELISA kit (ALPCO, Salem, NH, USA), mouse G6Pc (Glucose-6-Phosphatase, Catalytic) ELISA kit (G-Bioscience, St. Louis, MO, USA), glucose-6-phosphate assay kit (Sigma Aldrich, St. Louis, MO, USA), and glycogen assay kit (Sigma Aldrich). The kits were employed following the manufacturers’ instructions and values measured relative to the standards enclosed with the kits. 4.3. Blood Glucose

Blood glucose was measured with a handheld glucometer (Accuchek Advantage, Roche Diagnostics Corp. (Indianapolis, IN, USA), and AgaMatrix. Inc. (Salem, NH, USA)). Blood (25 to 50 μL) was loaded onto the appropriate test strip for each glucometer and read 15 s later. Blood glucose was always measured prior to the experimental procedure and immediately prior to sacrifice. In addition, repeated measurements were made on 68 mice, between 1 and 48 h post ricin injection.

4.4. Measurement of Ricin in TissuesAn antigen capture ELISA was used to measure ricin holotoxin in the tissues [18,39]. Tissue samples were obtained at necropsy and immediately frozen. The samples were weighed and mixed with a 5× volume of ice-cold PBS 0.1% Triton X-100 (Sigma). The samples were homogenized for 60 s with a handheld tissue homogenizer, while on ice. Following sedimentation, the supernatants were tested for ricin. The wells of a 96-well round-bottom Immulon 2HB plate (Thermo Scientific) were coated with the mAb RBC-11, then blocked with milk buffer (PBS, non-fat dry milk, Tween 20, 10% Thimersol) for >1 d. The plates were rinsed and incubated with supernatants of tissue homogenates or ricin standards overnight, washed, and incubated with biotinylated mAb RAC 18. Following an overnight incubation, the plates were washed and avidin-conjugated HRP was added. Forty-five min later, the plates were washed, and binding was detected by incubation with tetramethylbenzidine (Sigma) for 15 min, then stopping the process with 1 M sulfuric acid, and reading at 450 nm on an automated plate reader (BioTek, Winooski, VT, USA). 4.5. Flow Cytometry

Single cell suspensions of splenocytes were prepared and stained immediately post necropsy. The spleens were minced, passed through a nylon mesh (Falcon), and RBCs were removed with RBC lysis buffer (Sigma). The cells were resuspended in PBS/1%BSA/Azide. The cells (5 × 105) were stained in a total volume of 100 µL with an equal mixture of the following Abs: PE-Cy7-anti-CD45R/B220, PE-anti-CD3e, and APC anti-CD14. The cells were stained for 1 h on ice, washed with PBS, and fixed in PBS/2% paraformaldehyde. Analyses were performed on 10,000 cells gated by FSC and SSC using LSRII (BD Biosciences, San Jose, CA, USA). The data were analyzed with FlowJo software. A compensation algorithm was set using single stained cells and fluorescence-matched isotype controls (eBiosciences).

4.6. Multiplex Analyses

The plasma obtained from the mice at the time of necropsy was tested for a panel of cytokines, mediators, and hormones using Milliplex Murine Map kits (MilliporeSigma, Burlington, MA, USA). Following the manufacturer’s instructions, recommended dilutions of plasma were mixed with multiplex beads. Beads and plasma were vortexed and centrifuged for 3 min. The following panels of multiplexed beads were used: bone1A (MBN1A-41K), cytokine/chemokine I (MPXMCYTO70KPMX), soluble cytokine receptor (MSCR-42K-PMX), IGF binding protein (MIGFBP-43K), cardiovascular disease I (MCVD1-77AK), cytokine/chemokine II (MPXMCYP2-PMX12), cytokine/chemokine panel III (MPXMCYP3-PMX6), CRP Single Plex (MCVD77K1CRP), IGF-1 single plex (RMIGF187K), endocrine (MENDO-75K), gut hormone (MGT-78K), and TGFß 1,2,3 (TGFB-64K-03). Each sample was run in duplicate. The samples were then assayed using a Bioplex 200 system (BioRad, Hercules, CA, USA). The concentrations of ligand were determined with the standards included in the bead sets.

4.7. Glucose Tolerance Test (GTT)

We used a GTT to determine the ability of ricin-treated mice to clear glucose from the blood. A control group contained 6 mice and was injected with saline 24 h prior to GTT. Experimental groups of 4 mice each received ricin (30 µg/kg) 4 h, 6 h or 24 h prior to GTT. All animals were fasted 6 h prior to the beginning of the GTT, and water was allowed. Blood glucose was measured immediately prior to the start of the GTT. The mice were injected ip with 2 g glucose/kg body weight. Blood sugar was measured, using a glucometer as described previously, 15 m, 30 m, 60 m, and 120 m following glucose injection. Because blood sugar dropped with time post ricin injection, the area under the curve was calculated using a baseline normalized to blood sugar immediately prior to ricin injection.

4.8. ImmunoblottingSignaling by insulin was examined in the liver by immunoblotting as described elsewhere [55]. The liver was homogenized in cold lysis buffer (1% Nonidet P-40, 50 mM Hepes, pH 7.6, 250 mM NaCl, 10% glycerol, 1 mM EDTA, 20 mM β-glycerophosphate, 1 mM sodium orthovanadate, 1 mM sodium metabisulfite, 1 mM benzamidine hydrochloride, 10 μg/mL leupeptin, 20 μg/mL aprotinin, 1 mM phenylmethylsulfonyl fluoride). The supernatant was used for immunoblotting after centrifugation at 10,000× g for 10 min at 4 °C. Total protein (100 μg) in 50 μL of reducing sample buffer was used per lane. The product was resolved in 7% SDS-PAGE and transferred onto a polyvinylidene difluoride membrane for immunoblotting. Blotting of the membrane was conducted in a milk buffer with the first Ab for 1–24 h and HRP-conjugated secondary Ab for 30 m after. We used matched pairs of Abs recognizing the phosphorylated and total amounts of each protein: IRS-1, AKT, and GSK-3β. The phosphorylated signal and total signal strength were quantified with NIH Image J, and the ratio was calculated. The signals of treated samples are expressed as fold change over the untreated sample. 4.9. STZ-Induced DiabetesSTZ was used to induce diabetes as described elsewhere [44]. Prior to each STZ treatment, food was removed at 6 a.m. Four hours later, a fasting blood sugar was obtained, and STZ (40 mg/kg) was injected ip and the food was replaced. The mice received 5 daily doses of STZ. Fasting blood glucose was measured daily. When the blood glucose level was >250 mg/dL for 3 consecutive days, the mice were treated with ricin or saline. In the first experiment, three STZ-treated mice received ricin (30 or 15 µg/kg). The second experiment used 12 mice in 4 groups of 3. Two groups received STZ, two did not. The mice were injected with ricin (7.5 or 2.5 µg/kg). Blood glucose was measured twice daily post ricin. The mice were euthanized according to predetermined clinical criteria as outlined in the IACUC protocol and described above. 4.10. Microscopic Analyses

The liver and pancreas were fixed in formalin, embedded in paraffin, sectioned and deparaffinized. For pathologic examination, the tissues were stained with H&E. IHC was performed with rabbit anti-insulin or anti-glucagon primary Abs, alkaline phosphatase-conjugated goat anti-rabbit IgG (Invitrogen, Waltham, MA, USA), and aminoethyl carbazole substrate (Sigma). IHC slides were counterstained with hematoxylin. Images were visualized using the built-in widefield capabilities of a Zeiss LSM 510 microscope and analyzed with Axiovision software (Zeiss, White Plains, NY, USA). Islets were identified by morphologic criteria using a 40× objective, a region of interest (ROI) drawn around the circumference of the islet, and measurement of the ROI’s area was calculated by the software.

4.11. Isolation of Islets from Intact PancreasPancreatic islets were isolated as described [67,68]. The mice were anesthetized with ketamine/xylazine. Collagenase (type 4, Worthington, 2 mg/mL) in Medium 199 (Gibco) was injected into the pancreatic duct in situ. The pancreas was immediately removed and placed in a glass screw cap vial on ice. The mouse was euthanized. Collagenase solution was added to each of the pancreases, which were placed in a shaking water bath at 37 °C for 15 min. RPMI 1640 medium with 10% FCS was added to stop the digestion, and the vial was placed on ice. The cells were washed centrifugally ×3 in RPMI 1640 + 10% FCS and resuspended in Ficoll-Paque (GE Healthcare) buffer. The cells were shaken to achieve suspension of pellets. RPMI1640 was layered on top of the Ficoll-Paque and centrifuged at 800× g for 20 min. The RPMI layer, interface, and the top of Ficoll-Paque layer were removed, washed, and placed in RPMI1640 + 10% FCS in petri dishes. With gentle swirling under a dissecting microscope, the islets were identified and collected. Isolated islets were saved in RNAlater (Qiagen, Hilden, Germany) or teased into single cell suspensions for immediate study. 4.12. Reverse Transcriptase, Quantitative PCR (RT-PCR)Relative changes in gene expression were quantified using real-time RT-PCR. The primers were designed to have similar annealing temperatures and are shown in Table 1. Total RNA was isolated from the liver or isolated pancreatic islets with an RNeasy mini kit (Qiagen) following DNase treatment. Quantitative real-time RT-PCRs were preformed using an iScript one-step RT-PCR kit with SYBR green (Bio-Rad, Hercules, CA, USA) on the iCycler MyiQ single-color real-time PCR detection system (Bio-Rad). A total of 10 ng of RNA was used per reaction and the transcript levels were normalized to GAPDH levels. The relative mRNA expression levels were calculated according to the comparative CT (∆∆CT) method as described by the manufacturer (User Bulletin #2, Applied Biosystems, Waltham, MA, USA). All of the reactions were performed in triplicate. 4.13. StatisticsAll of the statistics and graphs were generated using GraphPad Prism version 8.4.1 for macOS (GraphPad Software). The Shapiro–Wilk test and Brown–Forsythe test were performed on mice included in Figure 1 to validate the use of parametric tests throughout the manuscript. We have previously shown that blood glucose was the only parameter that significantly changed following ricin administration [21]. We reasoned that if the blood glucose data following ricin administration was parametric, the remaining data would be as well. Student’s t tests were performed when two groups were compared. When more than two groups were examined, one-way ANOVA with Dunnett’s multiple comparisons test was administered. The significance level was set at 0.05 for all statistical experiments and all of the groups were compared to the control mice. The mean ± SEM are shown in all graphical representations. If no error bars are visible, they are smaller than the symbol indicating the mean. Representative images are shown for tracing the islets of Langerhans, immunoblotting, and tissue staining experiments.