Ionizing radiation (IR) has complex toxic biological effects on human tissues as an inducer of DNA damage and a threat to genomic integrity, causing radiation injuries^1,2and fatal diseases, including acute radiation syndrome, cancers, non-cancer diseases in long-term. These consequences are classified into deterministic effects or tissue reactions and stochastic effects^3,4, intensifying the need to promote radioprotection technology. Radiation protectors are significant and indispensable especially in medicine. Protection of normal tissues surrounding the targeted parts in diagnostic radiology and radiotherapies is necessary and has been focused on since application of these technologies in clinics. On the other hand, with the development of manned spaceflight in recent decades, complex cosmic radiation environment turned out to be among the serious obstacles to human space exploration. IR induces DNA damage directly by generating DNA double-strand breaks (DSB) via interactions between charged particles and DNA molecules⁵. The other indirect way is through oxidative stress by producing free radicals and reactive oxygen species (ROS)^5,6,7, leading to teratogenesis or carcinogenesis^8,9. IR also causes chromosomal aberrations, alterations in cell cycle and apoptosis^10,11.

Only a few compounds or cytokines as radiation protectors are available currently including amifostine, palifermin, filgrastim, sargramostim and romiplostim, which have been already used in radio and chemotherapies or in the countermeasure medicines list for radiological and nuclear emergencies. However, their application has been largely limited by a lack of diversity, dissatisfactory availability, potential toxicity and uncertain mechanisms of action¹². There is still an urgent need to identify novel protectors to combat fatal damage caused by IR.

The creation and acquisition of new and effective medical protectors for radiation injuries will require profound basic research to expand the radiobiological knowledge, as well as translational research to begin the transformation of fundamental knowledge gained into product development. The R&D cycle and cost of the traditional paradigm of new drug discovery are increasing, so that bringing a new drug to market is becoming more difficult and risky of a huge investment. Drug reposition is the effective identification of new indications for established drugs. It offers advantages including lower risk of failure, less investment and shorter period^13,14. At first, most new uses of old drugs were discovered occasionally in clinical settings, and then high-throughput screens. Prospective candidates are more likely to be proposed by the combination of computation-based prediction and experimental verification. Accumulation of large-scale genomic and pharmacochemical data and development of computational technologies present a huge opportunity to identify potential relationships between known drugs and diseases via targets, phenotypes and chemical constructions. The rapidly increasing approved chemical and biomedical drugs extend the repository for repositioning drug discovery. DrugBank is a comprehensive database that integrates detailed information of more than 17,430 drugs¹⁵. ChEMBL is a manually curated database of bioactive molecules with drug-like properties¹⁶, offering integrated chemical, biological and genomic data to aid their efficient translation into drugs. Combining data on compounds, targets and phenotypes from these resources may provide more comprehensive understanding of mechanisms of drug action and revelatory framework for drug reposition. Systematic computing methods identifying and analyzing complex interplays among diseases, drugs and their targets show excellent performance¹⁷and promote the whole process of drug reposition to a large extent¹⁸. A reported drug repositioning model HeTDR is based on heterogeneous networks and text mining, which combines multi-network drug and disease characteristics from literatures to reveal the potential associations between approved drugs and off-label diseases¹⁹. The AdaDR approach works by deeply integrating interactions between node features and topological structures with adaptive graph convolution operation network²⁰. It performs better than multiple baselines for drug repositioning and provides additional exploration in revealing drug-disease associations. Approved tyrosine kinase inhibitory anti-carcinogens dasatinib and imatinib were shown to be effective on radioprotection by promoting DSB repair²¹. Even though, application of rational design for drug reposition to radiation protector development is rarely reported.

In this paper, a computational method based on a comparison of biological similarity to predict repositioning drugs for protection against IR is presented. Gene expression profiles of γ-ray irradiation and different kinds of established compounds were employed. Potential compounds that might reverse the biological effects of radiation could be indicated by negative enrichment scores (ES). Lenalidomide, an U.S. Food and Drug Administration (FDA)-approved medicine mainly for treating the multiple myeloma and anemia caused by certain types of myelodysplastic syndromes, was selected finally and verified to protect against γ-ray radiation effectively in vivo and in vitro. It mitigated immune system impairment and intestinal radiation toxicity, and facilitated recovery of bone marrow hematopoiesis in irradiated mice.

Prediction of potential radioprotectors and selection of promising candidates

The flow diagram is shown in Fig.1. The gene expression profiles describing biological effects of γ-ray irradiation on human lymphocytes were obtained from GSE44245 dataset in GEO. The gene expression signatures of compounds applied to hematopoietic lymphoid cell lines in LINCS including NOMO- 1, PL- 21, SKM- 1 and WSU-DLCL2 were selected and extracted, which matched the lymphocytes in GSE44245. They provided us with a total of 280 known compounds. Gene ID conversion into official Gene Symbol was carried out according to annotation files of platforms of GPL96 and GPL570 for LINCS and GSE44245, respectively.

The schematic workflow of identification and verification of potential repositioning radiation protectors. They were screened out from Library of Integrated Network‑based Cellular Signatures (LINCS) by similarity comparison of their gene expression signatures with those of irradiation experiments from Gene Expression Omnibus (GEO), and verified in vivo and in vitro. ES: Enrichment scores. IR: Ionizing radiation.

Full size image

Comparison of biological effects between γ-ray irradiation and different chemical compounds bridged radiation response and established drugs. Opposite effects meant possibility of reversion of radiation damage, which implied the potential of repositioning to radioprotection of the corresponding known drugs. Gene expression signatures were considered as snapshots depicting biological effects. They were compared across the 4 hematopoietic lymphoid cell lines in LINCS with R codes and ES between ‑1 and 1 were calculated, characterizing similarity of biological effects between γ-ray irradiation and different chemical perturbations of cells in hemopoietic system²². ES > 0 meant that biological effects of the drugs are similar to those of γ-ray irradiation. While negative ES for compounds indicated the opposite biological effects. Absolute values of ES quantized the results. The selected threshold was − 0.25 for all of the 4 cell lines to ensure that < 1/3 of compounds were screened in each case. The intersection of screened drugs in the 4 cell lines was determined. Only FDA-approved drugs were considered for their accessibility and favorable safety profiles.

Menadione, lenalidomide and pimozide were 3 identified compounds in NOMO- 1 cells. The identified compound numbers were 9, 12 and 5 in PL- 21, SKM- 1 and WSU-DLCL2 cell lines, respectively (Supplementary Table 1). Predicting results in the 4 cell lines were integrated and a total of 7 compounds identified in more than 2 (≥ 2) cell lines were selected (Table1). We found that the only drug identified in all of the 4 cell lines was menadione. losartan was found in PL- 21, SKM- 1 and WSU-DLCL2 cell lines. Drugs identified in 2 cell lines included lenalidomide, rosiglitazone, pimozide, nimodipine and fluspirilene. Comprehensive investigation into physicochemical and pharmacological properties, mechanisms of action, and application of these drugs was conducted. We found that menadione and pimozide have been reported to be radiosensitizing by causing DNA damages^23,24. Losartan is an angiotensin receptor blocker used to treat hypertension and diabetic nephropathy, and reduces risk of stroke²⁵. It has been reported to attenuate IR-induced testis and lung injury^26,27. Lenalidomide is an immunomodulatory drug with potent anti-neoplastic, anti-angiogenic and anti-inflammatory properties. It inhibits NF-κB and MAPK related pathways. Lenalidomide has been approved by FDA and European Union for treatment of multiple myeloma, myelodysplastic syndromes, mantle cell lymphoma, follicular lymphoma and marginal zone lymphoma²⁸. We finally selected losartan and lenalidomide as candidates for the further validation in vivo.

Enrichment analyses were carried out on known targets of the 7 predicted repositioning drugs, aimed to explore their possible modes of action. The targets were retrieved from DrugBank in the form of Uniprot ID and converted into Entrez ID. Redundant entries were removed and 69 targets were obtained (Supplementary Table 2). Their GO enrichment analyses were conducted by R codes from the view of biological process (BP), cellular component (CC) and molecular function (MF). Their pathway enrichment analyses were also performed by R codes according to KEGG²⁹ (Fig.2). The enriched categories with adjustedP < 0.05 were shown in Supplementary Table 3.

The enrichment results of BP demonstrated that the most enriched categories of these targets were associated with xenobiotic metabolism and response, estrogen and steroid metabolic processes with the counts ≥ 11. For CC, 2 categories related to ion channel complex and transmembrane transporter complex were most enriched with the counts of 13. These known targets tended to enrich in 3 categories about binding activity from the view of MF with the counts of 15. The enriched category of steroid hydroxylase activity was ranked second according to adjustedP-value with the count of 10 (Fig.2A). In the KEGG pathway enrichment results shown in Fig.2B kinds of pathways were noteworthy. One was related to chemical carcinogenesis with 3 enriched categories and the other was steroid synthesis-relevant with 2 enriched categories.

Enrichment analyses of known targets of the 7 predicted repositioning radioprotectors. (A) Enriched categories of known targets from the view of biological process (BP), cellular component (CC) and molecular function (MF) according to Gene Ontology (GO). Top-ranked 8 categories for each were showed according to adjustedP-values. Categories with most counts in BP were response to xenobiotic stimulus (GO: 0009410, 33), cellular response to xenobiotic stimulus (GO: 0071466, 26) and xenobiotic metabolic process (GO: 0006805, 25). The most enriched categories in CC were cation channel complex (GO: 0034703), monoatomic ion channel complex (GO: 0034702) and transmembrane transporter complex (GO: 1902495) with the counts of 13. In the case of MF, the most enriched categories were heme binding (GO: 0020037), iron ion binding (GO: 0005506) and tetrapyrrole binding (GO: 0046906) with the counts of 15. (B) Enriched signaling pathways of known targets according to Kyoto Encyclopedia of Genes and Genomes (KEGG). Top-ranked 20 pathways were showed according to adjustedP-values. The most enriched pathways were chemical carcinogenesis-receptor activation (hsa05207, 20), chemical carcinogenesis-DNA adducts (hsa05204, 17) and drug metabolism-cytochrome P450 (hsa00982, 16).

Full size image

Figure3 shows the results of typical in vivo survival study of lenalidomide and losartan. Mice in experimental groups of lethal 9 Gy of TBI were administered with 50 mg kg⁻¹ (lenalidomide-L), 100 mg kg⁻¹ (lenalidomide-H) of lenalidomide or 40 mg kg⁻¹ (losartan-L), 80 mg kg⁻¹ (losartan-H) of losartan prior to irradiation (n = 9). We found that 3/9 and 1/9 of mice in the groups of lenalidomide-H and losartan-H survived up to day 30 respectively besides the normal control and positive control groups, implying their protective effects against the lethal dose of IR. More mice survived in the lenalidomide-H group than losartan-H. All mice in the irradiated control group untreated with any drugs succumbed within 13 days after irradiation (Fig.3A). In the case of 8 Gy of TBI (n = 10), both high and low concentrations of lenalidomide and losartan delayed death of mice and enhanced their survival rates at all time points till the end of the observation compared with the irradiated control group. It’s worth noting that as many as 90% of the mice in the lenalidomide-H group survived to day 30, which was the same with the positive control group (Fig.3B).

The effects of lenalidomide and losartan on body weights of mice after TBI are shown in Fig.3C and D. In the case of 9 Gy of TBI, the body weights of mice in all experimental groups decreased during the first 10 days after irradiation, and began to recover from then on. On day 30 the body weights of mice in lenalidomide-H group were similar to those of the positive control group, while those in losartan-H group were even better (Fig.3C). The body weights of mice decreased in all experimental groups after 8 Gy of TBI during the first week after irradiation. In the both lenalidomide pretreatment groups, the body weights kept higher than those in the irradiated control group since day 7. The body weights of mice in the lenalidomide-H group after day 13 were even higher than those in the positive control group (Fig.3D). As a result, it was suggested that 100 mg kg⁻¹ of lenalidomide was more effective for radioprotection. Thus, it was selected and would be used in the subsequent experiments.

Effects of losartan and lenalidomide on survival and body weights of irradiated mice. Mice were injected with losartan or lenalidomide 2 h before exposure to total body irradiation (TBI). Experimental groups: lenalidomide-L (50 mg kg⁻¹), lenalidomide-H (100 mg kg⁻¹), losartan-L (40 mg kg⁻¹), losartan-H (80 mg kg⁻¹). (A andB) Survival probability of mice after (A) 9 Gy and (B) 8 Gy of irradiation. (C andD) Body weights of surviving mice after (C) 9 Gy and (D) 8 Gy of irradiation. Amifostine is the positive control. NC refers to normal control.

Full size image

Figure4 shows the effects of 100 mg kg⁻¹ of lenalidomide on peripheral hemogram of mice after 6.5 Gy of irradiation. The lowest WBC counts of mice were observed on day 10 after irradiation for both irradiated control and lenalidomide groups. Their WBC counts began to recover since that day for the both groups. It was obvious that WBC counts of the lenalidomide-treated mice were higher than those of the irradiated control group throughout the observation. The counts of lenalidomide group were about 1.5 folds those of the irradiated control group (Fig.4A). The RBC counts and HGB content of mice in both groups decreased during the first 10 days after irradiation. Recovery in the lenalidomide groups started from then on while that of the irradiated control groups started on day 14. RBC counts and HGB of the lenalidomide groups were higher than those of the irradiated control groups from day 4 (Fig.4B and C). PLT counts of the lenalidomide group also stayed above those of the irradiated control group since day 4 although fluctuation was observed during the first week after irradiation (Fig.4D).

To further explore the radioprotective mechanism of lenalidomide from its effects on recovery of hematopoietic function, mouse pathological changes in bone marrow were evaluated on day 10 and 14 after 6.5 Gy of TBI, respectively (Fig.4E). The examination of longitudinal sections of the femur of mice after irradiation showed a marked increase in adipose tissue and cell debris, indicating severe bone marrow cell failure induced by IR. We found obvious microfocal regeneration of hematopoietic cells and notable recovery of megakaryocytes in femoral sections of lenalidomide-treated mice after irradiation, suggesting substantial protection and restoration of lenalidomide on hematopoietic function of bone marrow.

Effects of lenalidomide on hematopoiesis of irradiated mice by γ-ray (n = 8). Mice were pretreated with 100 mg kg⁻¹ of lenalidomide 2 h before 6.5 Gy of irradiation. (A-D) Peripheral blood cell counts of (A) white blood cells (WBC), (B) red blood cells (RBC), (C) hemoglobin (HGB) and (D) platelets (PLT) of mice after total body irradiation (TBI). (E) Hematoxylin and eosin (H&E) staining of femur of mice on day 10 and 14 after irradiation. The area of tissue lesion was indicated by arrows.^*P < 0.05,^**P < 0.01,^***P < 0.005.

Full size image

Gastrointestinal tract is sensitive to IR. Intestinal injury of mice induced by 6.5 Gy of TBI and protective effects of lenalidomide were observed on day 3.5 and 10 after irradiation, respectively (Fig.5). Intestinal tract of mice in the vehicle group was almost destroyed (Fig.5A). The villi became shorter and were largely exfoliated with severe inflammatory infiltration, indicating small intestinal injury. Small intestinal villi of lenalidomide-treated mice were obviously longer and more complete than that of the vehicle group (Fig.5B). Inflammatory infiltration was also alleviated. Compared to the vehicle, intestinal structures of the mice in lenalidomide group were more intact, with longer villi and more viable crypts (Fig.5C).

Protective effects of lenalidomide against intestinal injury induced by 6.5 Gy of irradiation. Hematoxylin and eosin (H&E) staining of small intestine treated with ionizing radiation (IR) and 100 mg kg⁻¹ of lenalidomide. The areas of tissue lesion were indicated by arrows. Quantitative Measurement of Villus Length (B) and Crypt Depth in the Mouse Small Intestine (C).^*P < 0.05,^**P < 0.01,^***P < 0.005,^****P < 0.001.

Full size image

Lenalidomide enhances cell viability after irradiation

AHH- 1 and IEC- 6 cells were treated with gradient concentrations of lenalidomide from 1 to 1000 µM. The cell survival at 24 and 48 h after drug administration are presented in Fig.6A and B. We found that survival rates of the both cell lines were greater than 80% when concentrations of lenalidomide were lower than 400 µM, which determined non-cytotoxic concentrations of the drug. The 2 kinds of cells were treated with 200 µM of lenalidomide. Its Effects on survival of AHH- 1 and IEC- 6 cells after 1, 2, 4 and 6 Gy of irradiation were determined by CCK- 8. Then, We plotted the radiation dose-response curve. It was shown that 200 µM of lenalidomide enhanced survival rates of AHH- 1 cells after 1, 2, 4 and 6 Gy irradiation respectively (Fig.6C). IEC- 6 cells showed similar tendency with AHH- 1 cells (Fig.6D). So 200 µM was selected and would be used in the following studies.

Colony formation assays were conducted in IEC- 6 cells irradiated by gradient doses of γ-ray of 2, 4, 6 and 8 Gy. The results showed that 200 µM of lenalidomide was non-toxic for the cells. Cell survival in both irradiated control and lenalidomide groups decreased with respect to doses of irradiation. Compared to the irradiated control group, proliferation of IEC- 6 cells was evidently improved by 200 µM of lenalidomide for all of the doses examined (Fig.6E).

Lenalidomide promotes survival of cells after irradiation by γ-ray. (A andB) Effects of lenalidomide on growth of (A) AHH- 1 and (B) IEC- 6 cells determined by CCK‑8 assays. (C andD) Protective effects of lenalidomide in (C) AHH- 1 and (D) IEC- 6 cells against 1, 2, 4 and 6 Gy of irradiation. (E) Colony formation assays demonstrated that 200 µM of lenalidomide enhanced survival of IEC- 6 cells after irradiation. The line chart showed colony formation of IEC- 6 cells irradiated by different doses of γ-ray.^*P < 0.05,^**P < 0.01,^***P < 0.005,^****P < 0.001.

Full size image

Rapid accumulation of data and advances in technology promoted drug R&D in recent decades, especially data of multi-omics and large-scale computational approaches. The ever-deepening insight into detailed mechanisms of DNA damage response and repair assists in discovery of radiation protectors. Rational identification of known compounds for their potential anti-radiation effects provides a strategy as timesaving and cost-effective innovation. Improved radioprotective technology may facilitate development and progress in many areas, such as rescue of nuclear accidents, exploration into deep space, extended application of radiological imaging and radiotherapy in medicine. Moreover, active development of regulators of radiation response including radioprotectors and radiosensitizers would further drive our understanding of DNA damage-related bioprocesses. This study constructed a workflow to predict and verify repositioning radiation protectors, and proposed a prospective candidate lenalidomide for further studies.

Lenalidomide is a thalidomide derivative approved by FDA in 2005 mainly for treatment of multiple myeloma and anemia in low-to-moderate myelodysplastic syndromes with the known targets TNF, CRBN, TNFSF11, CDH5 and PTGS2³⁰. Lenalidomide is more effective than thalidomide with less adverse reaction and toxicity. It is easily accessible for its biological tolerance and bioavailability have been verified during application of about 20 years. A known mechanism of lenalidomide is regulation of immune response, which is consist with the peripheral hemogram of mice in this study. Lenalidomide enhances NK cell-mediated cytotoxicity, and inhibits production of pro-inflammatory cytokines like TNF-α, IL- 1, IL- 6 and promotes production of anti-inflammatory cytokines like IL- 10 and IFN-γ, which regulate T cell function.

Our results showed that one of the possible mechanisms of radioprotection might be associated with anti-tumor for the known targets of predicted drugs tended to enrich in chemical carcinogenesis-related pathways. In addition, steroid hormone metabolism-relevant bioprocesses and pathways were also considered critical according to our enrichment analyses. The protective role against IR of vitamin D as a kind of secosteroids has been reported for its anti-oxidant property³¹. The anti-inflammatory corticosteroid dexamethasone has been proved effective in treatment of IR-induced ototoxicity via regulation of mTOR signaling pathway³². The other notable enriched category for BP was metabolic process of estrogen, which turned out to be potential in neuroprotection against IR-induced brain injury³³. Estrogen treatment after irradiation was also reported to reduce incidence and progression of cataracts³⁴. Single IP injection of lenalidomide before IR represents an attractive prophylactic and mitigating candidate for human use, which prevents IR-induced injury, providing a convenient and reasonable precaution. We found that lenalidomide was especially effective against 8 Gy of irradiation as its performance was closer to that of the positive control. Its performance in maintenance and restoration of body weights of mice after 8 Gy of irradiation was even better than that of the positive control. In addition to the lethal doses of irradiation, lenalidomide might also be promising in preventing medium or low doses of IR so as to be used probably in protection of occupational exposure. Drastically decreased WBC counts of mice in the first 24 h after irradiation demonstrated IR-induced immunosuppression. Pretreatment of lenalidomide may alleviate the injury by promoting immunological recovery. Lenalidomide has also been reported as an inducer of ubiquitination of casein kinase 1 A1 by an E3 ubiquitin ligase³⁵. The famous metformin is effective in alleviating IR-induced senescence and cardiotoxicity by modulating BRCA1-BARD1-RAD51 complex in ubiquitination³⁶. RFWD3-dependent ubiquitination on Rad51 even determines the radio-bidirectional regulatory effects of valproate³⁷. So it was suggested that post-translational modifications, especially ubiquitination and SUMOylation³⁸ might be crucial in drug action against IR of lenalidomide and some other compounds. They are likely to be a component of the mechanistic basis for regulation of radiation response.

Due to the many limitations of traditional drug discovery paradigm, drug reposition based on computational methods and rational design has been proved to be promising and efficient³⁹. Nevertheless, the current predictive methods and models are in need of further optimization to enhance the accuracy and reliability. Progression in multi-omics, including transcriptomics, epigenomics, metabolomics and signalosomes will provide a comprehensive landscape of integrated knowledge from a global view. Another revolutionary strategy is artificial intelligence (AI), which will play an important role in construction of novel predictive models and frameworks using approaches like machine learning and neural network. They will bring a wave of innovation in drug discovery and evaluation. R&D of radiation protectors will surely benefit from this trend in pharmaceutical industry.

In the future work, the precise radioprotective mechanism of lenalidomide needs to be further investigated by biological experiments in vivo and in vitro. The increase of apoptosis by Hesperidin administration can be attributed to the decreased expression of bax and significantly reduced expression of bcl- 2 and finally increasing the ration of bax/bcl- 2⁴⁰. We evaluated the effects of lenalidomide on post-irradiation apoptosis levels in AHH- 1 and IEC- 6 cells (FigureS1) and are currently investigating its radioprotective mechanisms. More details of lenalidomide on gastrointestinal tract and immune organs like spleen and thymus are still unclear. Its effects on other radiosensitive tissues including lung and angiocarpy should also be elucidated. Another critical consideration is the combinations of multiple drugs or diverse functional targets for improved therapeutic effects and reduced side effects. Drug interactions such as synergism and antagonism must be taken into consideration since they add complication and unpredictability to therapies, which have to be carefully evaluated and balanced. In future treatment options, it is believed that the appropriate combinations of multi-drugs or targets may provide potent radioprotective therapeutic regimens at lower doses.

In summary, our work proposed an efficient and economical workflow for precise identification of alternative repositioning compounds to combat IR based on computational comparison of similarity between biological effects. We found that lenalidomide represented a prospective radiation protector by inhibiting apoptosis, and improving immunoregulation and gastrointestinal reaction. Besides interpretation and revelation of molecular mechanisms on IR-induced DNA damage response and repair, the emerging multi-omics and AI technologies will accelerate discovery and development of more efficient repositioning radiation protectors.

Library of integrated network‑based cellular signatures (LINCS)

The cellular expression signatures of all compounds used in this study were collected and extracted from LINCS of National Institutes of Health (https://commonfund.nih.gov/LINCS). Currently the database contains gene expression profiles in response to pharmacological perturbations of 42,794 small molecules applied to 1175 cell lines at different time points and concentrations. Data of each disturbance precursor treatment and its corresponding control were downloaded from the portal of LINCS. A total of 55,129 gene expression profiles of 4617 chemicals and 4388 profiles of their controls were screened and selected, providing candidates for prediction of potential radioprotectors in this study.

Gene expression omnibus (GEO)

The primordial gene expression profile dataset GSE44245 used in this study was collected from GEO database (https://www.ncbi.nlm.nih.gov/geo/), containing 4348 datasets, 243,983 series and 26,892 platforms. The data in GSE44245 were obtained from 3 experimental groups of human lymphocytes irradiated by 5 Gy of γ-ray and their 3 control groups. They were detected by microarrays 24 h after irradiation. Platform GPL570 was employed so that the data can match those in LINCS. We downloaded Series Matrix Files of GSE44245 in “.txt” and obtained detailed gene expression profiles of γ-ray irradiated human lymphocytes.

Enrichment analysis

We performed enrichment analyses of known targets for predicted repositioning drugs. Gene ID conversion from Uniprot ID into Entrez ID was conducted using the “bitr” function in db package (3.17.0) in R software (4.3.2). The clusterProfiler package (4.10.0) in R was used for Gene Ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses. Plotting was completed by ggplot2 package (3.4.4) in R.

Reagents

Dulbecco’s modified eagle medium (DMEM) culture medium, RPMI1640 medium, pancreatic enzyme and penicillin-streptomycin dual resistance solution were purchased from Shanghai BioScience Company. Fetal bovine serum was purchased from Shanghai ExCell Company and crystal violet from Beijing Solarbio Technology Co., LTD. Annexin V-FITC/PI apoptosis detection kits were purchased from Dojindo Institute of Chemistry, Japan and CCK- 8 kits from Meilunbio. Lenalidomide and losartan were from Selleck, USA. Dimethyl sulfoxide (DMSO) was purchased from German Sigma Company.

Animal model

C57BL/6 J male mice were purchased from SPF (Beijing) biotechnology Co. Ltd., and experiments were conducted after 1 week of stabilization. All mice were used at about 6–8 weeks of age, weighting 21–24 g. The animals were kept in a specific pathogen-free environment (temperature 22 ± 2 °C, humidity 55 ± 5%, light/dark cycle 12 h) and had free access to standard animal feed and acidic water. All experimental procedures were approved by Animal Care and Use Committee of the Zoological Center Institution of Academy of Military Medical Sciences (IACUC-DWZX-2024 - 585). The authors confirm that all animal experiments in the present study have been performed in accordance with ARRIVE guidelines. All methods were carried out in accordance with relevant guidelines and regulations.

Cell culture

IEC- 6 cell line was purchased from iCell Bioscience Inc. It was cultured in DMEM (1.5 g L⁻¹ containing NaHCO3) medium with 10% fetal bovine serum and 1% penicillin/streptomycin. AHH- 1 cells were cultured in RPMI1640 medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. The cells were cultured in a 5% CO₂ humidified incubator at 37 ℃. The medium was changed every 48 h.

Irradiation and drug administration

The mice were placed in a clear plastic box and exposed to a lethal dose (8–9 Gy) or sublethal dose (6.5 Gy) of total body irradiation (TBI) of γ-ray by a⁶⁰Co irradiator (Beijing Institute of Radiation Medicine) at an average dose rate of 60 cGy min⁻¹. For the survival tests and peripheral blood cell counts, irradiated mice were scrutinized on day 30 after TBI. Mice were given lenalidomide, losartan or vehicle of one intraperitoneal (IP) injection of solvent 2 h before irradiation. Amifostine was used in positive control groups and mice were given 150 mg kg⁻¹ of amifostine of one IP 30 min before irradiation.

The cells of irradiation groups were irradiated with⁶⁰Co γ-ray at doses of 2, 4, 6, and 8 Gy at room temperature, respectively.

Peripheral blood cell counts

Mouse tail veins peripheral blood (PB) was collected. The numbers of white blood cells (WBC), red blood cells (RBC), hemoglobin (HGB) and platelets (PLT) were counted using MEK- 730 hemocytometer (NIHON KOHDEN Corp, Japan).

Histopathology

Femurs of mice were fixed, decalcified and paraffin-embedded at day 10 or day 14 after 6.5 Gy of TBI. 5 μm sections were prepared for hematoxylin and eosin (H&E) staining. Mouse intestines were harvested at day 3.5 or day 10 after 6.5 Gy of TBI, fixed in 4% PFA overnight at room temperature and then embedded in paraffin. 5-µm sections were prepared and stained with H&E. The sections were observed under a light microscope and histological changes were recorded.

Cell viability assay

AHH- 1 and IEC- 6 cells of logarithmic growth stage (2000 cells per well) were inoculated in 96-well plates (6 multiple wells per group). Then the cells were cultured for 24 h. DMSO and lenalidomide were added respectively 2 h before irradiation. And 100 µL of the culture base containing 10% CCK- 8 reagent was changed and incubated for 2 h away from light. The absorbance value at 450 nm was detected with an enzyme marker. The maximum and minimum values were removed from each group and the mean values were taken.

Colony formation assay

IEC- 6 cells of logarithmic growth stage were selected. They were adjusted to an appropriate concentration and inoculated into 6-well plates. The numbers of cells corresponding to 0, 2, 4, 6 and 8 Gy of irradiation were 400, 400, 600, 800 and 1000 per well, respectively. The culture plates were placed in an incubator. The stock solution was removed 2 h before irradiation. Culture medium containing DMSO and lenalidomide was added respectively. Each group was set with 3 double holes. After different doses of irradiation, cell culture was continued for 7–14 days. It was terminated when the cell communities were visible with naked eyes after changing the fluid every 3–4 days. The cells were fixed with 4% paraformaldehyde for 60 min. The staining solution was washed off 30 min after crystal violet staining. The numbers of clones of ≥ 50 cells were counted.

Apoptosis assay

Cells of logarithmic growth stage were selected and inoculated in 12-well plates. Then the plates were cultured in an incubator for 24 h. The stock solution was discarded 2 h before 6 Gy of irradiation, and culture medium containing DMSO and lenalidomide was added. Cells were digested with pancreatic enzymes 24 and 48 h after irradiation, respectively. They were centrifuged at 4 ℃, 1000 rpm min⁻¹ for 3 min, washed 3 times with phosphate buffer saline at 4 ℃, and stained with Annexin V-FITC/PI kit. The 10× binding buffer in the kit was diluted to 1× with ddH₂O. 100 µL of 1× binding buffer was added to each group to resuspend cells. 3.5 µL FITC dye and 3.5 µL PI dye were added to mix well. The reaction was carried out for 20 min at room temperature away from light. The staining was terminated by adding 400 µL of 1× binding buffer. The apoptosis ratio was detected by Percp and FITC channels using flow cytometry. The proportions of FITC+/PI- (early apoptotic cells) and FITC+/PI+ (late apoptotic cells) cells were analyzed.

Statistics and reproducibility

A permutation test procedure based on empirical phenotype was employed to conduct statistical analyses of GEO data in computational prediction. Phenotype labels were permuted randomly and ES of the gene set in the gene list of fold changes for the permuted data were recomputed. We set the number of permutations to 1000.P‑values of ES were calculated according to this distribution, andP < 0.05 was considered to be statistically significant.

The sample sizes of experiments in vivo were more than 8 and experiments in vitro were conducted at least in triplicate. Data from these independent experiments are presented as mean ± standard deviation (s.d.).P-values were calculated using a two-way analysis of ANOVA. Differences between groups were analyzed using Student’s t‑test. Statistical analyses were conducted by graphpad 9 software.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

The authors declare that the data supporting the findings of this study are available within the paper and its supplementary information files. The identified compounds in 4 hematopoietic lymphoid cell lines and their enrichment scores (ES) are available in Supplementary Table (1) Targets of predicted repositioning radioprotectors are available in Supplementary Table (2) Enriched categories and pathways of targets of predicted repositioning radioprotectors are available in Supplementary Table 3.

D.X. discloses support for the research of this work from Beijing Natural Science Foundation [grant number 7232107] and the Young Elite Program [grant number AMMS-QNTB- 2022 - 001]. X.H. discloses support for the research of this work from the National Natural Science Foundation of China [grant number 82203982]. P.Z. discloses support for the research of this work from the National Natural Science Foundation of China [grant number 32171238 and 82230108].

Authors and Affiliations

1. Department of Preventive Medicine, School of Public Health, University of South China, 421001, Hengyang, Hunan, China

   Qi Huang, Bo Huang & Pingkun Zhou

2. National Center of Biomedical Analysis, 100039, Beijing, China

   Xiaoyao Yin

3. Department of Radiation Biology, Beijing Key Laboratory for Radiobiology (BKLRB), Beijing Institute of Radiation Medicine, 100850, Beijing, China

   Hua Guan, Xin Huang, Dafei Xie & Pingkun Zhou

Contributions

P.Z. conceptualized and designed the study. X.Y., H.G. X.H. and B.H. conducted the investigation. X.Y. and D.X. contributed to data integration, computation and statistical analyses. Q.H., H.G. and B.H. contributed to experimental studies. Q.H. and D.X. created the visualizations. D.X. and Q.H. drafted the original manuscript. P.Z. reviewed and edited the manuscript. P.Z. and B.H. supervised the project.

Corresponding authors

Correspondence toDafei Xie orPingkun Zhou.

Competing interests

The authors declare no competing interests.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visithttp://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions