SEARCH

SEARCH BY CITATION

Keywords:

  • Mesenchymal stromal cells;
  • Hypoxia;
  • Irradiation;
  • DNA damage response;
  • DNA repair

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Author Contributions
  10. Disclosure of Potential Conflicts of Interest
  11. References
  12. Supporting Information

Mesenchymal stromal cells (MSCs) are radioresistant bone marrow progenitors that support hematopoiesis and its reconstitution following total body irradiation. MSCs reside in hypoxic niches within the bone marrow and tumor microenvironments. The DNA damage response (DDR) represents a network of signaling pathways that enable cells to activate biological responses to DNA damaging agents. Hypoxia-mediated alterations in the DDR contribute to the increased radioresistance of hypoxic cancer cells, limiting therapeutic efficacy. The DDR is important in mediating mouse MSC radioresistance. However, the effects of hypoxia on MSC radioresistance are currently unknown. In this report, hypoxia was found to (a) increase MSC proliferation rate and colony size; (b) increase long-term survival post-irradiation (IR), and (c) improve MSC recovery from IR-induced cell cycle arrest. DNA double-strand break (DSB) repair in MSCs was upregulated in hypoxia, accelerating the resolution of highly genotoxic IR-induced DNA DSBs. In addition, HIF-1α was found to contribute to this enhanced DSB repair by regulating (a) the expression of DNA ligase IV and DNA-PKcs and (b) Rad51 foci formation in response to DNA DSBs in hypoxic MSCs. We have demonstrated, for the first time, that hypoxia enhances mouse MSC radioresistance in vitro. These findings have important implications for our understanding of MSC functions in supporting allogeneic bone marrow transplantation and in tumorigenesis. Stem Cells 2014;32:2188–2200


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Author Contributions
  10. Disclosure of Potential Conflicts of Interest
  11. References
  12. Supporting Information

The bone marrow is essential for adult hematopoiesis and for regulating hemostasis in the periphery. These key processes are tightly controlled by a specialized microenvironment known as the hematopoietic stem cell (HSC) niche that consists of a network of immune and stromal cell types, including mesenchymal stromal cells (MSCs). MSCs are multipotent, adherent progenitors that generate individual colonies in vitro known as colony-forming unit fibroblasts [1]. Mouse MSCs are characterized phenotypically as CD45 Ter119 CD44+ CD90+ CD29+ [2-4]. MSCs provide physical support for HSCs and regulate hematopoiesis by controlling HSC self-renewal, differentiation, and retention [5-8]. MSCs are also an integral component of the tumor microenvironment [9, 10]. Signals from the tumor mass stimulate MSC exodus from the bone marrow to the tumor where they can develop into cancer-associated fibroblasts (CAFs) [11-13]. CAFs are a heterogeneous population of stromal cells distinguished by the expression of specific fibroblast markers, including α-smooth muscle actin and fibroblast activation protein. CAFs can arise from tissue adjacent to the primary tumor site and from the bone marrow [9-14]. They support tumorigenesis in multiple ways including (a) physical tumor support; (b) secretion of growth (e.g., epidermal growth factor) and proangiogenic factors (e.g., vascular endothelial-derived growth factor (VEGF)); (c) suppression of antitumor immune responses; and (d) inhibition of tumor cell apoptosis [9, 11, 12, 14].

MSCs are generally expanded in vitro at the inhaled ambient oxygen concentration of 21% O2, commonly referred to as “normoxia.” However, 21% O2 is physiologically hyperoxic for MSCs since the oxygen concentration of the bone marrow ranges from approximately 1% to 8% O2, referred to as hypoxia or physiologic normoxia [15, 16]. In addition, hypoxic regions (estimated to range between approximately 7 to <1% O2) are commonly found within solid tumors due to abnormal tumor vasculature [17-20].

Cellular adaptation to hypoxia is primarily mediated by the transcription factor, hypoxia inducible factor-1 (HIF-1). HIF-1 consists of a constitutively expressed HIF-1β-subunit and an oxygen-regulated HIF-1α subunit (or its paralogs HIF-2α and HIF-3α). HIF-1α is stabilized in hypoxia due to inhibited proteasomal degradation, a pathway involving prolyl hydroxylase domain-containing proteins, factor inhibiting HIF, and an E3 ubiquitin ligase termed product of the von Hippel-Lindau tumor suppressor. This enables dimerization with β subunits in the nucleus and downstream activation of hypoxia-responsive genes [19, 21-23]. Intratumoral hypoxia and HIF-1α overexpression are associated with increased malignancy and metastatic potential and with poor patient outcome [17, 18, 24]. This prognosis is highly correlated with the increased therapeutic resistance of hypoxic cancer cells [17, 18, 24, 25].

Radiotherapy destroys cells by inducing extensive DNA damage, particularly highly genotoxic DNA double-strand breaks (DNA DSBs) [26, 27]. This genotoxic stress activates a network of interacting signaling pathways, collectively termed the DNA damage response (DDR). Conceptually, the DDR consists of (a) sensor proteins which recognize damaged DNA; (b) transducer proteins which amplify the DNA damage signal, and (c) effector proteins which execute appropriate biological response(s) such as DNA repair, transient delays in cell cycle progression (i.e., checkpoints), transcription, and apoptosis [28, 29]. Cellular exposure to severe hypoxia (<1% O2) can activate the DDR independently of DNA damage, indicating an important role for the DDR in adaptation to low oxygen environments [30-35]. There is mounting evidence that hypoxia-mediated modulation of the DDR contributes to the resistance of hypoxic cancer cells to DNA damaging agents, including ionizing radiation (IR) [36, 37].

MSCs are radioresistant and can support hematopoietic reconstitution following total body irradiation [38-43]. We have previously shown that execution of the DDR contributes to the radioresistance of mouse MSCs [44]. However, whether hypoxia influences the radioresistance of MSCs is currently unknown. Herein, we have compared the effects of normoxia (21% O2) and hypoxia (5% and 2% O2) on mouse MSC radioresistance using primary bulk MSCs, two clonal authentic mouse MSC lines, MS5 and ST2 [44], and the mouse CD4+ CD8+ thymocyte line, ST4.5, as a radiosensitive control. For the first time, we demonstrate that hypoxia enhances mouse MSC radioresistance and this is due to alterations in their DDR that occur in a HIF-1α dependent manner.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Author Contributions
  10. Disclosure of Potential Conflicts of Interest
  11. References
  12. Supporting Information

Cell Culture and Treatment

C57BL/6 primary bulk mouse MSCs were provided by Prof. Matthew D. Griffin (REMEDI, NUI Galway) and cultured from passage 6 to 8 as previously described [44, 45]. Bone marrow was isolated from Fox Chase severe combined immuno-deficiency (SCID) Beige mice, provided by Dr. Maoija Xu (REMEDI, NUI Galway), by crushing femurs in sterile Dulbecco's modified Eagle's medium (DMEM) using a mortar and pestle. The resulting bone marrow suspension was filtered through 70 µm nylon mesh, centrifuged, and resuspended in ammonium-chloride-potassium buffer for 5 minutes to lyse erythrocytes. Following centrifugation, the cells were resuspended in growth medium and maintained at passage 0 in 10 cm culture dishes for 4 weeks in 5% O2, with frequent media changes to remove contaminating hematopoietic cells. MS5 and ST2 MSC lines were provided by Prof. Antonius Rolink (Department of Biomedicine, University of Basel) and were recloned prior to use. All MSC cultures were CD45, CD44+, and CD29+ as previously described [44, 45]. ST4.5 CD4+CD8+ thymocyte line was provided by Dr. Anne Wilson (Ludwig Institute for Cancer Research, Lausanne) and J774A.1 monocyte line by Prof. Benjamin Bradley (Department of Orthopaedics, University of Bristol). All cell types were continuously cultured in humidified incubators at 37°C containing 21% O2 (normoxia) or 2%/5% O2 (hypoxia) for at least 1 week prior to experimentation. Cells were γ-irradiated at the indicated doses using a Mainance Millennium Sample Irradiator containing a 137Cs source at a dose rate of approximately 102 cGy/minute.

Growth Curve Analysis

Cells were seeded into six-well plates (Nunc) at a concentration of 30,000 (ST4.5) or 50,000 (MSCs) cells per well. Individual cultures were harvested daily for 5 (ST4.5) or 7 days (MSCs and J774A.1), and cell counts were performed in duplicate in a hemocytometer using trypan blue exclusion. The average doubling time in normoxia was 19 hours (MS5), 18 hours (ST2), 29 hours (bulk MSCs), 12 hours (ST4.5), and 14 hours (J774A.1) and in hypoxia was 16 hours (MS5 and ST2), 20 hours (bulk MSCs), 14 hours (ST4.5), and 14 hours (J774A.1).

Clonogenic Survival Assay

MSCs were irradiated at 2–10 Gy and seeded into six-well plates (Nunc, Thermo Scientific UK, http://www.thermoscientific.com) at a concentration of 200 (MSC lines), 500 (J774A.1) or 1,000 (bulk MSCs) cells per well. Cells were cultured for 7 (MSC lines and J774A.1) or 14 (bulk MSCs) days in 2%, 5% or 21% O2 until colonies were clearly visible. Colonies were stained with Coomassie Blue solution and counted as previously described [44]. All colony images are representative of one of three independent experiments. Nonadherent ST4.5 thymocytes were irradiated (2–10 Gy) and seeded into six-well plates (Nunc) at a concentration of 50,000 cells per well, and cell numbers were counted 5 days post-IR using trypan blue exclusion. The percentage survival of each cell type was determined by normalizing the number of colonies per cells generated in irradiated cultures to the number of colonies per cells generated in control cultures. The plating efficiency in normoxia was 65% (MS5), 50% (ST2), 26% (bulk MSCs), and 30% (J774A.1) and in hypoxia was 70% (MS5), 64% (ST2), 31% (bulk MSCs), and 33% (J774A.1).

Flow Cytometry

Cell Cycle Analysis Using 5′-Bromo-deoxyuridine–Propidium Iodide Staining

Cells were labeled for 15 minutes (ST4.5) or 45 minutes (MS5) with 25 µM 5′-bromo-deoxyuridine (BrdU) (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), washed in PBS, and resuspended in growth medium. Cells were harvested at the indicated time points post-irradiation (10 or 0.5 Gy), fixed in ice-cold 70% ethanol, and stained with anti-BrdU and fluorescein isothiocyanate (FITC)-conjugated anti-mouse IgG antibodies and propidium iodide (PI)/RNase staining buffer (BD Biosciences, San Diego, CA, http://www.bdbiosciences.com) as previously described [44]. The progression of G1, S, and G2/M cells through the cell cycle was analyzed by measuring the percentage of cells in each phase until 36 hours post-IR.

G2/M Checkpoint Analysis Using Phospho-Histone H3 Ser10-PI Staining

MS5 and ST4.5 cells were harvested at the indicated time points post-IR (10 or 0.5 Gy), washed in PBS, and fixed in ice-cold 70% ethanol. Following permeabilization in PBS/0.25% Triton X-100, cells were sequentially stained with anti-phospho-histone H3(Ser10) and FITC-conjugated anti-rabbit IgG antibodies for 2 hours and 30 minutes at room temperature, respectively, separated by a washing step in PBS/1% bovine serum albumin (BSA). Cells were then resuspended in PI/RNase staining buffer (BD Biosciences), and mitotic index was analyzed up to 24 hours post-IR.

Analysis of Apoptosis Using Annexin-V–PI Staining

MS5 and ST4.5 cells were harvested at the indicated time points post-IR (10 or 0.5 Gy), incubated at 37°C for 5 minutes, then resuspended in Annexin-V binding buffer, and stained with FITC-conjugated Annexin-V solution and PI as previously described [44].

In Vivo Homologous Recombination Assay

MS5 cells were stably transfected with DR-GFP HR reporter [46, 47] provided by Prof. Ciaran Morrison (Centre for Chromosome Biology, NUI Galway), using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) according to manufacturer's instructions. Individual MS5 clones were selectively expanded in Puromycin (10 µg/ml). Positive clones were screened by amplifying a 500-bp fragment encompassing the Sce-gfp sequence with 1 µg genomic DNA using the following polymerase chain reaction (PCR) primers: forward: 5′-GTGAGCAAGGGCGAGGAG-3′ and reverse: 5′-ATCTTGAAGTTCACCTTGAT-3′ (Supporting Information Fig. 5A, 5B). DSBs were generated by transfecting cells with I-Sce I endonuclease expressing vector, pCBASce, and homologous recombination (HR) efficiency was measured 2 days later by analyzing green fluorescent protein (GFP) expression using flow cytometry (Supporting Information Fig. 5A) [46, 47]. Clone MS5 HR4.4 was selected based on screening of transfected clones for efficient GFP expression and expanded in 21% and 5% O2 prior to transfection.

All fluorescence activated cell sorted (FACS) samples were analyzed using BD FACS Canto and Accuri C6 flow cytometers (BD Biosciences) and FlowJo software (TreeStar Inc., OR). All flow cytograms shown are representative of one of three independent experiments.

Western Blotting

Whole cell extracts were prepared from harvested control or irradiated cells at the indicated time points post-IR as previously described [44]. Samples were separated using 4%–15% SDS-PAGE gels and transferred to nitrocellulose membranes. Chemiluminescence was detected using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, UK, http://www.thermoscientific.com) and medical X-ray film (Konica Minolta, Inc. Lijnden, The Netherlands, http://www.konicaminolta.com). β-Tubulin expression was used as a loading control, and all Western blots shown are representative of one of three independent experiments.

Immunofluorescence Staining

Cells were cultured on glass coverslips in 21% or 5% O2 for at least 24 hours prior to irradiation. All cultures were fixed in 4% paraformaldehyde (Sigma-Aldrich), permeabilized in 0.1% Triton-X 100 solution, and nuclei were stained for γ H2AX (γ-H2AX) and Rad51 IR-induced foci (IRIF) as previously described [44]. All images were captured using ×100 magnification on an IX51 Olympus fluorescent microscope using Olympus xcellence software (Olympus, Hamburg, Germany). The number of γ-H2AX and Rad51 IRIF per nucleus was manually quantified blind in a total of 50 cells per time point in each experiment. All images shown are representative of one of three independent experiments.

In Vitro DNA DSB End-Joining Assay

Briefly, nuclear extracts were prepared according to manufacturer's instructions (Active Motif, La Hulpe, Belgium, http://www.activemotif.com). The pmaxFP-Green-N (pMAX-GFP) plasmid (Amaxa) was linearized by cleavage with XmnI. End-joining efficiency of nuclear extracts prepared from MS5 cells cultured in 21% O2 and 5% O2 and of passage 0 bulk MSC cultures (5% O2) derived from Fox Chase SCID Beige mouse femurs was analyzed as previously described [44, 48]. Ligated DNA fragment intensity was quantified using ImageJ software. Ligated DNA fragment intensity values detected in normoxic and hypoxic MS5 samples were normalized to the background intensity detected in SCID bulk samples for three independent experiments.

Small Interfering RNA-Mediated Knockdown of HIF-1α Expression

MSC lines cultured in 5% O2 in Opti-MEM (Invitrogen) were transfected with 5 nM Silencer select negative control mouse small interfering RNA (siRNA) or with 50 nM Silencer select predesigned HIF-1α siRNAs (s67532 and n414564) (Ambion, Austin, TX, http://www.ambion.com) using Oligofectamine according to manufacturer's instructions (Invitrogen). High glucose DMEM (Sigma-Aldrich) supplemented with 30% fetal calf serum was added to cultures 3 hours post-transfection. Irradiation and harvesting of transfected cultures for required experiments was performed 2 days post-transfection. Approximately, 60%–80% HIF-1α knockdown efficiency was achieved in MSC lines (Fig. 5E).

Statistical Analysis

Statistical significance between (a) normoxic (21% O2) and hypoxic (5% or 2% O2) MSC samples and (b) MSCs transfected with scrambled or HIF-1α siRNA in the appropriate experiments was determined using two-way ANOVA with Bonferroni post-test on Graph Pad Prism 6 software. Statistical significance between (a) % GFP positive MS5 HR4.4 cells in normoxia and hypoxia and (b) ligated fragment intensity in normoxic and hypoxic MS5 samples was determined using Student's t test. Error bars in all figures represent mean ± SD, n = 3. p < .05 was considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Author Contributions
  10. Disclosure of Potential Conflicts of Interest
  11. References
  12. Supporting Information

Hypoxia Enhances Mouse MSC Radioresistance In Vitro

Growth curves of MSC lines (MS5 and ST2), primary bulk C57BL/6 MSCs, and CD4+ CD8+ thymocytes (ST4.5 cell line) cultured in normoxia and hypoxia were first analyzed to determine the effects of hypoxia on cell growth. In agreement with previous findings [49, 50], MSC growth rate was measurably increased in hypoxia (5% and 2% O2) (Fig. 1A, 1B; Supporting Information Fig. 1A). For example, on day 5, 136 × 104 MS5 cells, 134 × 104 ST2 cells, and 25 × 104 bulk MSCs were detected in 21% O2 compared with 160 × 104 MS5 cells (p < .01), 152 × 104 ST2 cells (p < .05), and 46 × 104 bulk MSCs in 5% O2 (p < .001) (Fig. 1A, 1B). Unlike MSCs, the growth of ST4.5 thymocytes was unaltered in hypoxia (Supporting Information Fig. 1B). Clonogenic survival assays revealed a twofold to threefold increase in MSC survival post-IR in hypoxia compared with MSCs cultured in normoxia (Fig. 1C, 1D; Supporting Information Fig. 1C). MSC colony size was also increased in hypoxia under control and irradiated conditions (Fig. 1E, 1F). These results demonstrate that MSC radioresistance was enhanced in hypoxia. In contrast to MSCs, the radiosensitivity of ST4.5 thymocytes was increased in hypoxia and in a dose-dependent manner (Fig. 1C). The effects of hypoxia on the moderately radioresistant J774A.1 monocyte line was also studied [44]. Similar to ST4.5 thymocytes, hypoxic conditions did not enhance the growth or the radioresistance of J774A.1 monocytes (Supporting Information Fig. 1D, 1E).

image

Figure 1. Hypoxia enhances MSC survival post γ-IR. Growth curves of (A) MS5 and ST2 and (B) bulk MSCs in 21%/5% O2. Radiation survival curves of (C) MS5, ST2, and ST4.5 cells and (D) bulk MSCs in 21%/5% O2. Images of (E) MS5 and (F) bulk MSC colonies generated in survival assays following 0/4/8 Gy irradiation. *, MS5/bulk MSC samples; #, ST2 samples; */#, p < .05; **/##, p < .01; ***, p < .001 compared with normoxic samples, two-way ANOVA with Bonferroni post-test. Abbreviations: IR, irradiation; MSC, mesenchymal stromal cell.

Download figure to PowerPoint

Unlike MSC lines, irradiated bulk MSCs exhibit a highly variable DDR, likely due to the inherent heterogeneity of primary uncloned MSC cultures [44]. MS5 and ST2 cells respond to irradiation in a similar manner to bulk MSCs [44]. Furthermore, the effects of hypoxia on the radioresistance of these MSC lines were comparable to that of bulk MSCs (Fig. 1A–1F). Therefore, for detailed molecular investigation of the effects of hypoxia on the DDR of MSCs, we focused upon cloned MS5 and ST2 MSC lines. Since culture in 5% and 2% O2 had similar effects on MSC radioresistance (Fig. 1C; Supporting Information Fig. 1C), subsequent experiments were primarily performed in 5% O2. In addition, as our previous characterization of irradiated mouse MSCs in normoxia was largely performed using 10 Gy irradiation [44], 10 Gy was used to compare the DDR of normoxic and hypoxic MSCs throughout this study.

Hypoxia Accelerates Mouse MSC Recovery from IR-Induced Cell Cycle Arrest

Increased MSC survival and colony size post-IR in hypoxia suggested that hypoxia could affect DNA damage checkpoints activated by MSCs. Irradiated mouse MSCs activate robust intra-S-phase and G2/M checkpoints in normoxia [44]. Here, we further characterized the G2/M checkpoint response of irradiated MSCs by analyzing mitotic index using intracellular phosphorylated histone H3Ser10 (pH3S10) staining by flow cytometry [51] (Fig. 2A, 2B). The absence of mitotic (pH3S10 positive) cells at early time points (2–8 hours) post-IR confirmed that irradiated MS5 cells activated the G2/M checkpoint in both normoxia and hypoxia (Fig. 2A, 2B). However, recovery from G2/M arrest occurred earlier in hypoxia, indicated by the presence of mitotic MS5 cells at 8 hours post-IR in hypoxia (Fig. 2A, black arrowhead, lower 8-hour panel; Fig. 2B), which were absent until 12 hours post-IR in normoxia (Fig. 2A, 2B).

image

Figure 2. Hypoxia accelerates mesenchymal stromal cell (MSC) recovery from IR-induced cell cycle arrest. Cytograms of MS5 cells stained for (A) histone H3Ser10 phosphorylation (pH3S10) or (C) BrdU incorporation and DNA content (propidium iodide) in 21%/5% O2 0–24 hours post 10 Gy irradiation. Black arrowheads indicate mitotic (A) or BrdU labeled G1/early S phase MS5 cells (C). Black box indicates BrdU labeled S phase population. Average % (B) mitotic index and (D) BrdU labeled G1/early S phase cells 0–36 hours post-IR. (E): Total p53 and β-tubulin expression levels in normoxic and hypoxic MSCs 0–4 hours post-IR. *, p < .05 compared with normoxic samples, two-way ANOVA with Bonferroni post-test. Abbreviations: BrdU, 5′-bromo-deoxyuridine; IR, irradiation; pH3S10, phosphorylated histone H3Ser10.

Download figure to PowerPoint

Due to the intrinsic radiosensitivity of ST4.5 thymocytes (Fig. 1A) [44], 0.5 Gy irradiation (equitoxic to 10 Gy in MSCs) was used to analyze DNA damage checkpoints in this cell line. In contrast to MS5 cells, irradiated ST4.5 thymocytes did not arrest at the G2/M checkpoint in either normoxia or hypoxia, indicated by the continued presence of mitotic cells post-IR (Supporting Information Fig. 2A, 2B). The reason(s) for this G2/M checkpoint deficiency are unknown but may be linked to the very low levels of ATM and Chk2, proteins involved in the G2/M checkpoint, expressed in this cell type [44, 51, 52].

To analyze the intra-S-phase/G2 checkpoint, the cell cycle progression of irradiated BrdU labeled MS5 cells (S phase population) (Fig. 2C, black box in upper left panel) was analyzed using a flow cytometry-based BrdU incorporation assay [44, 51]. Following irradiation, BrdU labeled MS5 cells accumulated as a cohort in late S/G2 under both oxygen tensions (Fig. 2C; 8-hour time point) and the proportions of these BrdU labeled populations remained largely unchanged up to 8 hours post-IR whereas the proportions of control BrdU labeled populations steadily declined (Supporting Information Fig. 3A). Cell cycle progression of unirradiated (control) BrdU labeled MS5 populations was comparable in normoxia and hypoxia (Supporting Information Fig. 3A; Fig. 2D). These findings demonstrate that DNA damage checkpoints were intact in hypoxic MSCs and were activated in a similar manner under both oxygen tensions (Fig. 2A–2D; Supporting Information Fig. 3A).

Interestingly, in comparison with normoxia, significantly increased proportions of BrdU labeled G1/S cells were present at 12 hours (7.6% in 21% O2 vs. 22.4% in 5% O2; p < .05) and at 24 hours (18.4% in 21% O2 vs. 32.6% in 5% O2; p < .05) post-IR in hypoxia (Fig. 2C, black arrowheads, 2D). This indicated that MSC recovery from IR-induced cell cycle arrest was accelerated in hypoxia. In contrast to MS5 cells, irradiated BrdU labeled ST4.5 thymocytes were recovered from cell cycle arrest at similar rates under both oxygen tensions (Supporting Information Fig. 2C, 2D). Similar proportions of apoptotic (Annexin-V positive) MS5 and ST4.5 cells were detected post-IR in normoxia and hypoxia (Supporting Information Fig. 3B, 3C). As previously shown in normoxia [44], p53 was transiently stabilized in irradiated MSCs (Fig. 2E) whereas it was maintained in irradiated ST4.5 thymocytes (Supporting Information Fig. 3D) and similar kinetics occurred in hypoxia (Fig. 2E; Supporting Information Fig. 3D).

DNA DSB Repair Is Enhanced in Mouse MSCs Exposed to Hypoxia

Improved MSC recovery from IR-induced cell cycle arrest in hypoxia, in combination with no effect on the extent of apoptosis or on p53 stabilization kinetics, indicated that the DNA DSB repair capacity of MSCs may be altered in hypoxia. Analysis of the DNA DSB marker, γ-H2AX [53-55], demonstrated that (a) H2AX Ser139 phosphorylation (Fig. 3A) and (b) γ-H2AX IRIF (Fig. 3B–3D) were resolved at a faster rate in irradiated MSCs cultured in hypoxia than in normoxia. For example, at 2 hours, 41 (21% O2) versus 33 (5% O2) γ-H2AX IRIF (p < .05) were present per MS5 nucleus (Fig. 3D, upper panel) and 33 (21% O2) versus 23 (5% O2) IRIF (p < 0.01) per ST2 nucleus (Fig. 3D, lower panel). Similar to previous findings [44], maximal H2AX phosphorylation was delayed in ST4.5 thymocytes until 2–4 hours post-IR and its resolution was not improved in hypoxia (Supporting Information Fig. 3E). Interestingly, Rad51 IRIF (indicative of DNA DSB repair by HR) formed at a faster rate in MSCs cultured in hypoxia than in normoxia (Fig. 3B, 3C, 3E). For example, at 1 hour, 2 (21% O2) versus 5 (5% O2) Rad51 IRIF (p < .05) were present per MS5 nucleus (Fig. 3E, upper panel) and 2 (21% O2) versus 6 (5% O2) IRIF (p < 0.01) per ST2 nucleus (Fig. 3E, lower panel).

image

Figure 3. Hypoxia accelerates resolution of DNA double-strand breaks by irradiated mesenchymal stromal cells (MSCs). (A): γ-H2AX, total H2AX, and β-tubulin expression levels in normoxic and hypoxic MSCs 0–24 hours post 10 Gy irradiation. Images of MS5 (B) and ST2 (C) nuclei stained for γ-H2AX and Rad51 IR-induced foci (IRIF) in 21%/5% O2 1–8 hours post 10 Gy irradiation. Average number of (D) γ-H2AX and (E) Rad51 IRIF per nucleus 0–24 hours post-irradiation. *, MS5 samples; #, ST2 samples; */#, p < .05; ##, p < .01, compared with normoxic samples, two-way ANOVA with Bonferroni post-test. Abbreviation: IR, irradiation.

Download figure to PowerPoint

The extent of γ-H2AX IRIF formation and H2AX phosphorylation at 1 hour post-IR was similar in normoxic and hypoxic MSC nuclei (Fig. 3B, 3C, 1 hour panels). Furthermore, direct analysis of DNA DSBs in irradiated MSCs via the Neutral Comet assay (Supporting Information Materials and Methods) demonstrated that similar numbers of DNA DSBs were generated in both oxygen tensions (Supporting Information Fig. 4A–4C). Therefore, the accelerated γ-H2AX resolution and Rad51 IRIF formation in hypoxic MSCs was likely due to alterations in their DNA DSB repair capacity.

To analyze the HR efficiency of MSCs, MS5 cells were stably transfected with the DR-GFP HR reporter, to generate the MS5 subclone, MS5 HR4.4 (Materials and Methods section; Supporting Information Fig. 3A, 3B). Transfection with pCBASce vector expressing the rare cutting endonuclease, I-SceI, generates DSBs within the GFP reporter substrate and successful HR-mediated repair results in GFP expression (Supporting Information Fig. 5A) [46, 47]. At 2 days post-transfection, 4.6% GFP-positive cells were detected in normoxia whereas this significantly increased to 6.6% in hypoxia (p < .05) (Fig. 4A, 4B). These results demonstrate that mouse MSCs can repair DNA DSBs using HR and that their HR efficiency is increased in hypoxia.

image

Figure 4. Hypoxia enhances DNA double-strand break repair in mesenchymal stromal cells (MSCs). (A, B): Flow cytometry analysis of GFP positive MS5 HR4.4 cells transfected with Lipofectamine 2000 alone (Mock) or with pCBASce in 21%/5% O2. (C, D): End-joining efficiency of control SCID-derived MSCs (SCID bulk) and MS5 nuclear extracts in 21%/5% O2 and corresponding DNA-PKcs and Rad51 expression levels in these extracts. (E): DNA ligase IV, DNA-PKcs, Rad51, and β-tubulin expression levels in control MS5, ST2, and ST4.5 cells cultured >3 days in 21%/5% O2. *, p < .05 compared with normoxic samples, Student's t test. Abbreviations: GFP, green fluorescent protein; SCID, severe combined immuno-deficiency; SSC, side scatter.

Download figure to PowerPoint

image

Figure 5. HIF-1α contributes to enhanced DNA double-strand break repair in hypoxic mesenchymal stromal cells (MSCs). HIF-1α, HIF-2α, DNA ligase IV, DNA-PKcs, Rad51, and β-tubulin expression levels in untreated MSCs (A) and in hypoxic MSCs transfected with scrambled (control) or HIF-1α siRNA (HIF-1α) (E) 0–24 hours following transfer from 21% to 5% O2. (B): Radiation survival curves of hypoxic transfected (scrambled or HIF-1α siRNA) MSCs. (C): γ-H2AX, total H2AX, and β-tubulin expression levels in hypoxic transfected MSCs 0–24 hours post 10 Gy irradiation. (D): Average number of γ-H2AX IRIF per nucleus 0–24 hours post-irradiation. *, MS5 samples; #, ST2 samples; *, p < .05, ##, p < .01; ###, p < .001 compared with normoxic samples, two-way ANOVA with Bonferroni post-test. Abbreviations: HIF-1α, hypoxia inducible factor 1α; IR, irradiation; siRNA, small interfering RNA.

Download figure to PowerPoint

Nonhomologous end-joining (NHEJ) efficiency was measured using a previously established in vitro assay in which the capacity of control nuclear extracts to religate digested plasmid DNA containing a single DNA DSB is analyzed [44, 48]. Analysis of ligated fragment intensity indicated that end-joining of monomeric DNA fragments was significantly increased (p < .05) in nuclear extracts of MS5 cells cultured in hypoxia compared with normoxia (Fig. 4C, 4D). End-joining was absent in the presence of SCID-derived bulk MSC nuclear extracts (SCID bulk), used as negative control, demonstrating that plasmid religation was NHEJ-dependent (Fig. 4C). Western blot analysis of MSC extracts revealed that DNA ligase IV and DNA-PKcs were expressed at higher levels in MSCs cultured in hypoxia than in normoxia whereas Rad51 expression levels were comparable (Fig. 4C, 4E; Supporting Information Fig. 1F). The expression levels of these proteins in ST4.5 thymocytes, although low, were unaltered by oxygen tension (Fig. 4E; Supporting Information Fig. 1F).

HIF-1α Is a Mediator of Enhanced MSC Radioresistance in Hypoxia

To confirm that this increased DNA ligase IV and DNA-PKcs expression was hypoxia-dependent, MSCs continuously cultured in normoxia were transferred to hypoxia and their expression levels were analyzed over a 24-hour time course (Fig. 5A). In both MSC lines, DNA ligase IV and DNA-PKcs expression levels began to increase at 4 and 8 hours, respectively, in hypoxia whereas Rad51 expression levels remained stable (Fig. 5A). HIF-1α and HIF-2α were stabilized in MSC lines, beginning at 4 and at 24 hours, respectively, in hypoxia (Fig. 5A). HIF-1α stabilization coincided with increasing levels of DNA ligase IV and DNA-PKcs in MSC lines exposed to hypoxia (Fig. 5A). In bulk MSCs, DNA ligase IV and DNA-PKcs expression levels fluctuated over time in hypoxia but were strongly increased at 24 hours, coinciding with HIF-1α and HIF-2α stabilization, whereas Rad51 levels remained unaltered (Supporting Information Fig. 5C, 5D). The (a) delayed HIF-1α and HIF-2α stabilization kinetics and (b) varying levels of DNA ligase IV and DNA-PKcs in hypoxic bulk MSCs could be due to the inherent heterogeneity of primary uncloned MSC cultures in which various progenitor cell types are likely to differentially respond to hypoxia exposure.

The correlation between HIF-1α stabilization and increasing expression levels of DNA-PKcs and DNA ligase IV indicated a potential role for HIF-1α in regulating DNA DSB repair in hypoxic MSCs. siRNA-mediated knockdown of HIF-1α caused a 1.5-fold and a twofold reduction in the radioresistance of MS5 and ST2 cells, respectively, in hypoxia (Fig. 5B). MSC colony size was unaffected by HIF-1α knockdown (data not shown). In addition, (a) H2AX phosphorylation (Fig. 5C) and (b) γ-H2AX IRIF (Fig. 5D; Supporting Information Fig. 6A) persisted for longer in irradiated HIF-1α depleted MSCs than in control MSCs. For example, at 4 hours, 19 (control siRNA) versus 24 (HIF-1α siRNA) γ-H2AX IRIF were present per MS5 nucleus (Fig. 5D, upper panel) and 22 (control siRNA) versus 30 (HIF-1α siRNA) IRIF (p < .001) were present per ST2 nucleus (Fig. 5D, lower panel). Also, Rad51 IRIF formed less efficiently and persisted for longer in irradiated HIF-1α depleted MSC nuclei (Supporting Information Fig. 6A–6C). For example, at 1 hour, MS5 nuclei contained 6 (control siRNA) versus 3 (HIF siRNA) Rad51 IRIF (Supporting Information Fig. 6B) and at 8 hours, ST2 nuclei contained 7 (control siRNA) versus 11 (HIF siRNA) Rad51 IRIF (Supporting Information Fig. 6C). Finally, HIF-1α knockdown prevented hypoxia-dependent increases in DNA ligase IV and DNA-PKcs expression in MSCs, whereas Rad51 expression levels were unaltered (Fig. 5E). These results confirm that HIF-1α contributed to enhanced MSC radioresistance in hypoxia by altering their DNA DSB repair capacity.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Author Contributions
  10. Disclosure of Potential Conflicts of Interest
  11. References
  12. Supporting Information

MSCs are radioresistant bone marrow-derived progenitors that reside in hypoxic niches in vivo [5, 6, 15, 16, 56]. Hypoxia is emerging as an important modulator of stem/progenitor cell behavior [15, 16]. Previous studies have demonstrated that in hypoxia, MSCs exhibit (a) an increased proliferative capacity; (b) increased expression of pluripotency genes and telomerase activity; (c) improved differentiation potential; and (d) altered cytokine and growth factor profiles [49, 50, 57-61]. Similarly, we found that mouse MSC (a) growth rate (Fig. 1A, 1B; Supporting Information Fig. 1A) and (b) colony size was increased in hypoxia (Fig. 1E, 1F). However, the effects of hypoxia on MSC radioresistance are largely unknown.

We have demonstrated herein that mouse MSC radioresistance is enhanced in hypoxia in vitro. Cellular adaptation to hypoxia results in alterations in various biological processes that can consequently promote resistance to DNA damaging agents, including IR. These mechanisms include (a) increased resistance to apoptotic stimuli; (b) induction of cell cycle arrest; (c) suppression of cellular senescence; and (d) altered DNA repair capacity [19, 30, 36, 62-68]. In this study, hypoxia did not affect (a) MSC cell cycle distribution under normal conditions (Fig. 2C, 2D; Supporting Information Fig. 3A); (b) IR-induced p53 stabilization kinetics (Fig. 2E); or (c) the extent of IR-induced apoptosis in mouse MSCs (Supporting Information Fig. 3B). However, the DNA DSB repair capacity of MSCs was altered in hypoxia (Figs. 3, 4).

DNA DSBs are the most genotoxic lesions caused by IR and are repaired via NHEJ or HR. HR is reported to be suppressed in multiple hypoxic cell types whereas NHEJ can be either suppressed or enhanced in hypoxia [36, 62]. However, herein, both HR (Fig. 4A, 4B) and NHEJ (Fig. 4C, 4D) were found to be upregulated in hypoxic MSCs. HIF-1α is implicated in regulating the expression and/or post-translational modification of various DNA DSB repair proteins in hypoxic cancer cells [34, 65-68]. However, whether HIF-1α modulates DNA DSB repair in stem/progenitor cells, such as MSCs, is currently unknown. Herein, HIF-1α stabilization contributed to the increased expression of DNA-PKcs and DNA ligase IV in hypoxic MSCs (Figs. 4E, 5A, 5E). Although reduced Rad51 expression in hypoxia has been reported for multiple cell types [65-67], Rad51 levels in MSCs were unaffected by hypoxia (Figs. 4E, 5A, 5E). However, Rad51 IRIF formation in MSCs was accelerated in hypoxia (Fig. 3B, 3C, 3E) and these kinetics were altered by HIF-1α depletion (Supporting Information Fig. 6A–6C). This suggests that HIF-1α may influence MSC HR activity by affecting the recruitment of repair proteins, such as Rad51, to DNA DSBs. Overall, this study indicates that HIF-1α contributes to the increased radioresistance of hypoxic MSCs, at least in part, by promoting the upregulation of DNA DSB repair in this cell type (summarized in Fig. 6).

image

Figure 6. HIF-1α contributes to the enhanced DNA DSB repair capacity of hypoxic mouse MSCs. Irradiated MSCs activate DNA damage checkpoints and DNA DSB repair mechanisms, namely NHEJ and HR, promoting MSC survival. In hypoxia (right panel), HIF-1α stabilization results in (a) increased expression of DNA-PKcs and DNA ligase IV and (b) accelerated Rad51 IRIF formation. Consequently, the DNA DSB repair capacity of MSCs is increased, promoting cell cycle recovery and long-term survival. Abbreviations: DSB, double stranded break; HIF-1α, hypoxia inducible factor 1α; HR, homologous recombination; IR, irradiation; IRIF, IR-induced foci; MSC, mesenchymal stromal cell; NHEJ, nonhomologous end-joining.

Download figure to PowerPoint

Enhanced cellular radioresistance in hypoxia is also associated with the generation of less reactive oxygen species (ROS)-mediated DNA damage, compared with normoxic cells irradiated at an equivalent IR dose [69]. However, DNA DSB generation in irradiated MSCs was unaffected by oxygen tension (Supporting Information Fig. 4A–4C). In addition, while MSCs expressed considerably higher levels of the ROS scavenging enzyme, superoxide dismutase-1 (SOD-1), than ST4.5 thymocytes, SOD-1 levels in all cell types were unaffected by oxygen tension (Supporting Information Fig. 4D). However, irradiation leads to the generation of numerous forms of DNA damage, for example, single-strand breaks and base damage. Therefore, whether altered regulation of other DNA repair mechanisms in hypoxic MSCs could also contribute to their enhanced radioresistance remains to be investigated. It is also known that the activation of cellular senescence in MSCs is suppressed in hypoxia [50, 60]. Therefore, for future studies it may be interesting to determine whether DNA damage-induced senescence is also suppressed in hypoxic MSCs.

The effects of hypoxia on MSC radiobiology have numerous important implications. Conditional deletion of MSC subpopulations in the bone marrow results in defective hematopoiesis [70-74]. Total-body irradiation (TBI) is commonly used in preparative regimens preceding allogeneic bone marrow transplantation in order to destroy the radiosensitive host hematopoietic system [75, 76]. Therefore, increased MSC radioresistance in hypoxia is likely to promote host-/patient-derived MSC survival in the bone marrow post-TBI, thereby improving hematopoietic reconstitution.

MSCs are capable of homing to the hypoxic tumor microenvironment where they can contribute to tumorigenesis and metastasis [13, 77-81]. MSCs produce increased amounts of proangiogenic factors (e.g., VEGF) and protumorigenic cytokines (e.g. CXCL12) in hypoxia [57, 59]. Furthermore, MSCs promote the therapeutic resistance of multiple hematological malignancies, including chronic myeloid leukemia and chronic lymphocytic leukemia [82-84]. In combination with our finding that MSC radioresistance is increased in hypoxia, it is likely that MSC-derived CAFs are resistant to radiotherapy within the tumor microenvironment and thereby support tumor survival.

Recent studies indicate biological response(s) modulated in hypoxia are cell type-dependent [85-87]. By comparing the radiation response of MSCs and double positive thymocytes, we have also shown that hypoxia-mediated modulation of the DDR is cell type-dependent. Therefore, our understanding of how hypoxia influences the radioresistance and chemoresistance of not only cancer cells but also of their supportive stroma, including MSCs, will also likely be essential for developing effective strategies to overcome hypoxia-induced therapeutic resistance in cancer.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Author Contributions
  10. Disclosure of Potential Conflicts of Interest
  11. References
  12. Supporting Information

In conclusion, we have demonstrated that hypoxia enhances the radioresistance of mouse MSCs. This increased radioresistance was found to be due to hypoxia-mediated alterations in the DDR of mouse MSCs, including an increased DNA DSB repair capacity, and HIF-1α was identified as a key mediator in this process. Our findings have important implications for advancing our understanding of MSC functions in supporting hematopoietic reconstitution following myeloablative radiotherapy and in promoting tumor survival and therapeutic resistance.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Author Contributions
  10. Disclosure of Potential Conflicts of Interest
  11. References
  12. Supporting Information

We thank all members of the Ceredig and Lowndes laboratories for valuable discussions; Prof. Matthew Griffin, Claas Baustian, and Senthilkumar Alagesan for providing primary bulk mouse MSC cultures; Dr. Stephen Elliman for use of the hypoxia incubators; Drs. Michael D. Rainey and Alessandro Natoni for flow cytometry techniques and reagents; Dr. Shirley Hanley for running the flow cytometry core facility; Dr. Maoija Xu for Fox Chase SCID Beige murine femurs; Prof. Ciaran Morrison for DR-GFP plasmid; Janna Luessing and Danielle Hamilton for expertise with in vivo HR assay; Dr. Silvia Maretto for siRNA transfection protocol; Drs. Panagiota Sotiropoulou and Stephen Rea for technical assistance and reagents for in vitro DNA DSB end-joining assay; and Dr. Michelle Duffy for help with statistical analysis. T.S. was supported by an Irish Research Council Government of Ireland Postgraduate Scholarship (Grant No. RS20102702) and by the Thomas Crawford Hayes Research Fund, NUI Galway; N.F.L. by Science Foundation Ireland (Grant No. SFI 07/IN.1/B958); and Rh.C. by Science Foundation Ireland (Grant No. SFI 09/SRC/B1794) and Stokes Professorship.

Author Contributions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Author Contributions
  10. Disclosure of Potential Conflicts of Interest
  11. References
  12. Supporting Information

T.S.: conception and design, collection and assembly of data, data analysis and interpretation, and manuscript writing; N.F.L. and R.C.: conception and design, financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Author Contributions
  10. Disclosure of Potential Conflicts of Interest
  11. References
  12. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgments
  9. Author Contributions
  10. Disclosure of Potential Conflicts of Interest
  11. References
  12. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
stem1683-sup-0001-Suppinfo.doc29KSupporting Information
stem1683-sup-0002-Suppfig1.tif387KSupporting Information Figure 1
stem1683-sup-0003-Suppfig2.tif310KSupporting Information Figure 2
stem1683-sup-0004-Suppfig3.tif644KSupporting Information Figure 3
stem1683-sup-0005-Suppfig4.tif317KSupporting Information Figure 4
stem1683-sup-0006-Suppfig5.tif638KSupporting Information Figure 5
stem1683-sup-0007-Suppfig6.tif405KSupporting Information Figure 6

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.