Abstract
Human embryonic brain development is highly sensitive to ionizing radiation. However, detailed information on the mechanisms of this sensitivity is not available due to limited experimental data. In this study, differentiation of human embryonic stem cells (hESCs) to neural lineages was used as a model for early embryonic brain development to assess the effect of exposure to low (17 mGy) and high (572 mGy) doses of radiation on gene expression. Transcriptomes were assessed using RNA sequencing during neural differentiation at three time points in control and irradiated samples. The first time point was when the cells were still pluripotent (day 0), the second time point was during the stage of embryoid body formation (day 6), and the third and final time point was during the stage of neural rosette formation (day 10). Analysis of the transcriptomes revealed neurodifferentiation in both the control and irradiated cells. Low-dose irradiation did not result in changes in gene expression at any of the time points, whereas high-dose irradiation resulted in downregulation of some major neurodifferentiation markers on days 6 and 10. Gene ontology analysis showed that pathways related to nervous system development, neurogenesis and generation of neurons were among the most affected. Expression of such key regulators of neuronal development as NEUROG1, ARX, ASCL1, RFX4 and INSM1 was reduced more than twofold. In conclusion, exposure to a 17 mGy low dose of radiation was well tolerated by hESCs while exposure to 572 mGy significantly affected their genetic reprogramming into neuronal lineages.
INTRODUCTION
There continues to be a great deal of uncertainty and anxiety among the general population regarding the possible health effects from various types of ionizing radiation, including diagnostic imaging procedures such as computed tomography (CT) scans. In particular, the effects of low-dose radiation on fetal development from common diagnostic radiological studies such as CT and X-ray imaging procedures performed on women in the early stages of pregnancy are not well understood. It is generally accepted that a radiation exposure at the pre-implantation stage may result in either miscarriage or may be inconsequential (1–3). However, the molecular mechanisms that govern these outcomes, the effects of the dose and the stages of development during the exposure, are unclear.
Most of our knowledge of radiation-induced biological effects on early human development is predominately based on retrospective studies from atomic bomb (A-bomb) survivors exposed prenatally (1, 4–6). In those studies, it was found that embryo exposure resulted in fetal brain malformations such as microcephaly and growth retardation (6). Animal studies also showed that lower doses of radiation could cause defects in brain development (7–9). However, since neural development in mice differs considerably from development in humans, caution is warranted when extrapolating results from animal models to humans (10).
Further in vitro work could be performed using human organoid systems to investigate the effects of radiation on embryogenesis (11). Such three-dimensional (3D) cell culture models would allow for control of both the dose and time of radiation exposure, and thus, could potentially provide insights into the underlying molecular mechanisms.
In our previously published study (12), we used directed differentiation of pluripotent human embryonic stem cells (hESCs) into neuronal lineages as a model for early brain development. We found that exposure to the 17 mGy dose typically received by a single abdominal/pelvic CT scan did not significantly affect either the morphology of the resulting neural progenitors or the expression of neurodevelopment markers. However, exposure to 572 mGy radiation resulted in a nearly twofold decrease in the expression of early neuronal differentiation marker paired box protein (PAX6). Herein, we extended our studies to a whole transcriptome analysis of gene expression on early stages of differentiation to neuronal lineages after exposure to CT scans.
MATERIALS AND METHODS
Differentiation of hESCs into Neural Progenitor Cells
H9 hESCs (passage 21; WiCell® Research Institute, Madison, WI) were cultured in feeder-free complete mTeSR-1 culture medium (STEMCELL™ Technologies Inc., Vancouver, Canada) in Matrigelt® (Corning®, Corning, NY) coated flasks. Cells were passaged every 5 to 7 days with collagenase IV (STEMCELL Technologies). H9 hESCs were seeded onto T-75 Matrigel-coated flasks 48 h prior to irradiation. After irradiation and 24-h recovery, hESCs were directed to differentiate into neuronal lineages. We used the STEMdiff™ Neural Induction protocol, with reagents and supplies from STEMCELL Technologies as was described in detail elsewhere (12). Briefly, the colonies of hESCs were dissociated into a single cell suspension and seeded onto AggreWelle 800 plates in 2 ml of Neural Induction Medium (NIM) supplemented with 10 μM of Y-27632 ROCK inhibitor. On day 5 of differentiation, embryoid bodies were harvested and plated onto Matrigel-coated six-well plates. Neural rosettes were harvested on day 9 of differentiation. The differentiation procedure was repeated three times, generating three biological replicates.
CT Scan Irradiation
Cell irradiations were performed using a single X-ray tube of a dual-source CT scanner (SOMATOM Definition Flash; Siemens Healthcare, Forchheim, Germany). Flasks with hESCs were placed into a 3D torso phantom (CIRS, Norfolk, VA). Cells in the low-dose group received a dose of 15 mGy CT dose index (CTDI), approximating the dose for abdominal/pelvic CT examinations in adults, delivered as a single axial scan at 120 kVp and a tube current of 200 mAs. As positive controls, cells in the high-dose group received a dose of 500 mGy CTDI, delivered as 10 consecutive axial scans at 120 kVp and a tube current of 666 mAs.
The CTDI values do not generally reflect the true absorbed dose to the samples placed in the gantry. The absorbed dose inside the phantom was measured using a calibrated, 10-cm-long ionization chamber (Victoreen® 660 with 660–6 probe; Fluke® Biomedical, Columbus, OH). The doses measured using the probe were calculated as 17 mGy and 572 mGy for low- and high-dose radiation, respectively. The 0 mGy control samples were mock-irradiated with the same scan procedure, only with no actual scan activation.
Immunocytochemistry
The samples were fixed on microscope slides with 4% paraformaldehyde and permeabilized with 0.1% Triton™ X-100 on days 5 and 10 during embryoid body and neural rosette formation, respectively. The samples were blocked with 5% bovine serum albumin and stained with PAX6 (1:50; Invitrogen™, Waltham, MA) and nestin (1:1,000; STEMCELL Technologies) primary antibodies overnight at 4°C followed by Alexa Fluort 488 Goat anti-mouse IgG (1:100; Invitrogen) and Alexa Fluor 594 Goat anti-rabbit IgG (1:100; Invitrogen) secondary antibodies at room temperature for 60 min. All samples were imaged using the Axiovert 200 fluorescence microscope (Carl Zeiss MicroImaging Inc., Thornwood, NY).
RNA Sequencing
Embryoid bodies and neural rosettes were removed using a cell scraper and manually broken down to a single cell suspension using a pestle followed by passing the cells through a 21-g needle. Total RNA was extracted and purified using the PureLink® RNA Mini Kit (Invitrogen), and DNA-free™ DNA Removal Kit (Invitrogen) to remove genomic DNA. RNA quality was assessed using an Agilent Technologies RNA 2100 Bioanalyzer (Santa Clara, CA).
NEBNext® Library Prep Kit (New England Biolabs® Inc., Ipswitch, MA) was used for library preparation. First, mRNA was purified from total RNA using poly-T oligo-attached magnetic beads. Then mRNA was fragmented randomly, and first-strand cDNA was synthesized using random hexamer primer and M-MuLV reverse transcriptase (RNase H-). Second-strand cDNA synthesis was subsequently performed using DNA polymerase I and RNase H. Double-stranded cDNA was purified using AMPure XP beads (Beckman Coulter Inc., Beverly, MA). Remaining overhangs of the purified double-stranded cDNA were converted into blunt ends via exonuclease/polymerase activities. After adenylation of the 3’ ends of DNA fragments, NEBNext Adaptor with hairpin loop structure was ligated to prepare for hybridization. To select cDNA fragments of preferentially 150– 200 bp in length, the library fragments were purified using the AMPure XP system. Finally, the final library was obtained by PCR amplification and purification of PCR products using AMPure XP beads.
The library was sequenced using an Illumina® NovaSeq™ sequencer (San Diego, CA) with PE150 platform by Novogene (Sacramento, CA). Spliced transcripts alignment to a reference (STAR), a fast NGS read aligner, was used for alignment of RNA-seq data. Differential expression analysis of two conditions/groups was performed using the DESeq2 R package (42). The resulting P values were adjusted using the Benjamini and Hochberg’s approach for controlling the false discovery rate (43). The difference in gene expression between two conditions/groups was considered statistically significant if Padj < 0.05.
For gene ontology (GO) enrichment analysis we used the PANTHER 14.1 online tool (44, 45). Overrepresentation test was performed using Fisher’s exact test and Bonferroni correction for multiple testing; adjusted P values are shown in Tables 1–4. A GO biological process was considered statistically significantly up- or downregulated if Bonferroni-corrected Padj < 0.05.
TABLE 1.
Top 20 Most Affected Biological Processes According to Gene Ontology Analysis Based on Sets of Statistically Significant Downregulated Genes on Day 6 vs. Day 0 (0 Dose)
| GO biological process complete | Padj value |
|---|---|
| Nervous system development | 5.69E–39 |
| Anatomical structure morphogenesis | 1.62E–32 |
| Neurogenesis | 6.97E–31 |
| Generation of neurons | 1.07E–29 |
| Anatomical structure development | 5.08E–29 |
| System development | 1.41E–28 |
| Multicellular organism development | 5.81E–28 |
| Developmental process | 3.50E–27 |
| Neuron differentiation | 6.74E–25 |
| Central nervous system development | 1.52E–23 |
| Cell projection organization | 4.81E–23 |
| Cellular developmental process | 5.58E–22 |
| Plasma membrane bounded cell projection organization | 1.33E–21 |
| Cell differentiation | 4.79E–21 |
| Cell development | 5.64E–21 |
| Regulation of developmental process | 1.39E–20 |
| Cell morphogenesis | 7.62E–19 |
| Regulation of nervous system development | 9.28E–19 |
| Cellular component morphogenesis | 9.87E–19 |
| Regulation of cell differentiation | 1.68E–18 |
Note. The processes directly related to neural differentiation are indicated in bold face.
TABLE 4.
Top 20 Most Affected Biological Processes According to Gene Ontology Analysis Based on Sets of Statistically Significant Downregulated Genes on Day 10, High Dose vs. 0 Dose
| GO biological process complete | Padj value |
|---|---|
| Animal organ development | 6.66E–08 |
| Anatomical structure development | 1.06E–07 |
| Multicellular organism development | 1.17E–07 |
| System development | 2.88E–07 |
| Regulation of developmental process | 4.25E–07 |
| Developmental process | 6.33E–07 |
| Cellular developmental process | 0.0000255 |
| Animal organ morphogenesis | 0.0000283 |
| Regulation of cell differentiation | 0.0000314 |
| Nervous system development | 0.0000394 |
| Anatomical structure morphogenesis | 0.0000611 |
| Cell differentiation | 0.0000973 |
| Multicellular organismal process | 0.000121 |
| Epithelium development | 0.00014 |
| Tissue development | 0.000145 |
| Negative regulation of developmental process | 0.000233 |
| Central nervous system development | 0.000347 |
| Neurogenesis | 0.000418 |
| Neuron differentiation | 0.000486 |
| Forebrain development | 0.000692 |
Note. The processes directly related to neural differentiation are indicated in bold face.
Ampliseq
We used the Ion AmpliSeq™ RNA Stem Cell Research Panel that targets 930 gene expression markers to characterize hESCs using Ion Ampliseq RNA workflow (Thermo Fisher Scientific™ Inc., Waltham, MA). The sequencing library was constructed using the Ion AmpliSeq Library Kit 2.0. Template preparation and chip loading was done with the automated Ion Chef™ System. Samples were barcoded and sequenced by three samples per ion 318 chip with the Ion PGM sequencer producing more than 106 reads per sample. Relative gene expression was calculated in terms of reads per million (RPM) using torrent server with ampliSeqRNA plugin software.
Statistical Analysis of AmpliSeq Data
For each gene, data were collected for 26 trial-by-day-by-dose combinations, as shown in Table 5. For each of these combinations, the sum of read counts across genes was calculated; and then each read count was divided by the sum, and multiplied by 106, resulting in a measure expressed in RPM.
TABLE 5.
Combinations of Days and Doses for Trials 1–3 Used in Statistical Analyses
| Dose: | Trial 1 | Trial 2 | Trial 3 | ||||||
|---|---|---|---|---|---|---|---|---|---|
| 0 mGy | 17 mGy | 572 mGy | 0 mGy | 17 mGy | 572 mGy | 0 mGy | 17 mGy | 572 mGy | |
| Day 0 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | - | Yes |
| Day 4 | Yes | Yes | Yes | Yes | Yes | Yes | - | - | - |
| Day 6 | Yes | Yes | Yes | Yes | Yes | Yes | - | - | - |
| Day 10 | Yes | - | Yes | Yes | - | Yes | Yes | - | Yes |
These RPM measures were assumed to be independent between trials, as well as between doses. However, they were assumed to be correlated across days within the same trial and dose. As a result of the assumed correlation between some of the measures, a mixed model was used. The response (dependent) variable was the RPM; and the explanatory (independent) variables were: trial (1, 2 or 3); dose (0, 15 and 500 mGy); day (0, 4, 6 and 10); and dose-by-day interaction. A statistical model was fitted separately for each gene, with N = 26. The categorical day effect and dose-by-day interaction allow individually fitted RPMs for each time point (day) and each dose, without any assumption on the shape of the time trend. To show a full picture of RPM estimated values over time, per dose, from all trials combined, RPM least-squares means calculated from the model were plotted with 95% confidence intervals (Supplementary Figs. S2–21; https://doi.org/10.1667/RR15535.1.S1). Differences in RPM least-squares means between days and between doses were tabulated (Supplementary Tables S1–S3; https://doi.org/10.1667/RR15535.1.S1). The SAS statistical software (Cary, NC) was used for this analysis.
The datasets generated in the current study have been deposited into the NCBI Gene Expression Omnibus, accession no. GSE137800.
RESULTS
Colonies of H9 hESCs received mock (0 Gy), low-dose (17 mGy) or high-dose (572 mGy) CT exposure, and were then directed to differentiate into neuronal lineages via an embryoid body stage protocol as described elsewhere (12). Total RNA was purified 24 h postirradiation (day 0), on day 6 from embryoid bodies and on day 10 from neural rosettes. Figure 1 shows that, on day 10, both irradiated and mock-irradiated hESCs grown in STEMdiff Neural Progenitor Medium formed neural rosettes. The cells expressed a neural precursor marker, PAX6, and a neural differentiation marker, nestin (NES), and the NES-positive cells were concentrated in the centers of the neural rosettes. The differentiation experiments were independently repeated three times, giving rise to three biological replicate RNA samples.
FIG. 1.
Staining of hESC on day 10 of neural differentiation with PAX6 (red) and nestin (green) antibodies. Panel A: Mock-irradiated samples. Panel B: 17 mGy irradiated samples. Panel C: 572 mGy irradiated samples.
Gene expression in the RNA samples was assessed using RNA sequencing (RNA-seq). Comparison of gene expression on days 6 and 10 to day 0 showed a robust differentiation into neuronal lineages in mock-, low-dose, and high-dose irradiated samples (Fig. 2). In all of the samples, approximately 3,000 genes were upregulated, and 3,000 genes were downregulated. GO analysis of upregulated genes showed that the most upregulated biological processes were nervous system development, neurogenesis and generation of neurons (Tables 1 and 2). One of the most downregulated biological processes was stem cell population maintenance with expression of the key pluripotency markers POU5F1 and NANOG reduced by more than 300 times in all of the samples on day 10. This confirms our previous findings that even exposure to a relatively high dose (572 mGy) of ionizing radiation did not prevent hESCs from differentiating into neuronal lineages.
FIG. 2.
Volcano diagrams of differential expression of genes; pairwise comparisons at day 0 and day 10 for mock- (panel A), low-dose (panel B) and high-dose (panel C) irradiated samples. Horizontal axis shows the fold change of gene expression in different samples. Vertical axis shows the statistical significance of the degree of changes in gene expression levels. The difference in gene expression between two conditions/groups was considered statistically significant if Padj < 0.05 (−log10(Padj) > 1.3). Each point represents a gene. Blue dots indicate no significant difference in gene expression, red dots indicate upregulated differential gene expression and green dots indicate downregulated differential gene expression.
TABLE 2.
Top 20 Most Affected Biological Processes According to Gene Ontology Analysis Based on Sets of Statistically Significant Downregulated Genes on Day 10 vs. Day 0 (0 Dose)
| GO biological process complete | Padj value |
|---|---|
| Nervous system development | 3.76E–44 |
| Anatomical structure morphogenesis | 1.93E–40 |
| Anatomical structure development | 1.30E–38 |
| System development | 1.43E–38 |
| Developmental process | 1.74E–38 |
| Multicellular organism development | 2.09E–35 |
| Neurogenesis | 4.81E–34 |
| Generation of neurons | 3.48E–32 |
| Cellular developmental process | 5.18E–28 |
| Neuron differentiation | 1.74E–27 |
| Cell development | 4.64E–27 |
| Cell differentiation | 6.31E–27 |
| Cell projection organization | 2.04E–26 |
| Plasma membrane bounded cell projection organization | 1.21E–25 |
| Animal organ morphogenesis | 2.25E–23 |
| Animal organ development | 6.57E–23 |
| Cellular component morphogenesis | 8.69E–23 |
| Regulation of developmental process | 3.24E–22 |
| Multicellular organismal process | 4.20E–22 |
| Cell morphogenesis | 6.00E–22 |
Note. The processes directly related to neural differentiation are indicated in bold face.
We then compared gene expression between irradiated and mock-irradiated cells at different time points postirradiation. Samples receiving low-dose irradiation showed no statistically significant differences in gene expression compared to mock-irradiated cells at days 0, 6 and 10 of differentiation (Supplementary Fig. S1; https://doi.org/10. 1667/RR15535.1.S1). Samples receiving high-dose irradiation showed changes in expression of a number of genes that progressively increased from day 0 to day 10 (Fig. 3). On day 0, at 24 h postirradiation, we detected statistically significant changes in expression of just four genes (Fig. 3A). Interestingly, the only upregulated gene was BTG2, a tumor suppressor that can be activated by both TP53 and NF-κB pathways (13) in response to oxidative stress to stimulate repair of DNA double-strand breaks (14). On day 6, at the embryoid body stage, changes in expression of 35 genes were detected, 32 of which were downregulated. Among the genes downregulated by exposure to radiation were those that play a key role in nervous system development, such as PAX7, ELAVL4, DLL1, NHLH1, ROBO3 and ASIC4 (15–17, 18, 19). GO analysis showed that, among the biological processes most significantly downregulated, by high-dose radiation exposure, on day 6, were nervous system development, neurogenesis and generation of neurons (Table 3).
FIG. 3.
Volcano diagrams of differential expression of genes; pairwise comparisons of mock- and high-dose irradiated samples at day 0 (panel A), day 6 (panel B) and day 10 (panel C). Horizontal axis shows the fold change of gene expression in different samples. Vertical axis shows the statistical significance of the degree of changes in gene expression levels. The difference in gene expression between two conditions/groups was considered statistically significant if Padj < 0.05 (−log10(Padj) > 1.3). Each point represents a gene. Blue dots indicate no significant difference in gene expression, red dots indicate upregulated differential gene expression and green dots indicate downregulated differential gene expression.
TABLE 3.
Top 20 Most Affected Biological Processes According to Gene Ontology Analysis Based on Sets of Statistically Significant Downregulated Genes on Day 6, High Dose vs. 0 Dose
| GO biological process complete | Padj value |
|---|---|
| Nervous system development | 1.27E–32 |
| Neurogenesis | 1.92E–25 |
| Anatomical structure morphogenesis | 9.89E–25 |
| Generation of neurons | 1.79E–24 |
| Anatomical structure development | 4.05E–24 |
| Developmental process | 1.72E–23 |
| Multicellular organism development | 1.95E–23 |
| System development | 3.74E–23 |
| Plasma membrane bounded cell projection organization | 6.61E–23 |
| Central nervous system development | 1.00E–22 |
| Cell projection organization | 1.64E–22 |
| Head development | 1.49E–21 |
| Brain development | 7.34E–21 |
| Neuron differentiation | 6.27E–20 |
| Sensory organ development | 6.56E–18 |
| Cellular developmental process | 5.36E–17 |
| Regulation of cellular process | 5.39E–17 |
| Regulation of transcription by RNA polymerase II | 8.10E–17 |
| Regulation of biological process | 3.50E–16 |
| Cell differentiation | 3.96E–16 |
Note. The processes directly related to neural differentiation are indicated in bold face.
At the stage of neural rosette formation, day 10 of differentiation, exposure to high-dose radiation affected expression of 68 genes; 25 up- and 43 downregulated. Such transcription factors and major regulators of neuronal development as NEUROG1, ARX, ASCL1, INSM1, RFX4, FGFR3, MEF2C and PAX6 were the most significantly downregulated by radiation exposure (20–27). Among the most significantly downregulated biological processes were regulation of cell differentiation, nervous system development, cell differentiation and neurogenesis (Table 4). GO analysis of upregulated genes did not reveal any statistically significantly affected biological processes. Thus, we concluded that, while early stages of directed differentiation into neuronal lineage cells showed expected morphological features, exposure to the high-dose radiation caused statistically significant reduction in the expression of the genes involved in nervous system development.
Cluster analysis of the samples (Fig. 4) showed that, as expected, the day 6 and day 10 differentiated samples showed gene expression patterns that are quite similar but radically different from that of undifferentiated cells on day 0. Interestingly, within the clusters for each day, mock- and low-dose irradiated samples clustered together and separated from the high-dose samples.
FIG. 4.
Hierarchical clustering heatmap of differential expression genes. Red represents high-expression genes, blue represents low-expression genes. Color descending from red to blue, indicated log10(FPKM + 1) from large to small.
We used an AmpliSeq RNA Stem Cell Research panel of genes to verify our RNA-seq results. The comparisons of RNA-seq and Ampliseq data for the genes most considerably downregulated by radiation that are involved in neurogenesis (Fig. 5) show that the data are in good agreement. Ampliseq expression results on the selected genes receiving mock-, low- and high-dose irradiation at days 1, 4, 6 and 10 are shown in Supplementary Figs. S2–21 (https://doi.org/10.1667/RR15535.1.S1). The genes can be separated into three groups. The first group consists of genes that are consistently downregulated at all time points, such as ARX, PAX6, ASCL1, NEUROG1 and LEF1 (20). The second group consists of genes that became downregulated only at the latest time point, on day 10, such as INSM1, RFX4, NEUROD1, MEF2C, TBX5 and FGFR3. The third group of genes are those that had decreased expression at earlier time points (days 4 and 6) relative to the mock-irradiated samples, which then became almost equal to the mock-irradiated samples on day 10, such as PAX7, WNT1, DLL1, CRABP2, CLU and ASIC4 (28–30). Interestingly, the expression of most of the genes in the third group in the mock-irradiated samples was only transiently elevated on days 4 or 6 and then decreased on day 10. Notably, the Ampliseq data showed downregulation of several genes after low-dose irradiation, such as ARX, NEUROG1, DLL1, CLU, ASIC4, and especially CRABP2 when considered individually. However, downregulation of these genes was not statistically significant; because of the multiple comparisons, only P values that are approximately 0.0001 or less reflect evidence of true differences (Supplementary Tables S1–S3).
FIG. 5.
Comparisons of the changes in expression [log(fold change)] of selected genes on day 6 (panel A) and day 10 (panel B) relative to day 0 obtained by RNA-seq (blue) and Ampliseq (orange) experiments.
DISCUSSION
Ionizing radiation exposure in the early stages of in utero development has the potential to cause immediate effects such as microcephaly and growth retardation, as well as delayed effects including various types of cancer (1–7). Epidemiological studies based on A-bomb survivors and pregnant women exposed to radiation, particularly from CT scans during the early stages of pregnancy, have suggested that exposure of a blastocyst results in “all or none” phenomena, leading to either miscarriage or with no detectable consequences at an estimated threshold of approximately 200 mGy (2, 3).
Several published retrospective epidemiological studies have shown that CT exposure during childhood increases the risks of central nervous system tumors, leukemia and other cancer types. However, in other studies these results have been called into question, suggesting that CT-related brain cancer risk was confounded by insufficient controls for cancer susceptibility syndromes and potential methodological biases (31).
The use of in vitro models such as cell cultures and organoids could provide valuable insights into the mechanisms of both early effects and carcinogenesis. This approach has been widely used in toxicology studies (32). Recently, with the use of the cerebral organoid model, it was shown that pleiotropic drugs could mitigate the effect of radiation on organoid development (33). Of course, caution should be taken when extending results obtained in experiments with organoids to human embryo development. However, the use of organoids made of human cells may complement the data obtained with mouse embryos, and in many respects has an advantage over mouse models because the neurogenesis in the latter has been shown to be different from that in human embryos (10).
We employed an in vitro neuro-development model to assess the effects of CT radiation on gene expression in neuronal lineage cells. In particular, we focused on early stages of hESC differentiation into neuronal precursors. We found that exposure of hESCs to a high dose (572 mGy) considerably downregulated expression of multiple genes involved in neurogenesis. However, exposure to a relatively low dose (17 mGy) associated with a single abdominal/pelvic CT scan did not cause notable changes in levels of gene expression. At the same time, we did not notice any morphological changes in the formation of neuronal rosettes nor even in neural progenitor cells that further progressed along the differentiation process (12). As previously shown, exposure of hESCs to 572 mGy of radiation resulted in considerable cell death via apoptosis (12). This could have delayed the differentiation program, and in principle, the “repaired” cells could still differentiate into normal neural progenitor cells in culture. However, such a delay in utero, where each step of development is precisely timed, could have catastrophic consequences.
The samples exposed to the higher dose (572 mGy) revealed a significant downregulation of a number of genes that play essential roles in neurogenesis. Mutations that disrupt functions of genes such as NEUROD1, ARX, ASCL1, RFX4, MEF2C, PAX7, ELAVL4 and ROBO3 have been linked to various neurodevelopmental and neuropsychiatric disorders (34–41). This raises concerns for the pregnant population exposed at therapeutic doses typically used in radiotherapy, especially because differentiation events that contribute to human embryonic brain development start at the time when pregnancy cannot be reliably confirmed in most women. Experiments with longer differentiation times leading to formation of brain organoids may provide further insights on how gene expression changes translate to morphological abnormalities.
In contrast, the lower dose, representative of a single abdominal/pelvic CT scan (17 mGy) exposure, did not affect expression of genes involved in neurogenesis. However, we cannot exclude the possibility of low-dose radiation effects, especially over the long term, on neurogenesis in the considerably more complex system of a human embryo. Future studies following 3D organoids to later stages of brain development have the potential to advance our understanding of the effects of low-dose radiation at various stages of embryogenesis and may offer improved estimations of dose effects and thresholds. This would ultimately provide a better overall assessment of the risks of in utero radiation exposure.
Supplementary Material
Fig. S1. Volcano diagrams of differential expression of genes; pairwise comparisons of mock- and low-dose irradiated samples at days 0, 6 and 10.
Fig. S2. Analysis for gene ARX.
Fig. S3. Analysis for gene ASCL1.
Fig. S4. Analysis for gene INSM1.
Fig. S5. Analysis for gene NEUROG1.
Fig. S6. Analysis for gene PAX7.
Fig. S7. Analysis for gene RFX4.
Fig. S8. Analysis for gene WNT1.
Fig. S9. Analysis for gene DLL1.
Fig. S10. Analysis for gene CXCR4.
Fig. S11. Analysis for gene NEUROD1.
Fig. S12. Analysis for gene MEF2C.
Fig. S13. Analysis for gene TBX5.
Fig. S14. Analysis for gene FGFR3.
Fig. S15. Analysis for gene LEF1.
Fig. S16. Analysis for gene ELAVL3.
Fig. S17. Analysis for gene CRABP2.
Fig. S18. Analysis for gene CLU.
Fig. S19. Analysis for gene COL1A2.
Fig. S20. Analysis for gene ASIC4.
Fig. S21. Analysis for gene PAX6.
ACKNOWLEDGMENT
This study was supported by the Intramural Research Program of the Clinical Center, National Institutes of Health.
Footnotes
Editor’s note. The online version of this article (DOI: https://doi.org/10.1667/RR15535.1) contains supplementary information that is available to all authorized users.
REFERENCES
- 1.Otake M, Schull WJ, Lee S. Threshold for radiation-related severe mental retardation in prenatally exposed A-bomb survivors: a re-analysis. Int J Radiat Biol 1996; 70:755–63. [DOI] [PubMed] [Google Scholar]
- 2.Brent RL. Saving lives and changing family histories: appropriate counseling of pregnant women and men and women of reproductive age, concerning the risk of diagnostic radiation exposures during and before pregnancy. Am J Obstet Gynecol 2009; 200:4–24. [DOI] [PubMed] [Google Scholar]
- 3.Guilbaud L, Beghin D, Dhombres F, Blondiaux E, Friszer S, Ducou Le Pointe H, et al. Pregnancy outcome after first trimester exposure to ionizing radiations. Eur J Obstet Gynecol Reprod Biol 2019; 232:18–21. [DOI] [PubMed] [Google Scholar]
- 4.Blot WJ, Miller RW. Mental retardation following in utero exposure to the atomic bombs of Hiroshima and Nagasaki. Radiology 1973; 106:617–9. [DOI] [PubMed] [Google Scholar]
- 5.Otake M, Schull WJ. In utero exposure to A-bomb radiation and mental retardation; a reassessment. Br J Radiol 1984; 57:409–14. [DOI] [PubMed] [Google Scholar]
- 6.Wood JW, Johnson KG, Omori Y. In utero exposure to the Hiroshima atomic bomb. An evaluation of head size and mental retardation: twenty years later. Pediatrics 1967; 39:385–92. [PubMed] [Google Scholar]
- 7.De Santis M, Cesari E, Nobili E, Straface G, Cavaliere AF, Caruso A. Radiation effects on development. Birth Defects Res C Embryo Today 2007; 81:177–82. [DOI] [PubMed] [Google Scholar]
- 8.Saha S, Woodbine L, Haines J, Coster M, Ricket N, Barazzuol L, et al. Increased apoptosis and DNA double-strand breaks in the embryonic mouse brain in response to very low-dose X-rays but not 50 Hz magnetic fields. J R Soc Interface 2014; 11:20140783. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Barazzuol L, Jeggo PA. In vivo sensitivity of the embryonic and adult neural stem cell compartments to low-dose radiation. J Radiat Res 2016; 57:i2–i10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Zhang X, Huang CT, Chen J, Pankratz MT, Xi J, Li J, et al. Pax6 is a human neuroectoderm cell fate determinant. Cell Stem Cell 2010; 7:90–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Arlotta P, Pasca SP. Cell diversity in the human cerebral cortex: from the embryo to brain organoids. Curr Opin Neurobiol 2019; 56:194–8. [DOI] [PubMed] [Google Scholar]
- 12.Hanu C, Loeliger BW, Panyutin IV, Maass-Moreno R, Wakim P, Pritchard WF, et al. Effect of ionizing radiation from computed tomography on differentiation of human embryonic stem cells into neural precursors. Int J Mol Sci 2019; 20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Imran M, Lim IK. Regulation of Btg2(/TIS21/PC3) expression via reactive oxygen species-protein kinase C-NuFkappaBeta pathway under stress conditions. Cell Signal 2013; 25:2400–12. [DOI] [PubMed] [Google Scholar]
- 14.Choi KS, Kim JY, Lim SK, Choi YW, Kim YH, Kang SY, et al. TIS21(/BTG2/PC3) accelerates the repair of DNA double strand breaks by enhancing Mre11 methylation and blocking damage signal transfer to the Chk2(T68)-p53(S20) pathway. DNA Repair (Amst) 2012; 11:965–75. [DOI] [PubMed] [Google Scholar]
- 15.Monsoro-Burq AH. PAX transcription factors in neural crest development. Semin Cell Dev Biol 2015; 44:87–96. [DOI] [PubMed] [Google Scholar]
- 16.Grassi E, Santoro R, Umbach A, Grosso A, Oliviero S, Neri F, et al. Choice of alternative polyadenylation sites, mediated by the RNA-binding protein Elavl3, plays a role in differentiation of inhibitory neuronal progenitors. Front Cell Neurosci 2018; 12:518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kageyama R, Shimojo H, Ohtsuka T. Dynamic control of neural stem cells by bHLH factors. Neurosci Res 2019; 138:12–8. [DOI] [PubMed] [Google Scholar]
- 18.Couch JA, Chen J, Rieff HI, Uri EM, Condron BG. robo2 and robo3 interact with eagle to regulate serotonergic neuron differentiation. Development 2004; 131:997–1006. [DOI] [PubMed] [Google Scholar]
- 19.Hoshikawa M, Kato A, Hojo H, Shibata Y, Kumamoto N, Watanabe M, et al. Distribution of ASIC4 transcripts in the adult wild-type mouse brain. Neurosci Lett 2017; 651:57–64. [DOI] [PubMed] [Google Scholar]
- 20.Velkey JM, O’Shea KS. Expression of Neurogenin 1 in mouse embryonic stem cells directs the differentiation of neuronal precursors and identifies unique patterns of down-stream gene expression. Dev Dyn 2013; 242:230–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ehashi T, Suzuki N, Ando S, Sumida K, Saito K. Effects of valproic acid on gene expression during human embryonic stem cell differentiation into neurons. J Toxicol Sci 2014; 39:383–90. [DOI] [PubMed] [Google Scholar]
- 22.Ashique AM, Choe Y, Karlen M, May SR, Phamluong K, Solloway MJ, et al. The Rfx4 transcription factor modulates Shh signaling by regional control of ciliogenesis. Sci Signal 2009; 2:ra70. [DOI] [PubMed] [Google Scholar]
- 23.Zhang D, Zeldin DC, Blackshear PJ. Regulatory factor X4 variant 3: a transcription factor involved in brain development and disease. J Neurosci Res 2007; 85:3515–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Monaghan CE, Nechiporuk T, Jeng S, McWeeney SK, Wang J, Rosenfeld MG, et al. REST corepressors RCOR1 and RCOR2 and the repressor INSM1 regulate the proliferation-differentiation balance in the developing brain. Proc Natl Acad Sci U S A 2017; 114:E406–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Cho EG, Zaremba JD, McKercher SR, Talantova M, Tu S, Masliah E, et al. MEF2C enhances dopaminergic neuron differentiation of human embryonic stem cells in a parkinsonian rat model. PLoS One 2011; 6:e24027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Mayer C, Hafemeister C, Bandler RC, Machold R, Batista Brito R, Jaglin X, et al. Developmental diversification of cortical inhibitory interneurons. Nature 2018; 555:457–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yamamizu K, Piao Y, Sharov AA, Zsiros V, Yu H, Nakazawa K, et al. Identification of transcription factors for lineage-specific ESC differentiation. Stem Cell Reports 2013; 1:545–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Chou AH, Zheng S, Itsukaichi T, Howard BD. Wnt-1 inhibits nerve growth factor-induced differentiation of PC12 cells by preventing the induction of some but not all late-response genes. Brain Res Mol Brain Res 2000; 77:232–45. [DOI] [PubMed] [Google Scholar]
- 29.Chaerkady R, Kerr CL, Marimuthu A, Kelkar DS, Kashyap MK, Gucek M, et al. Temporal analysis of neural differentiation using quantitative proteomics. J Proteome Res 2009; 8:1315–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Cordero-Llana O, Scott SA, Maslen SL, Anderson JM, Boyle J, Chowhdury RR, et al. Clusterin secreted by astrocytes enhances neuronal differentiation from human neural precursor cells. Cell Death Differ 2011; 18:907–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Bernier MO, Baysson H, Pearce MS, Moissonnier M, Cardis E, Hauptmann M, et al. Cohort Profile: The EPI-CT study: A European pooled epidemiological study to quantify the risk of radiation-induced cancer from paediatric CT. Int J Epidemiol 2019; 48:379–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Suzuki N, Ando S, Sumida K, Horie N, Saito K. Analysis of altered gene expression specific to embryotoxic chemical treatment during embryonic stem cell differentiation into myocardiac and neural cells. J Toxicol Sci 2011; 36:569–85. [DOI] [PubMed] [Google Scholar]
- 33.Shakhbazau A, Danilkovich N, Seviaryn I, Ermilova T, Kosma cheva S. Effects of minocycline and rapamycin in gamma-irradiated human embryonic stem cells-derived cerebral organoids. Mol Biol Rep 2019; 46:1343–8. [DOI] [PubMed] [Google Scholar]
- 34.Zhou Y, Qing J, Dong Y, Nie J, Li J, Wang C, et al. The role of transcription factors of neurosensory cells in non-syndromic sensorineural hearing loss with or without inner ear malformation. Acta Otolaryngol 2016; 136:277–82. [DOI] [PubMed] [Google Scholar]
- 35.Marsh E, Fulp C, Gomez E, Nasrallah I, Minarcik J, Sudi J, et al. Targeted loss of Arx results in a developmental epilepsy mouse model and recapitulates the human phenotype in heterozygous females. Brain 2009; 132:1563–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Oliveira MAP, Balling R, Smidt MP, Fleming RMT. Embryonic development of selectively vulnerable neurons in Parkinson’s disease. NPJ Parkinsons Dis 2017; 3:21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Glaser B, Kirov G, Bray NJ, Green E, O’Donovan MC, Craddock N, et al. Identification of a potential bipolar risk haplotype in the gene encoding the winged-helix transcription factor RFX4. Mol Psychiatry 2005; 10:920–7. [DOI] [PubMed] [Google Scholar]
- 38.Harrington AJ, Raissi A, Rajkovich K, Berto S, Kumar J, Molinaro G, et al. MEF2C regulates cortical inhibitory and excitatory synapses and behaviors relevant to neurodevelopmental disorders. Elife 2016; 5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Proskorovski-Ohayon R, Kadir R, Michalowski A, Flusser H, Perez Y, Hershkovitz E, et al. PAX7 mutation in a syndrome of failure to thrive, hypotonia, and global neurodevelopmental delay. Hum Mutat 2017; 38:1671–83. [DOI] [PubMed] [Google Scholar]
- 40.Haugarvoll K, Toft M, Ross OA, Stone JT, Heckman MG, White LR, et al. ELAVL4, PARK10, and the Celts. Mov Disord 2007; 22:585–7. [DOI] [PubMed] [Google Scholar]
- 41.Bosley TM, Salih MA, Jen JC, Lin DD, Oystreck D, Abu-Amero KK, et al. Neurologic features of horizontal gaze palsy and progressive scoliosis with mutations in ROBO3. Neurology 2005; 64:1196–203. [DOI] [PubMed] [Google Scholar]
- 42.Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014; 15:550. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Hochberg Y, Benjamini Y. More powerful procedures for multiple significance testing. Stat Med 1990; 9:811–8. [DOI] [PubMed] [Google Scholar]
- 44.Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 2000; 25:25–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.The Gene Ontology Consortium. The gene ontology resource: 20 years and still GOing strong. Nucleic Acids Res 2019; 47:D330–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Fig. S1. Volcano diagrams of differential expression of genes; pairwise comparisons of mock- and low-dose irradiated samples at days 0, 6 and 10.
Fig. S2. Analysis for gene ARX.
Fig. S3. Analysis for gene ASCL1.
Fig. S4. Analysis for gene INSM1.
Fig. S5. Analysis for gene NEUROG1.
Fig. S6. Analysis for gene PAX7.
Fig. S7. Analysis for gene RFX4.
Fig. S8. Analysis for gene WNT1.
Fig. S9. Analysis for gene DLL1.
Fig. S10. Analysis for gene CXCR4.
Fig. S11. Analysis for gene NEUROD1.
Fig. S12. Analysis for gene MEF2C.
Fig. S13. Analysis for gene TBX5.
Fig. S14. Analysis for gene FGFR3.
Fig. S15. Analysis for gene LEF1.
Fig. S16. Analysis for gene ELAVL3.
Fig. S17. Analysis for gene CRABP2.
Fig. S18. Analysis for gene CLU.
Fig. S19. Analysis for gene COL1A2.
Fig. S20. Analysis for gene ASIC4.
Fig. S21. Analysis for gene PAX6.