56 Fe IONS ON HUMAN MESENCHYMAL STEM CELLS KYLE KURPINSKI, PH.D.,* DEOK-JIN JANG, PH.D., y SANCHITA BHATTACHARYA, M.S., y BJORN RYDBERG, PH.D., y JULIA CHU, B.S.,* JOANNA SO, B.S.,* ANDY WYROBEK, PH.D., y SONG LI, PH.D.,* AND DAOJING WANG, PH.D. y *Department of Bioengineering, University of CaliforniaBerkeley, Berkeley, CA; and y Life Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA Purpose: Stem cells hold great potential for regenerative medicine, but they have also been implicated in cancer and aging. How different kinds of ionizing radiation affect stem cell biology remains unexplored. This study was designed to compare the biological effects of X-rays and of highlinear energy transfer (LET) 56 Fe ions on hu- man mesenchymal stem cells (hMSC). Methods and Materials: A multi-functional comparison was carried out to investigate the differential effects of X-rays and 56 Fe ions on hMSC. The end points included modulation of key markers such as p53, cell cycle progression, osteogenic differentiation, and pathway and networks through transcriptomic proling and bioinfor- matics analysis. Results: X-rays and 56 Fe ions differentially inhibited the cell cycle progression of hMSC in a p53-dependent man- ner without impairing their in vitro osteogenic differentiation process. Pathway and network analyses revealed that cytoskeleton and receptor signaling were uniquely enriched for low-dose (0.1 Gy) X-rays. In contrast, DNA/RNA metabolism and cell cycle regulation were enriched for high-dose (1 Gy) X-rays and 56 Fe ions, with more signicant effects from 56 Fe ions. Specically, DNA replication, DNA strand elongation, and DNA bind- ing/transferase activity were perturbed more severely by 1 Gy 56 Fe ions than by 1 Gy X-rays, consistent with the signicant G2/M arrest for the former while not for the latter. Conclusions: 56 Fe ions exert more signicant effects on hMSC than X-rays. Since hMSC are the progenitors of osteoblasts in vivo, this study provides new mechanistic understandings of the relative health risks associated with low- and high-dose X-rays and high-LET space radiation. 2009 Elsevier Inc. Radioresponse, High-LET, Cell cycle, Osteogenic differentiation, Transcriptomics. INTRODUCTION Stem cells hold great potential for regenerative medicine be- cause they can self-renew and differentiate along tissue-spe- cic lineages. Mesenchymal stem cells (MSCs), the multipotent stromal cells from bone marrow, play important roles in the maintenance and repair of various tissues, includ- ing bone, cartilage, and muscle (1). The differentiation of MSCs into osteoblasts is controlled by the transcription factor Runx2 (2). Stem cells also respond to genotoxic stresses, for example, ionizing radiation (IR), which may lead to the induc- tion of mutations and chromosome aberrations, transient cell- cycle arrest, apoptosis, mitotic catastrophe, cellular senes- cence, or malignant transformation. Abnormalities in stem cell biology have been implicated in cancer and aging, but the mechanisms are poorly understood (3). For example, MSCs have been shown as a target for neoplastic transforma- tion following manipulations (4, 5). However, major chal- lenges still remain to dene the normal molecular signaling and tumorigenic pathways in stem cells. Because of their high tissue-penetrating property, IR such as X-rays/gamma rays is used in CT scans, cancer Reprint requests to: Daojing Wang, Ph.D., Life Sciences Divi- sion, Lawrence Berkeley National Laboratory, 1 Cyclotron Road, MS 977-225A, Berkeley, CA 94720. Tel: (510) 486-6592; Fax: (510) 495-2535; E-mail: djwang@lbl.gov Conict of interest: none. AcknowledgmentsThe authors thank Dr. M. Bissell for advice and Drs. B. Sutherland and P. Guida for help with the logistics of Na- tional Aeronautics and Space Administration (NASA) Space Radi- ation Laboratory (NSRL) experiments. This work was supported by the Low Dose Radiation Research Program jointly funded by the U.S. Department of Energy (DOE) and NASA (to D.W.). A.W. acknowledges the funding support from the DOE Low Dose Radi- ation Research Program. S.L. acknowledges the funding support from the National Institutes of Health (HL079419). This work was performed under the auspices of the U.S. DOE, at the University of California/Lawrence Berkeley National Laboratory under contract No. DE-AC02-05CH11231. Supplementary material for this article can be found at www.redjournal.org. Received May 12, 2008, and in revised form Aug 29, 2008. Accepted for publication Sept 3, 2008. 869 Int. J. Radiation Oncology Biol. Phys., Vol. 73, No. 3, pp. 869877, 2009 Copyright 2009 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/09/$see front matter doi:10.1016/j.ijrobp.2008.10.002 radiotherapy, and nuclear medicine. However, IR is also a well-known DNA-damaging agent, clastogen, and carcino- gen (6, 7). During deep space missions, space crews encounter an elevated radiation environment that includes high-energy charged (HZE) particles, such as 56 Fe ions from the galactic cosmic rays (8). Such densely ionizing radiation induces more severe clustering of damage on DNA compared to X-rays, and gives rise to complex DNA double-strand breaks (DSBs) that are more prone to be misrepaired and cause permanent genomic changes (9). Another complicatingfactor duringspace travel is the possibility of synergistic relationships with other risk factors, such as microgravity. The possibility that HZE radiation affects the differentiation process of human mesen- chymal stem cells (hMSC) into osteoblasts, important for bone regeneration and loss in space because of microgravity, has not been investigated. Advances in systems biology have opened up new possi- bilities for elucidating the molecular mechanisms underlying cellular responses to IR. Global gene expression changes in response to IR have been reported for Saccharomyces cerevi- siae (10), murine brain (11), normal human skin broblasts (12), and human lung broblasts (13). These studies sug- gested that IR regulates gene expression at the transcriptional level in a dose-dependent manner and depending on radiation quality, for example, X-rays vs. HZE particles. On the other hand, proteomic and transcriptomic studies of hMSC have provided important insights into their proliferation and differ- entiation (14, 15). However, no direct correlations between phenotypic radioresponses and pathway regulations have been reported for hMSC, particularly in the context of osteo- genic differentiation. Here we describe the rst comprehensive study of the radioresponses of hMSC to X-rays and high energy 56 Fe ions. We show that neither X-rays nor 56 Fe ions impaired the in vitro osteogenic differentiation process of hMSC. We further show that they differentially regulated the cell cycle progression of hMSC in a dose-dependent and p53- dependent manner. Pathway and network analyses of transcriptomic proles revealed the underlying mechanisms. MATERIALS AND METHODS Cell culture Human mesenchymal stem cells were obtained from Lonza Group (Walkersville, MD). The log number was 4F1560, which was derived from the bone marrow of a healthy 23-year-old woman. We characterized expanded hMSC (up to passage 10) again by their surface markers and differentiation potential as we described previ- ously (14, 15). For cell proliferation without differentiation, hMSC (typical passage number, ve to eight) were cultured in MSCgrowth medium, supplemented with prescreened fetal bovine serum, L-glu- tamine, and a penicillin/streptomycin mix (MSCGM, Lonza). For osteogenic differentiation after irradiation, hMSC were expanded to 70% conuency and then supplemented with osteogenic differ- entiation medium containing dexamethasone, b-glycerophosphate, and ascorbate (Lonza) approximately 1.5 h before exposure. After irradiation, cells were maintained in osteogenic differentiation medium for 1, 3, 6, or 7 days and characterized. For growth and proliferation after irradiation, hMSC were expanded to 35% conuency before exposure. After irradiation, cells were maintained in MSCGM for 3, 5, or 24 h and characterized. Irradiation with X-rays and 56 Fe ions X-ray irradiation was performed at the Lawrence Berkeley National Laboratory (LBNL) with 320 kVp X-rays (Pantak Inc., Branford, CT) with 0.5 mm copper ltration. Acalibrated ion cham- ber with temperature and pressure correction (RadCal Corporation, Monrovia, CA) was used for dosimetry. 56 Fe ion irradiation was per- formed at the National Aeronautics and Space Administration (NASA) Space Radiation Laboratory (NSRL) radiation facility at Brookhaven National Laboratory (BNL) with 1 GeV/amu 56 Fe ions, LET = 150 keV/mm. Dosimetry was carried out by NSRL per- sonnel, as previously described (16). Triplicates of individual hMSC cultures were exposed to 0 Gy (sham control), 0.1 Gy (low dose), and 1 Gy (high dose) for X-rays and 56 Fe ions, respectively. For both X-rays and 56 Fe ions, the dose rate used was 0.1 Gy/min and 1 Gy/min for 0.1 Gy and 1 Gy, respectively. To estimate the number of 56 Fe traversals per cell nucleus, the average nuclear area perpendicular to the beam was measured using live Hoechst 33342 staining and the number of traversals per cell was calculated as N = 6.24 dose (Gy) area (mm 2 ) / LET (keV/mm) (17). The nuclear diameter of hMSC grown on plastic substrates was found under our growth conditions to be 20.2 3.2 mm (SD), corresponding to a nuclear area of 330 110 mm 2 (SD). The average number of traversals per cell is then calculated to be 1.36 per nucleus for the dose of 0.1 Gy and 13.6 per nucleus for the dose of 1 Gy. Applying Poisson statistics, 75% of the cell nu- clei were hit by at least one 56 Fe particle for a dose of 0.1 Gy, with 25% of the cell nuclei exposed only to a low dose of delta rays. At 1 Gy, 100% of the cells were traversed by multiple 56 Fe particles. Microscopy, alkaline phosphatase activity assay, and calcium staining Live-cell phase-contrast images were collected using a Nikon TMS microscope and a Hamamatsu Orca100 cooled digital CCD camera, usingC-ImagingSystemsoftware (CompixInc., Irvine, CA). Osteogenic differentiation was assessed using a colorimetric ALP activity assay and Alizarin Red S staining. Cells were lysed with 0.5% Triton X-100. SIGMAFAST p-nitrophenyl phosphate solution (N1891-5SET, Sigma, St. Louis, MO) was added to the supernatant. Absorbance at 405 nm was taken at several intervals from 0 to 60 min immediately after adding the substrate. Alkaline phosphatase (ALP) activity was calculated as the rate of increasing absorbance over time (slope value), and was normalized to the overall protein concentration as determined by DC Protein Assay (Bio-Rad, Hercu- les, CA). For calcium staining, cells after differentiation for 3 weeks were xed with 70% ethanol and stained with 1% Alizarin Red S solution (Sigma) at room temperature. Flow cytometry and immunoblotting analysis To determine the effects of IR on hMSC cell cycle progression, a standard FACS cell cycle analysis protocol was used (15). Immu- noblotting analysis of total cell lysates was performed as described (14). The mouse monoclonal antibody against total actin (including all isoforms) was from Chemicon (Temecula, CA). The phospho- specic rabbit polyclonal anti-pSer15-p53 and mouse monoclonal anti-pErk1/2 antibodies were from Cell Signaling (Danvers, MA). 870 I. J. Radiation Oncology d Biology d Physics Volume 73, Number 3, 2009 Microarray analysis and data processing Global mRNA expression was proled by Affymetrix (Santa Clara, CA) microarrays at the LBNL HTArray Genomics Core Fa- cility. Samples were prepared, labeled, and hybridized according to standard protocols (15). X-raytreated samples were hybridized to U133AAofAv2 chips containing 22,944 probe-sets per chip, and 56 Fe ions-treated samples were hybridized to HT_HG-U133A chips containing 22,277 probe-sets per chip. U133AAofAv2 was the ear- lier version of HT_HG-U133A and was used because of chip-type availability at the time. Signal intensities were obtained for all probe-sets and were organized and sorted using GeneTrafc v.3.2 microarray analysis software (Iobion, La Jolla, CA). Validation of microarray results by qRT-PCR. To verify gene changes obtained from microarray analysis, we used quantitative re- verse transcriptionpolymerase chain reaction (qRT-PCR) as previ- ously described (15). Primers for cyclin A2 (CCNA2), cyclin B1 (CCNB1), cyclin E2 (CCNE2), cell division cycle 2 (CDC2), G-2 and S-phaseexpressed 1 (GTSE1), helicase, lymphoid-specic (HELLS), antigen identied by monoclonal antibody Ki-67 (MKI67), RAD51 homolog (RAD51), topoisomerase II alpha (TOP2A), and 18S ribosomal RNA (18S) were designed, using the ABI Prism Primer Express software v.2.0 (Applied Biosystems, Foster City, CA). A full list of primer sequences is shown in Table E1. Standard curves for each gene were generated from untreated hMSC. All gene levels for a given sample were normal- ized to 18S levels from the same sample. Data were analyzed using ABI Prism 7000 SDS software. Statistical analysis Data were rst analyzed with analysis of variance (ANOVA), fol- lowed by a Holms t-test to nd the specic difference between irradiation doses (p < 0.05, n = 3). To account for baseline variations between experiments, values from each individual treated sample (0.1 or 1 Gy) were rst normalized to values from the corresponding sham control sample (0 Gy). After normalization, ratio values (radi- ation/control) were pooled among all experiments and analyzed for statistical signicance using a log-transformed one-sample t test (p < 0.05, n = 3). For qRT-PCR, differences were compared between radiation (0.1 or 1 Gy) and control (0 Gy). Data are presented as mean SD. Bioinformatics analysis To evaluate molecular interactions among regulated genes identi- ed by microarray experiments, we performed network and pathway analysis using Ingenuity program (Ingenuity Systems, Redwood City, CA). The dataset contains the signicantly up- or downregu- lated (at least twofold) genes. To select the most signicant networks and pathways from our gene sets, we used the HG- U133A_2 (www.affymetrix.com) as the reference sets to generate sets of over-represented pathways that are known to be affected by X-rays or 56 Fe ions at different doses. A p value for each network and canonical pathway was then calculated. The score of network was displayed as the negative log of the p value (-log 10 (p value)). RESULTS X-rays and 56 Fe ions inhibited proliferation without impairing in vitro osteogenic differentiation process of hMSC Figure 1a demonstrates clear morphologic changes in hMSC over a 7-day period when stimulated with osteogenic differentiation media. These changes became noticeable starting on Day 3 with an increasingly three-dimensional, cobblestone-like morphology potentially indicative of osteo- genic differentiation. However, no obvious morphologic dif- ferences were detected in irradiated cultures when comparing between radiation doses or types at any given time point. Quantitative ALP activity data in Fig. 1b supported this qual- itative observation. The hMSC showed minimal basal ALP activity after 1 day (data not shown), moderate ALP activity after 3 days, and a marked increase in ALP activity after ap- proximately 1 week. Statistical analysis revealed no signi- cant differences between radiation doses for any given time point. The ALP is the early marker for osteogenic differenti- ation; therefore, our results suggested that even after high doses of radiation, the in vitro osteogenic differentiation process of hMSC was not impaired. This was conrmed by Alizarin Red S staining of hMSC in osteogenic media for 3 weeks, showing that calcium phosphate deposition (an os- teogenic maturation marker) remained intact even after high doses of IR (data not shown). Despite the lack of morphologic changes and differences in ALP activity after radiation exposure, the phase contrast Fig. 1. Alkaline phosphatase (ALP) activity of human mesenchy- mal stem cells (hMSC) in osteogenic media after exposure to X-rays or 56 Fe ions. (a) Phase contrast images of hMSC after irradi- ation from Day 0 to Day 7. (b) Bar graph representation of ALP ac- tivity after irradiation. Activity is normalized to total protein concentration. No statistically signicant differences were found between doses (*p < 0.05 using Holms t test, n = 3). X-ray and 56 Fe ions effects on mesenchymal stem cells d K. KURPINSKI et al. 871 images in Fig. 1a did reveal lower conuency of hMSC with 1 Gy X-rays and 56 Fe ions after several days in culture (most noticeable at Days 3 and 7). This suggested that high doses of IR inhibited the proliferation of hMSC, whereas the remain- ing cells differentiated normally. X-rays and 56 Fe ions induced differential effects on cell cycle progression of hMSC Figure 2a displays representative FACS dot plots for distri- butions of G1/G0, S, and G2/M phases in hMSC, analyzed at 24 h after exposure to 56 Fe ions. Figures 2b and 2c showcom- piled dot plot data for X-rays and 56 Fe ions in a dose-depen- dent manner. Low-dose IR (0.1 Gy) did not cause any detectable change in hMSC cell cycle distribution, whereas 1 Gy exposure induced signicant cell cycle modulation. Both X-rays and 56 Fe ions at 1 Gy caused a decrease in the population of S phase cells, suggesting a strong hindrance to entering S-phase. 1 Gy 56 Fe ions also caused a signicant increase in G2/M phase cells, whereas X-rays did not. In ad- dition, 1 Gy 56 Fe ions decreased the fraction of cells in G1/G0 phase, whereas 1 Gy X-rays showed the opposite trend. These ndings suggested that 1 Gy 56 Fe ions hindered the cell cycle progression by impairing both the entry into S phase and mitosis (G2/M), whereas 1 Gy X-rays mainly inhibited the entry into S phase. The impaired mitotic function (G2/M arrest) may account for the reduced number of cells in G1/G0 after exposure to 1 Gy 56 Fe ions. Conversely, the increased number of cells in G1/G0 after exposure to 1 Gy X-rays may be caused by a reduced ability of these cells to enter S phase (G1/G0 arrest), consistent with the decreased number of cells in S phase. We further characterized the kinetics of cell cycle pro- gression of hMSC after exposure to IR. As shown in Fig. E1, there was an apparent G2/M arrest starting at 5 h for 1 Gy X-rays but not for 1 Gy 56 Fe ions. More detailed kinetics for X-rays is shown in Fig. E2. G2/M arrest started 5 h for both 1 Gy and 3 Gy X-rays. For 1 Gy X-rays at 24 h, a portion of cells recovered from G2/M arrest and moved into G1, whereas for 3 Gy X-rays, cells were in the persis- tent G2/M arrest even after 24 h. Although 3 Gy X-rays may give comparable levels of cell killing to 1 Gy 56 Fe ions, their effects on cell cycle arrest were apparently different. The 3 Gy X-rays induced G2/M arrest at early PI intensity b 0 0.2 0.6 0.4 0.8 1.0 * * X-rays F r a c t i o n
o f
c e l l s F r a c t i o n
o f
c e l l s 0.1 Gy 1 Gy 0 Gy G1/G0 G2/M c G1/G0 0 0.2 0.6 0.4 0.8 1.0 S G2/M 0.1 Gy 1 Gy 0 Gy * * * 56 Fe ions S a 0 Gy 0.1 Gy 1 Gy S (1.7%) G1/ G0 (69.6%) G2/M (28.6%) S (13.6%) G1/ G0 (77.9%) G2/M (8.5%) S G1/ G0 (73.3%) G2/M (9.4%) F I T C
( B r d U )
i n t e n s i t y 10 1 10 2 10 0 10 1 10 2 10 0 10 1 10 2 10 0 10 1 10 2 10 0 10 1 10 2 10 0 10 1 10 2 10 0 (17.3%) Fig. 2. Cell cycle distribution analysis of human mesenchymal stemcells (hMSC) exposed to ionizing irradiation. (a) Rep- resentative dot plots of hMSC 24 h after exposure to 56 Fe ions. (b) Bar graph representation of cell cycle distribution 24 h after exposure to X-rays. (c) Bar graph representation of cell cycle distribution 24 h after exposure to 56 Fe ions. *p < 0.05 using Holms t test (n = 3). 872 I. J. Radiation Oncology d Biology d Physics Volume 73, Number 3, 2009 time points (e.g., 5 h) but 1 Gy 56 Fe ions did not. Further- more, the G2/M arrest at 24 h increased only 1.3-fold for 3 Gy X-rays (Fig. E2) while increasing more than 2.0- fold for 1 Gy 56 Fe ions (Fig. 2c) when normalized with their respective 0 Gy controls. Taken together, our results suggested that although both X-rays and 56 Fe ions seriously impaired the entry into S phase, 56 Fe ions were a more potent inhibitor of mitosis than were X-rays. X-rays and 56 Fe ions induced differential p53 but not Erk1/ 2 activation in hMSC Figure 3 shows that p53 was activated as early as 1 min at high dose but not at low dose, and 56 Fe ions induced higher p53 activation than X-rays irrespective of growth media. Our results conrmed that hMSC used in this study have normal p53 functions. In contrast, Erk1/2 was not activated even at high dose for either X-rays or 56 Fe ions. The apparent lack of Erk1/2 activation was consistent with no apparent effects of IR on in vitro osteogenic differentiation of hMSC (Fig. 1 and Discussion). X-rays and 56 Fe ions induced differential transcriptome responses in hMSC The numbers of signicantly regulated genes are summa- rized in Table 1. A total of 81 genes showed common changes between X-rays and 56 Fe ions at 1 Gy. Complete de- tails for each gene are listed in Table E2 (5 h) and Table E3 (24 h). For the 5-h gene set, fewer than 10 genes were up- or downregulated. One of the major changes was the downregu- lation of cyclin E2 (CCNE2), consistent with the cell cycle arrest at this early time point. For the 24-h gene set, visual inspection suggested that many of the common genes were involved in cell cycle, DNA damage response and repair, and other DNA interactions (e.g., MDM2, p21, and Rad51). Some genes showed differential responses when comparing between X-rays and 56 Fe ions (e.g., PIP5K1A, DDX17, and AP1S1). The master osteogenic transcription factor, Runx2, was not signicantly modulated by either X- rays (0.951.14 fold) or 56 Fe ions (1.01.25 fold), based on the averages of its duplicate probe sets. These results sug- gested that X-rays and 56 Fe ions shared common damage re- sponses, as well as having some distinct differential effects on transcriptional regulation of hMSC. qRT-PCR conrmed differential transcriptional regulation of select genes in hMSC We veried nine genes involved in cell cycle and DNAme- tabolism by qRT-PCR (Fig. 4) The results were consistent with the microarray ratios shown in Table E3, and conrmed that IR induced downregulation of the cell cyclerelated genes CCNA2, CCNB1, CCNE2, CDC2, GTSE1, and MKI67 as well as HELLS and TOP2A, which are involved in maintaining and modifying the coiled superstructure of the DNA. Use of RT-PCR analysis also conrmed the down- regulation of RAD51 which is involved in several facets of DNA repair and cell cycle control (18). The higher dose of irradiation had larger overall effects on gene expression for all genes shown in Fig. 4, suggesting that both X-rays and 56 Fe ions acted in a dose-dependent manner for these genes. Several genes such as CCNB1, CCNE2, and HELLS, were signicantly downregulated by 0.1 Gy of 56 Fe ions but showed minimum response to the same dose of X-rays. Furthermore, at a dose of 1 Gy, many genes were signicantly downregulated with both radiation types, but the overall changes were not of the same magnitude. For example, CCNE2 and HELLS showed a greater average decrease for 56 Fe ions than X-rays. Also, RAD51 was signicantly downregulated with 1 Gy 56 Fe ions, but the cor- responding decrease with X-rays was small and not statisti- cally signicant. Bioinformatics analysis shed lights on the underlying mechanism for differential regulation of hMSC by X-rays and 56 Fe ions Figure 5 shows the comparisons of the top ranked path- ways that were enriched in hMSC in response to IR. The sig- nicance of over-representation of the top ranked pathways Fig. 3. Immunoblotting analysis of human mesenchymal stem cells (hMSC) exposed to ionizing radiation. Cells were irradiated with 4 Gy X-rays and lysed at different time points (a) or at 30 min after exposure to different doses of X-rays (b). To compare the effects of X-rays and 56 Fe ions, cells grown in basal media (c) or osteogenic media (d) were exposed to IR and lysed at 3 to 5 h or 24 h after ion- izing radiation (IR). Western blots were probed with anti-pSer15- p53 and anti-pErk1/2 antibodies. Actin was used as the loading control. Table 1. Summary of microarray results: Number of genes upregulated ($2.00 fold) or downregulated (#0.50 fold) in human mesenchymal stem cells 24 h after exposure to ionizing radiation X-rays 56 Fe ions Common between X-rays and 56 Fe ions at 1 Gy 0.1 Gy 1 Gy 0.1 Gy 1 Gy Upregulated 6 29 6 14 3 Downregulated 138 100 23 160 78 Total 144 129 29 174 81 X-ray and 56 Fe ions effects on mesenchymal stem cells d K. KURPINSKI et al. 873 was expressed as -log 10 (p value), and the pathways with -log 10 (p value) above the threshold (p = 0.05; pink line) were considered signicant. Several pathways, including cytoskeleton signaling, IGF-1 signaling, integrin signaling, estrogen receptor signaling, and insulin receptor signaling, were enriched for low-dose (0.1 Gy) X-rays, and were not af- fected signicantly under the other three radiation conditions. In contrast, DNA/RNA metabolism and cell cycle regulation were enriched for both high-dose (1 Gy) X-rays and 56 Fe ions, with more signicant effects for 56 Fe ions. This is con- sistent with a previous study showing that X-rays at low dose preferentially modulate cytoskeleton genes (12). The merged top-two networks shows similarities and dif- ferences between 1 Gy X-rays (Fig. 6a, p < 10 63 10 32 ) and 1 Gy 56 Fe ions (Fig. 6b, p < 10 69 10 57 ). Specically, the 56 Fe ions network included genes involved in the follow- ing: (1) chromosomal DNA replication and DNA repair: DNA-directed polymerases (e.g., POLE2); (2) DNA replica- tion and cell proliferation: DNA replication licensing factors (e.g., MCM2); (3) DNA elongation: replication factors (e.g., RFC5); and (4) cell cycle control: cyclins (e.g., cyclin A) and cyclin kinases/inhibitors (e.g., CDKN1A). The X-rays net- work contained only a subset of the functions found in the 56 Fe ions network, specically genes associated with cell cy- cle control, DNA replication, and cell proliferation. Both net- works contained the DNA replication licensing factors MCM 2, 5, and 10. The MCM proteins are critical for DNA synthe- sis and replication (19). Of particular interest is MCM2, which is required for the entry into S phase and for cell divi- sion. Reduced MCM2 expression was recently shown to re- sult in severe stem/progenitor cell deciency and cancer (20). We showed that MCM2 was downregulated by both 1 Gy X-rays and 56 Fe ions, consistent with the S-phase reduction in both cases, as shown in Fig. 2. The persistent G2/M arrest for 56 Fe ions but not for X-rays may be caused by the more severe downregulation of the four gene classes related to DNA replication and cell proliferation, as shaded in Fig. 6. Most of the genes in merged networks were downregu- lated in hMSC after IR. One of the few upregulated genes was CDKN1A, cyclin-dependent kinase inhibitor 1, or CCNA2 * * * * * * * 0.8 0.6 0.4 0.2 1.0 1.2 m R N A
r a t i o ( i r r a d i a t i o n
/
c o n t r o l ) 0 CCNB1 CCNE2 CDC2 GTSE1 HELLS MKI67 RAD51 TOP2A * * * * * * * * * * * * * * * * * 1 Gy 56 Fe ions 0.1 Gy 56 Fe ions 1 Gy X-rays 0.1 Gy X-rays 1.4 Fig. 4. Verication of select microarray genes using quantitative reverse transcriptionpolymerase chain reaction (qRT- PCR). Bar graph shows relative changes in gene expression 24 h after irradiation when compared to respective 0 Gy control (dotted line at 1.0 represents no change). *p < 0.05 using log-transformed one-sample t test (n = 3), for comparison between irradiated samples and 0 Gy control only. Fig. 5. Comparative pathway enrichments for human mesenchymal stem cells (hMSC) exposed to low and high doses of X-rays and 56 Fe ions. Signicance is expressed as -log10 (p value). The threshold of p = 0.05 are shown as the pink line. The bars that are above the line indicate signicant enrichment for the corresponding radiation exposure group. 874 I. J. Radiation Oncology d Biology d Physics Volume 73, Number 3, 2009 p21 cip1/waf1 (21). P21 prevents phosphorylation of critical cy- clin-dependent kinase substrates and blocking cell cycle pro- gression. We showed that p21 was upregulated by 2.1-fold by 1 Gy 56 Fe ions compared with 1.8-fold by 1 Gy X-rays, rel- ative to sham control (0 Gy) (Table E3). These microarray ndings were consistent with the data showing phenotypic cell cycle arrest (Fig. 2), and p53 activation through Ser15 phosphorylation (Fig. 3). DISCUSSION Ionizing radiation induces a variety of lesions to DNA, in- cluding single-strand breaks, double-strand breaks, and base damages, depending on radiation quality and dose. 56 Fe ions induce denser clustering of DNA lesions than X-rays, in- creasing the likelihood of failed repairs. In addition, DSBs in close proximity can misrejoin to give rise to chromosome aberrations and permanent genetic damage (22). Our results indicated that these biophysical events were associated with perturbed DNA synthesis, DNA replication, and cell cycle progression. Consistent with this view, drastic G2/M arrest has been reported for other cell types, for example, Chinese hamster cells after irradiation with other heavy-ion beams ( 13 C, 20 Ne, 40 Ar) (23). We showed that the majority of hMSC exit G2/M arrest at 24 h after 1 Gy X-rays, whereas a large portion is still in the persistent G2/M arrest from 1 Gy 56 Fe ions, causing more severe damage. We reason that this may be caused by more efcient DNA damage re- sponse and repair of hMSC after X-ray exposure. It has been observed clinically that hMSC obtained from patients after bone marrow transplantation originate from host tissue, suggesting hMSC are inherently radioresistant in vivo (24, 25). Indeed, the radioresistance to g-rays of human bone mar- rowderived, clonally expanded hMSC has been reported (26). On the other hand, we observed minimum effects of 0.1 Gy 56 Fe ions on hMSC. This was consistent with the fact that the cell nuclei of hMSC are large so that most cells (75%) are hit at the dose of 0.1 Gy. For other smaller cells, the majority of cells will not be hit but those that are hit get a much higher dose than 0.1 Gy. For example, if the cells are smaller and therefore only 25% of the cells are hit, the hit cells will receive 0.4 Gy and the rest close to 0 Gy, for an average of 0.1 Gy. In our case, the hit cells will receive just a bit more than 0.1 Gy (0.133 Gy), which is still a low dose. Our study linked changes in gene expression at single pro- tein and genome-wide levels with changes in cellular re- sponses after IR. Phosphorylation of p53 at Ser15 is a key indicator of DSBresponse and repair, and results in increased activity and stability of p53 (27). The major p53 responses in- cluded induction of growth arrest in the G1 and/or G2 phases of the cell cycle through p21 and the induction of apoptosis (28, 29). We showed that phosphorylation of p53 occurred within minutes in response to IR. Our transcriptomic studies showed that hMSC coordinated the transcriptional networks of DNA replication, DNA strand elongation, and DNA bind- ing/transferase activity for G2/Mand G1/G0 arrests 24 h after a high dose of IR. Therefore, it is tempting to speculate that the initial cellular response (minutes to hours) to IR is at the posttranslational level (e.g., phosphorylation of p53), whereas subsequent effects (hours to days) are mostly at the transcriptional and translational levels after activation Fig. 6. Comparison of interaction networks for genes regulated in human mesenchymal stem cells (hMSC) exposed to 1-Gy X-rays (a) and 1 Gy 56 Fe ions (b). Upregulated genes are coded in pink, whereas downregulated genes are coded in green. Genes in shaded regions are four different function groups associated with cell cycle control, DNA strand elongation, DNA binding/transferase activity, and replication fork formation. X-ray and 56 Fe ions effects on mesenchymal stem cells d K. KURPINSKI et al. 875 of p53 and other transcription factors. Using a phosphopro- teomics strategy to identify ATM/ATR substrates 1 h after IR, it was demonstrated that interconnecting networks coor- dinate during DNA damage responses (30). The work chal- lenged the old paradigm of distinct DNA repair pathways responding individually to different lesions after IR and val- idated the merit of a systems biology approach. Our results have implications for space exploration. We found that IR, even 1 Gy 56 Fe ion irradiation, did not alter the in vitro osteogenic differentiation process of hMSC. Runx2, the master osteogenic transcription factor, is regu- lated by pErk1/2 (2, 31). Lack of changes in osteogenic dif- ferentiation was consistent with our observation that Runx2 expression and Erk1/2 activation did not change in hMSC af- ter exposure to either X-rays or 56 Fe ions. Our results raise the possibility of osteosarcoma formation after high doses of 56 Fe ion irradiation. Osteosarcoma develops during physio- logical growth froman expanding cell population, suggesting that the transformed cell is a bone-forming progenitor (32). MSCs are physiological osteogenic progenitors, and osteo- sarcoma may originate from MSCs. In vivo, hMSC differen- tiate into osteoblasts which eventually become post-mitotic bone cells. Because irradiated hMSC do not lose their in vitro osteogenic differentiation potentials, we speculate that repo- pulation of osteoblast/bone cells by these damaged MSCs in vivo might trigger osteosarcoma. In fact, osteosarcoma induction by high-LET radiation has been documented (3336). We realized that the hMSC used in this study were a subset of multipotent bone morrow stromal cells and might contain heterogeneous populations of stem cells and lineage-committed cells. Furthermore, IR may have effects on in vivo bone formation because both proliferation and osteogenic differentiation of MSC in vivo contribute to the process. For example, both direct and indirect effects of high-LET radiations on bone loss in rodents have been reported (37, 38). Future studies of radioresponses of MSCs in vivo using murine models and also along other lineages (e.g., to adipocytes and chondrocytes) may provide more insights into the origin of bone loss during space travel, as well as cancer risks fromexposure to galactic cosmic rays (39). CONCLUSION In conclusion, we showed both phenotypically and mech- anistically the differential effects of X-rays and high-energy 56 Fe ions on hMSC. Our results will assist the Department of Energy with risk estimations for low-dose radiation, and the National Aeronautics and Space Admininstration with HZE risk estimations and countermeasures for space exploration. REFERENCES 1. Grove JE, Bruscia E, Krause DS. Plasticity of bone marrow- derived stem cells. Stem Cells 2004;22:487500. 2. Ducy P. Cbfa1: A molecular switch in osteoblast biology. Dev Dyn 2000;219:461471. 3. Beausejour CM, Campisi J. Ageing: Balancing regeneration and cancer. Nature 2006;443:404405. 4. Rubio D, Garcia-Castro J, Martin MC, et al. Spontaneous human adult stem cell transformation. Cancer Res 2005;65: 30353039. 5. Serakinci N, Guldberg P, Burns JS, et al. Adult human mesen- chymal stem cell as a target for neoplastic transformation. Oncogene 2004;23:50955098. 6. Hall EJ, Hei TK. Genomic instability and bystander effects in- duced by high-LET radiation. Oncogene 2003;22:70347042. 7. Morgan WF. Is there a common mechanismunderlying genomic instability, bystander effects and other nontargeted effects of exposure to ionizing radiation? Oncogene 2003;22:70947099. 8. Schimmerling W, Cucinotta FA, Wilson JW. Radiation risk and human space exploration. Adv Space Res 2003;31:2734. 9. Nelson GA. Fundamental space radiobiology. Gravit Space Biol Bull 2003;16:2936. 10. Birrell GW, Brown JA, Wu HI, et al. Transcriptional response of Saccharomyces cerevisiae to DNA-damaging agents does not identify the genes that protect against these agents. Proc Natl Acad Sci U S A 2002;99:87788783. 11. Yin E, Nelson DO, Coleman MA, et al. Gene expression changes in mouse brain after exposure to low-dose ionizing radiation. Int J Radiat Biol 2003;79:759775. 12. Ding LH, Shingyoji M, Chen F, et al. Gene expression proles of normal human broblasts after exposure to ionizing radia- tion: A comparative study of low and high doses. Radiat Res 2005;164:1726. 13. Sokolov MV, Smirnova NA, Camerini-Otero RD, et al. Micro- array analysis of differentially expressed genes after exposure of normal human broblasts to ionizing radiation from an external source and from DNA-incorporated iodine-125 radionuclide. Gene 2006;382:4756. 14. Wang D, Park JS, Chu JS, et al. Proteomic proling of bone marrowmesenchymal stemcells upon transforming growth fac- tor b stimulation. J Biol Chem 2004;279:4372543734. 15. Kurpinski K, Chu J, Hashi C, et al. Anisotropic mechanosens- ing by mesenchymal stem cells. Proc Natl Acad Sci US A 2006; 103:1609516100. 16. Zeitlin C, Heilbronn L, Miller J. Detailed characterization of the 1087 MeV/nucleon ion-56 beam used for radiobiology at the alternating gradient synchrotron. Radiat Res 1998;149: 560569. 17. Groesser T, Chun E, Rydberg B. Relative biological effective- ness of high-energy iron ions for micronucleus formation at low doses. Radiat Res 2007;168:675682. 18. Baumann P, West SC. Role of the human RAD51 protein in homologous recombination and double-stranded-break repair. Trends Biochem Sci 1998;23:247251. 19. Maiorano D, Lutzmann M, Mechali M. MCM proteins and DNA replication. Curr Opin Cell Biol 2006;18:130136. 20. Pruitt SC, Bailey KJ, Freeland A. Reduced Mcm2 expression results in severe stem/progenitor cell deciency and cancer. Stem Cells 2007;25:31213132. 21. Xiong Y, Hannon GJ, Zhang H, et al. p21 is a universal inhibitor of cyclin kinases. Nature 1993;366:701704. 22. Rydberg B, Cooper B, Cooper PK, et al. Dose-dependent misrejoining of radiation-induced DNA double-strand breaks in human broblasts: Experimental and theoretical study for high- and low-LET radiation. Radiat Res 2005;163: 526534. 23. Lucke-Huhle C, Blakely EA, Chang PY, et al. Drastic G2 arrest in mammalian cells after irradiation with heavy-ion beams. Radiat Res 1979;79:97112. 24. Rieger K, Marinets O, Fietz T, et al. Mesenchymal stem cells remain of host origin even a long time after allogeneic 876 I. J. Radiation Oncology d Biology d Physics Volume 73, Number 3, 2009 peripheral blood stem cell or bone marrow transplantation. Exp Hematol 2005;33:605611. 25. Dickhut A, Schwerdtfeger R, Kuklick L, et al. Mesenchymal stemcells obtained after bone marrowtransplantation or periph- eral blood stem cell transplantation originate from host tissue. Ann Hematol 2005;84:722727. 26. Chen MF, Lin CT, Chen WC, et al. The sensitivity of human mesenchymal stem cells to ionizing radiation. Int J Radiat Oncol Biol Phys 2006;66:244253. 27. Fei P, El-Deiry WS. P53 and radiation responses. Oncogene 2003;22:57745783. 28. Waldman T, Kinzler KW, Vogelstein B. p21 is necessary for the p53-mediated G1 arrest in human cancer cells. Cancer Res 1995;55:51875190. 29. Bunz F, Dutriaux A, Lengauer C, et al. Requirement for p53 and p21 to sustain G2 arrest after DNA damage. Science 1998;282: 14971501. 30. Matsuoka S, Ballif BA, Smogorzewska A, et al. ATM and ATR substrate analysis reveals extensive protein networks responsive to DNA damage. Science 2007;316:11601166. 31. Franceschi RT, Xiao G. Regulation of the osteoblast-specic transcription factor, Runx2: Responsiveness to multiple signal transduction pathways. J Cell Biochem 2003;88:446454. 32. Lengner CJ, Steinman HA, Gagnon J, et al. Osteoblast differen- tiation and skeletal development are regulated by Mdm2-p53 signaling. J Cell Biol 2006;172:909921. 33. Koshurnikova NA, Gilbert ES, Sokolnikov M, et al. Bone can- cers in Mayak workers. Radiat Res 2000;154:237245. 34. Rosemann M, Lintrop M, Favor J, et al. Bone tumorigenesis in- duced by alpha-particle radiation: Mapping of genetic loci inu- encing predisposition in mice. Radiat Res 2002;157:426434. 35. Oghiso Y, Yamada Y. The specic induction of osteosarco- mas in different mouse strains after injections of 239Pu citrate. J Radiat Res (Tokyo) 2003;44:125132. 36. Virtanen A, Pukkala E, Auvinen A. Incidence of bone and soft tissue sarcoma after radiotherapy: A cohort study of 295,712 Finnish cancer patients. Int J Cancer 2006;118:10171021. 37. Hamilton SA, Pecaut MJ, Gridley DS, et al. Amurine model for bone loss from therapeutic and space-relevant sources of radia- tion. J Appl Physiol 2006;101:789793. 38. Willey JS, Grilly LG, Howard SH, et al. Bone architectural and structural properties after 56Fe26+ radiation-induced changes in body mass. Radiat Res 2008;170:201207. 39. Cucinotta FA, Durante M. Cancer risk fromexposure to galactic cosmic rays: Implications for space exploration by human beings. Lancet Oncol 2006;7:431435. X-ray and 56 Fe ions effects on mesenchymal stem cells d K. KURPINSKI et al. 877