Radiation makes cells select the form of death dependent on external or internal exposure: apoptosis or pyroptosis

Internal radiation exposure from neutron-induced radioisotopes environmentally activated following atomic bombing or nuclear accidents should be considered for a complete picture of pathologic effects on survivors. Acute and localized high dose radiation exposure from hot particles taken into the body must induce cell death and severe damage to tissues, whether they are proliferating or not. However, very little the cellular and molecular mechanisms underlying this internal radiation pathology has been investigated. Male Wistar rats were internally exposed to 56MnO2 powder by inhalation. Small intestine samples were investigated by histological staining at acute phase (6 h, 3 days and 14 days) and late phase (2, 6 and 8 months) after the exposure. Histological location and chemical properties of the hot particles embedded in small intestinal tissues were analyzed by synchrotron radiation—X-ray fluorescence—X-ray absorption near-edge structure (SR–XRF–XANES). Hot particles located in the intestinal cavity were identified as accumulations of Mn and iron. Pathological changes showed evidence of crypt shortening, massive cell death at the position of stem cell zone, including apoptosis and pyroptosis from 6 h through 8 months in the internal exposed rats.


Results
Early event damage to small intestinal tissue for internal exposure is uncharacteristically rapid and severe at the stem cell position. The average radiation doses of each organ received in 56 Mn(4×) and 56 Mn(2×) groups were substantially almost same or lower than those received in 56 Mn(1×) group. The tissue average accumulated absorbed doses in the small intestine were 1.33 ± 0.17 Gy in 56 Mn(1×), while 1.48 Gy ± 0.37 in 56 Mn(4×) group and 0.58 Gy ± 0.15 in 56 Mn(2×) group 6 (Table 1, Supplementary Table S2) at the time the activity of 56 Mn (half-life 2.58 h) were almost totally attenuated (3 days, 14 days, 2 months, 6 months, 8 months). The doses of Hour 6 are 80% of the doses.
After 56 Mn(1×) internal exposure, apoptosis stained by TUNEL method was observed scattering in the small intestine of rats at the position of transit cell at 3 days and mainly at the position of stem cell at 14 days, no change was observed in the Mn stable group, the unexposed control group and the γ2Gy externally exposed group except that the expression of apoptosis was slightly increased at the position of transit cell in the externally exposed group on the 3rd day (Fig. 2).
Pyroptosis observed after 56 Mn(1×) internal exposure detected by absent in melanoma 2 (AIM2) staining prominently scattered at the positions of stem and transit cell in the small intestine, from top to bottom of the crypt, AIM2-positive cells were stacked together to form a mass at 3 days and in the same positions at 14 days. Aberrant crypt was formed and structural abnormality of crypt was observed at 14 days by H&E stain. AIM2-positive abnormal crypts (non-straight and distorted crypts) were formed as structural abnormalities. AIM2-positive abnormal crypts are not seen in the γ2Gy external exposure whereas the dose is about the same of the 56 Mn internal exposures on average (Fig. 3). The internally exposed groups had consistently higher levels of AIM2 expression compared to the externally exposed groups and unexposed control. AIM2 expression in 56 Mn(2×) exposed and 56 Mn(4×) exposed groups peaked at 6 h and 14 days post-exposure, respectively, whereas it increased slightly on 3 days and returned on 14 days in the externally exposed group. In the internally exposed group, AIM2 expression was significantly higher in the 56 Mn(4×) group than in the 56 Mn(2×) group at stem cellrich-locations from 3 days through 2 months after exposure. No change was observed in the Mn stable group and unexposed control group (Fig. 4C, Supplementary Table S1).
The histological changes in the small intestine of the γ2Gy externally exposed group showed slight increase in apoptosis at 6 h, an increase in mitosis at 3 days, slight increase in pyroptosis at 3 days and slight decrease in crypt length at 14 days. In contrast, in the 56 Mn(4×) internally exposed group, apoptosis increased prominently at 6 h and slightly increased at 3 days, increase of mitosis was similar to the γ2Gy externally exposed group at 3 days, pyroptosis increase from 6 h through 6 months in 56 Mn(4×) group (Figs. 4, Supplementary Table S1), and crypt length (ratio of crypt/villous) severely decreased from 6 h through 2 months, which was less pronounced in 56 Mn(2×) group (Table 1, Supplementary Table S2, Figs. 1B, 4D, Supplementary Table S1). At 14 days in 56 Mn(1×) group, a morphologically abnormal crypts with intense pyroptosis were found, where was formed a "non-straight and distorted crypts" structure clearly observed by H&E stain. Slight hyperemia was also observed in the interstitium (Fig. 3).
To compare the effect of internal and external radiation on radiation-induced injury in the rat small intestine, ratio of crypt vs. villous length per crypts were measured in samples taken from 0 to 8 months after exposure (Table 1, Supplementary Table S2). The ratio in internally exposed rats showed prominently decreases at 6 h (no lower, 87% lower), 3 days (72%, 71% lower) and 14 days (78%, 78% lower) in 56 Mn(2×) and 56 Mn(4×) groups, respectively, while the ratio in externally exposed γ2Gy rats slightly decreased at 14 days (89% lower). For groups internally exposed with 56 Mn, the crypts were significantly shorter (crypt shortening) than externally exposed or unexposed control groups after 6 h in 56 Mn(4×), 14 Table S1). AIM2 expression was observed at the position of stem cell and transit cell 2 months after 56 Mn(1×) internal exposure (Fig. 3), and higher level of pyroptosis, which persisted from 6 h onwards, was observed in activation level dependent manner for internally exposed groups of 56 Mn(2×) and 56 Mn(4×) (Fig. 4C, Supplementary Table S1), while no change was observed in the Mn stable group and the unexposed control group, and in the γ2Gy external exposure group, apoptosis and pyroptosis were slightly increased at 2 and 6 months, respectively (Fig. 4 From histological scored findings in Fig. 4, Supplementary Table S1, early effects of internal exposure foreshadowed the warning signs of late effects. High score levels for apoptosis, mitosis and pyroptosis, with prominent crypt shortening at 2× and 4× 56 Mn exposure at 6 h continued to 8 months after irradiation. There were no pathological findings continued in the γ2Gy group, except for a small decrease in crypt shortening (94, 90%) after 6 and 8 months (Table 1, Supplementary Table S2 Table S1).
From the corresponding serial sections stained with H&E, phosphorylated histone H2AX (γ-H2AX) and AIM2 immunohistochemistry, and TUNEL method, pyroptosis was clearly detected by H&E stain and AIM2 protein expression in 56 Mn(4×) group at 2 months. Apoptosis was not so obviously observed in the position of

SR-XRF analysis revealed that masses of manganese and iron located in the small intestine cavity.
Early elementary profile and histopathologic image were shown in small intestine tissue 6 h after 2× 56 Mn internal exposure. Sample2, Sample3 and Sample3-fine were identified as condensed accumulations of manganese and iron (Fig. 6A,B). From the elemental distribution, we estimated the elementary profile for Sample2, Sample3 and Sample3-fine. The focus of Sample3-fine is a condensed accumulation of manganese and iron. It was located in the small intestine cavity. The somewhat anomalous presence, extra-tissue luminal side of these inhaled particles may have originated from a larger mass of manganese and iron in the process of digestive excretion with food or other matter. Fe elements were extremely abundant in sample3-fine in the small intestine, as in shot1 in the lungs, located in the bronchiole cavity 6 .
XANES spectroscopy for the analysis of particles embedded in small intestine tissue samples (Sample2, Sample3 and Sample3-fine). Particles embedded in small intestinal tissue samples (Sam-ple2, Sample3 and Sample3-fine in Fig. 6A,B) were analyzed by XANES spectroscopy and found to be Mn; XANES spectroscopy shows the absorption energy position and intensity of Mn compounds, so these samples were metabolized Mn. Mn particles with a 4-valent chemical species as MnO 2 changed to 2-valent when deposited in small intestinal tissue, as in lung tissue 6 (Fig. 6C).

Discussion
In small intestine of rat in 56   www.nature.com/scientificreports/ moreover, prevalently resulted in an increase of apoptosis, mitosis and genomic instability such as DNA DSB at the position prominently including stem cells at 2 months. Their stem cells may initially increase their proliferation by delayed arrest (stem cell exhaustion) and re-proliferation (cancer predisposition). Our results indicated an increasing possibility of digestive system failure and morbidity from the internal exposure. At external doses of 12 Gy and above, the mortality rate of GI syndrome exceeds that of hematopoietic syndrome 1 . Radiation induces loss of intestinal crypts and disruption of the mucosal barrier. These changes cause abdominal pain, diarrhoea, nausea and vomiting and render the patient infectious. However, the cellular targets of GI syndrome and the mechanisms of radiation-induced cell death remain controversial.
We have already reported that (1) after 5 Gy, 50% lethal dose, whole body external radiation exposure, the mucosal length of the small intestine was 85% lower at 3 day but returned normal level at 7 day 18 and that (2) after 8 Gy, 100% lethal dose, whole body external radiation exposure, the crypt length of the small intestine was 63% lower at 3 day but returned normal level at 5 day 19 . These results of high dose; 5 Gy and 8 Gy external exposure experiments were similar to the pathological findings at 6 h, 3 days and 14 days (early) and 2 months (late) after 56 Mn internal exposure in our study, namely the crypt shortening (71% lower) at 3 day, but returned normal level at 6 months. Compared to 5 Gy external exposure, the 56 Mn internal exposure of 2× and 4× in our study resulted in similar or more severe late pathological findings, crypt shortening by delayed arrest (stem cell exhaustion) and AIM2 expression of indicator of pyroptosis.
Although AIM2 is an innate immune sensor 20,21 , Hu B et al. 22 found that AIM2-deficient mice were protected from lethality and intestinal damage caused by lethal doses of subtotal body irradiation (SBI). They showed that endogenous AIM2 formed nuclear punctures upon irradiation and considerable co-localization with γ-H2AXpositive foci was observed in the nucleus, suggesting that radiation can mobilize AIM2 to dsDNA cleavage sites.
DSBs pose a major threat to genetic integrity and consequently are a leading cause of chromosomal aberrations and cancer in cells 23,24 . Genomic instability results from dysregulation of cell cycle checkpoints or defects in DNA repair 25 and it has also been reported that ageing induced by oncogenes is part of the barrier to tumorigenesis imposed by the DNA damage checkpoint 26,27 .
In our experiment, apoptosis was not so obviously observed in the position of stem cell, DNA DSB by γ-H2AX expression, also senescence marker, clearly observed in the position of stem cell (stem cell exhaustion). Pyroptosis by AIM2 expression as inflammation marker, was observed in the wide range of positions of stem cell at 2 months after radiation (Fig. 5). Mechanistically, loss of clonogenic (stem/progenitor) cells in the crypts has been suggested to be responsible for radiation-induced intestinal damage 28 . These suggested that internal exposure potentiate senescence in stem cell and pro-inflammatory programmed cell death. That's geroconversion; it converts reversible arrest to irreversible senescence, which leads to hyper-secretory, hypertrophic and pro-inflammatory cellular phenotypes, hyperfunctions and malfunctions. On organismal level, geroconversion leads to age-related diseases and death 29,30 .  www.nature.com/scientificreports/ Early event damage to small intestinal tissue for the internal exposure is uncharacteristically rapid and severe compared with external radiation exposure. Whole crypt pyroptosis was not observed in the 60 Co external exposed group in our study, nor even at the significantly higher dose of 14.2 Gy in Hu et al. 22 . No increase of apoptosis confined to the position of transit cells in the rat small intestine and no decrease of apoptosis, mitosis and AIM2 expressions were observed with the internal exposure compared with external exposure, suggesting a unique radiation pathology.
Intestinal pyroptosis is characterized histologically by the destruction of small intestinal crypt with massive cell death 22 . The pyroptosis proposed by Cookson et al. 31 represents inflammatory programmed cell death, and activation of the AIM2 inflammasome in response to cytoplasmic DNA can induce massive caspase-1-dependent cell death and severe damage 20,32 .
Apoptosis and pyroptosis by AIM2 immunohistochemistry were observed in the small intestine of rats at the position of stem cells for both 1× and 4× 56 Mn internal exposure, with DNA DSB detected by γ-H2AX expression. Taken together, our data suggest that the AIM2 inflammasome mediated pyroptosis of clonogenic cells in the intestinal crypts plays a critical role in the internal radiation-induced GI syndrome. These are similar to models of apoptosis-28 and pyroptosis-dependent cell death in the small intestine 22,31 , as well as in high-dose ionizing radiation external exposures 20,22,32 . Our findings demonstrated that the mechanism of internal radiation injury in the small intestine involves apoptotic and pyroptotic damage at the stem cell position. Apoptosis and pyroptosis was clearly and persistently evident in the small intestinal crypt at the position of stem cell in the 56 Mn group ( 56 Mn(1×);1.33 Gy) until two months later after internal exposure. Moreover, on the 14th day after internal exposure, cytoplasmic AIM2-positive structural abnormal crypts (non-straight and distorted crypts) were formed in the 56 Mn group ( 56 Mn(1×);1.33 Gy), which are not seen in the external exposure (γ2Gy). Crypt structural abnormalities may be induced, triggered by persistent apoptosis and pyroptosis at the position of stem cell in the crypt of the small intestine (stem cell exhaustion), as seen after internal exposure by the 56 Mn group.
After 3 days, γ2Gy group (small intestinal cell turn-over rate) showed similar low AIM2 expression as the (1) inflammation; 56 Mn(2×) and 56 Mn(4×) (Fig. 4C, Supplementary Table S1), (2) tissue repair; roughly equivalent A previous study has shown that over 50% of tumors from patients with small bowel cancer have frameshift mutations in the gene encoding AIM2 34 . Our results may be related to the anti-tumor effects of AIM2, as the both 2× and 4× internal exposures resulted in high score levels for mitosis and pyroptosis, with prominent crypt shortening and crypt distortion at day 3 continued to 8 months. In contrast, there were no pathological findings for mitosis or pyroptosis continued to occur in the γ2Gy group. High levels of mitosis and pyroptosis triggered a unique pathology for internal exposure, such as crypt shortening or crypt distortion. Abnormal crypts are the earliest morphologically identifiable precancerous lesions in the human colon [35][36][37][38] . In the human disease ulcerative colitis (UC), apoptosis and shortened or abnormal crypts are observed; long-term cases of UC are at increased risk of developing colorectal cancer arising from chronically inflamed mucosa 39,40 .
The frequencies of cancer incidence in response to low doses of direct radiation from atomic bombs draw complex non-linear curves, which may reflect the effects of internal exposure 41 .
Among the sources of residual radiation, an understanding of residual radiation from neutron activated soil materials on the ground is particularly important in assessing the risks for people who may have moved to these cities soon after the explosion and inhaled radioactive dust [42][43][44] . Such people have been reported to suffer from various syndromes, such as gastrointestinal disorders, as well as acute radiation exposure 45 . The incidence rate of acute radiation disease including diarrhea has been reported to be 50%, and Sawada calculated the cumulative effective dose as 1.49 ± 0.38 Gy from the incidence of acute radiation disease for an entrant within 1 km of the hypocenter of Hiroshima on 6 August 1945, immediately after the explosion 5 . Sawada also described for survivors of the bombing beyond 1.5-1.7 km that the incidence of acute radiation disease by the internal exposure due to radioactive fallout are more severe than those from external exposure to primary radiation. Sasaki et al. showed that the average total body absorbed dose can be estimated from the rate of chromosomal aberrations. However, the relation between the estimated dose and distance from the hypocenter cannot be explained by primary external irradiation and it might suggest that some people more than 2.4 km away received radiation from sources other than the primary rays 46 . www.nature.com/scientificreports/ The radiation-induced diarrhea observed in A-bomb survivors, as described in Introduction, is consistent with the results of the present experiments. Internal exposure in the experiment involves localized high-dose exposure and severe radiation damage to the small intestine.
Chemical species of inhaled and deposited 56 Mn particles were identified (Sample2, Sample3 and Sample3fine) as Mn 2+ using XANES spectroscopy. A 56 Mn particle that had a valence of four chemical species as MnO 2 changed to a valence of two after deposition in small intestine tissue as well as lung tissue 6 . Sample3-fine, being composed largely of Fe and located in a small intestine cavity, is likely to be an ejected blood clot around Mn particle in the process of digestive excretion with food or other matter. In contrast to the standard elementary profile of small intestine, Sample3-fine include a large percentage of Fe likely owing to hemorrhage in the early event. Our findings point to the impact of highly localized early radiation effects of internal exposure and the pathologic chain of events initiated by them, whereas the focus of external radiopathology has primarily been on the long-term effects so exclusively.
Since local ultra-high doses from internal exposure are likely to have effects comparable to those of high LET radiation 6,11,[47][48][49] , which also gives a local ultra-high dose, interphase death, i.e. damaged cells undergoing apoptosis or pyroptosis, may contribute to the pathological progression caused by internal exposure. It is possible that tissue cells severely damaged by external exposure to 60 Co-γ radiation may turn over with new differentiated cells, as the stem cell damage is less severe 1,18,50,51 . The pathology may therefore gradually lighten. Implanted hot particles emit radiation of varying intensities, from low to very high doses, which can cause extremely serious damage to stem cells that are not affected by low doses from external sources. Stem cells affected by internal exposure resulted in potentially triggering serious pathologies, such as crypt shortening and abnormal crypts (non-straight and distorted crypt precancerous lesions). Tissue surrounding hot particles suffers a variety of radiation injuries resulting from various radiation doses of exposures mixed in close proximity.

Materials and methods
Chemicals and radiation. The size distribution of MnO 2 powder particles used in this experiment is same as that previously reported on Stepanenko et al. 14 . MnO 2 powder containing 56 Mn (T1/2 = 2.58 h) was produced by neutron activation of 100 mg 55 MnO 2 powder (Rare Metallic Co., Ltd., Japan) at the IVG.1 M ("Baikal-1") nuclear reactor 52 using a neutron of an irradiation time of 2000s and fluence of 4 × 10 14 n/cm 2 (1×), 4000 s (2×), 8000 s (4×). Briefly, the 100 mg of activated powder with 56 Mn activities of 2.74 × 10 8 Bq ( 56 Mn(1×)), 2× 2.74 × 10 8 Bq ( 56 Mn(2×)) and 4× 2.74 × 10 8 Bq ( 56 Mn(4×)) was sprayed pneumatically over rats located in an experimental box. γ2Gy group was externally exposed to 2.0 Gy of 60 Co-γ -ray at a dose rate of 2.6 Gy/min using a Teragam K2 unit (UJP Praha, Praha-Zbraslav, Czech Republic). Control group was unexposed. The initial specific activities (activity per mass) of neutron-activated MnO 2 powder were 4× higher in 56 Mn(4×) group and 2× higher in 56 Mn(2×) group. The average radiation doses were also estimated in the same way as Stepanenko et al. 14 . Animals and treatment. Ten-week-old male Wistar rats were purchased from Kazakh Scientific Center of Quarantine and Zoonotic Diseases, Almaty, Kazakhstan. They were housed in plastic cages under climatecontrolled conditions at 22 ± 2 °C with a relative humidity of 50% ± 10% and a constant day/night cycle (light 0.700-19.00 h). They were maintained with free access to basal diet and tap water. For the study of 8 months, 56 Mn(1×) rats (3 days; n = 4, 14 days; n = 4, 2 months; n = 4, 3 months; n = 4, 8 months; n = 8) were compared with a group of rats exposed to Mn-stable (not activated) (3 days; n = 4, 14 days; n = 4, 2 months; n = 4, 3 months; n = 4, Pathology. Small intestines, middle portion, were collected, dissected and fixed in 10% neutral buffered formaldehyde and embedded in paraffin. Sections of 4 μm thickness were prepared and stained with H&E. For pathologic examination of the small intestinal tissue, the number of apoptotic cells and mitotic cells per crypt were counted as in 10,11 . 20 longitudinal crypt sections per mouse were selected and counted. Sections were used for counting the number of mitotic figures with hematoxylin and H&E stain 18 . Ratio of crypt vs. villous length per crypts of the small intestine in rats were measured as in 18,19 , and were analyzed by Olympus cellSens Dimension using 5 images per sample. Briefly, villus length, defined as the length from the apex of the brush border to the base of the crypt in the small intestine, and crypt depth (along the long axis of the elliptical crypt) were measured using a 100× magnification stage micrometer. The lengths of more than five random villi or crypts were measured (Supplementary Table S2), and the measurements were averaged 18 .
The expression of AIM2 (bs-5986R, Bioss Antibodies) and γ-H2AX at (Ser139) (Cell Signaling Technology, Danvers, MA, USA) proteins in small intestinal tissues was assessed using immunohistochemistry. Briefly, paraffin sections were deparaffinized and pretreated with microwave heating for antigen retrieval in 0.01 mol/l citrate buffer (pH 6.0). The sections were reacted with 0.3% H 2 O 2 in deionized water for 10 min to inhibit endogenous peroxidase activity and incubated with antibodies overnight. After washing with PBS, the sections were incubated for 30 min using an LSAB-2 system-HRP for use on rat specimens (Dako). Antibody binding was visualized by