Abstract
Radiotherapy increases tumor cure and survival rates; however, radiotherapy-induced bone damage remains a common issue for which effective countermeasures are lacking, especially considering tumor recurrence risks. We report a high-specificity protection technique based on noninvasive electromagnetic field (EMF). A unique pulsed-burst EMF (PEMF) at 15 Hz and 2 mT induces notable Ca2+ oscillations with robust Ca2+ spikes in osteoblasts in contrast to other waveforms. This waveform parameter substantially inhibits radiotherapy-induced bone loss by specifically modulating osteoblasts without affecting other bone cell types or tumor cells. Mechanistically, primary cilia are identified as major PEMF sensors in osteoblasts, and the differentiated ciliary expression dominates distinct PEMF sensitivity between osteoblasts and tumor cells. PEMF-induced unique Ca2+ oscillations depend on interactions between ciliary polycystins-1/2 and endoplasmic reticulum, which activates the Ras/MAPK/AP-1 axis and subsequent DNA repair Ku70 transcription. Our study introduces a previously unidentified method against radiation-induced bone damage in a noninvasive, cost-effective, and highly specific manner.
INTRODUCTION
Radiotherapy, either alone or in combination with surgery, is one of the most common approaches for oncologic treatment. Nearly 10 million cancer patients worldwide (approximately half of new cancer cases per year) are estimated to undergo radiotherapy annually (1). The rapid development of radiotherapy techniques is continuously prolonging the survival time of patients and increasing the cure rate of malignant tumors (for example, the 10-year survival rates of prostate cancer and breast cancer have reached ~95 and 85%, respectively) (2, 3), and radiotherapy-related adverse effects have received increasing attention. The skeleton has the capacity to absorb much more radiation energy than other tissues, owing to its higher mineral component content, while radiotherapy-induced bone damage also remains a common and tough clinical issue (4, 5). Radiation-induced bone damage includes the rapid loss of bone mass, increased bone fragility and susceptibility to fractures, and elevated risks of osteonecrosis in the radiation field (6). Clinical data suggest that approximately 20% of breast cancer patients experience pathologic rib fractures after radiotherapy (7, 8). Pelvic fractures were noted in 16 to 37% of cervical cancer patients following focal radiotherapy (9, 10). Moreover, focal radiotherapy has been found to induce a progressive systemic decrease in bone mineral density (BMD) in cancer patients (11, 12). Considering that most cancer patients are middle-aged or older and may suffer from various primary diseases (e.g., osteoporosis and diabetes), radiotherapy aggravates the risks of fragility fractures and osteonecrosis and subsequent challenges of medical care.
The major mechanism by which radiation induces bone damage is through the inhibition of bone formation rather than the stimulation of bone resorption (13). Osteoblasts are highly sensitive to ionizing radiation. The growth, survival, and functional maturation of osteoblasts are markedly suppressed when the bone is exposed to radiation (14, 15). Osteoblasts are also much more vulnerable to ionizing radiation than undifferentiated mesenchymal stem cells (16). Radiation primarily targets DNA molecules in osteoblasts and induces a series of DNA lesion events, among which DNA double-strand breaks (DSBs) are the most deleterious, resulting in massive loss of genomic information and cell death (17, 18). Cells endeavor to initiate nonhomologous end joining (NHEJ) to repair DSBs, which is required for the interaction between the heterodimer of Ku70/80 and DNA-dependent protein kinase catalytic subunit (DNA-PKcs) (17). In particular, Ku70, located proximal to the broken DNA end, plays a critical role in recognizing, binding, and bridging the damaged DNA, and maintaining the stabilization of the NHEJ complex (19). In addition, Ku70 is able to prevent cell death independent of DSB repair (20, 21). The common agent for the osteoporosis treatment in clinics includes antiresorptive bisphosphonates and anabolic parathyroid hormone (PTH). In contrast to bisphosphonates with limited efficacy against radiation-induced bone fragility as well as increased risks of osteonecrosis (22), treatment with PTH for radiation bone damage is more confirmative (13, 23). However, PTH not only is expensive but also is known to stimulate tumor growth in addition to protection against osteoporosis (24, 25) and is listed as a banned drug for cancer patients by the U.S. Food and Drug Administration (FDA). Thus, it is of great clinical importance to develop approaches for resisting radiation-related bone damage with safe, cost-effective, and highly specific nature.
The discovery of the piezoelectric properties of bone raised the possibility of introducing exogenous electromagnetic signals to regulate bone cell biology. The use of the electromagnetic field (EMF) with low frequency and low intensity was found to greatly accelerate fracture healing over four decades ago (26), leading to subsequent approval by the FDA in orthopedic applications. Accumulating evidence suggests that EMF exposure (5 to 100 Hz, 0.5 to 10 mT) stimulates osteoblast growth and differentiation and promotes osteogenesis (27–30). However, the major obstacles of studies with regard to EMF are the difficulties in identifying the optimal waveform and parameter scheme for a certain specific bone disease and in clarifying the mechanism of the EMF sensation and signal transduction of bone cells. Primary cilia, as solitary, nonmotile, hair-like protrusions that emanate from the surface of cells, can detect external biophysical signals (e.g., mechanical stimuli) (31). It has been shown that osteoblasts are rich in the expression of primary cilia, while most cancer cells may be devoid of primary cilia (32–37). Thus, on the basis of the potential differential expression of primary cilia, it will be interesting to assess whether EMF can specifically resist radiotherapy-induced bone loss in the context of not altering tumor cell activity.
In this study, we identified that a previously unidentified waveform of pulsed-burst EMF (PEMF) at 15 Hz and 2 mT induced the most notable signaling response in irradiated osteoblasts characterized by unique intracellular calcium (Ca2+) oscillations with multiple robust Ca2+ spikes in contrast to other EMF waveforms. This waveform parameter significantly inhibited radiotherapy-induced bone loss by specifically regulating the activities of osteoblasts among various bone cells. The differential expression of primary cilia was found to dominate the distinct sensitivity between osteoblasts and tumor cells to PEMF. Mechanistically, primary cilia acted as the “PEMF sensor” in irradiated osteoblasts, initiated Ca2+ dynamics via the interaction between polycystin-1/2 (PC-1/2) and endoplasmic reticulum (ER), and promoted subsequent Ras/mitogen-activated protein kinase (MAPK)/AP-1 axis–mediated nuclear Ku70 activation. Moreover, PEMF-induced bone gain was blunted in irradiated mice with Ku70 deficiency in osteoblasts.
RESULTS
EMF triggers a specific signaling response in irradiated bone cells highly dependent on the waveform parameters
Intracellular Ca2+ is a critical and ubiquitous second messenger and one of the earliest signal transduction events in response to external biophysical stimuli (e.g., mechanical loading), thereby regulating a variety of subsequent downstream cellular functions (38). Thus, we examined real-time intracellular Ca2+ signaling responses in irradiated bone cells under EMF stimulation with various waveforms that were reported to have potential biological effects, including sinusoidal EMF (sin-EMF), single-pulsed EMF (sPEMF) (39), and unique PEMF. As shown in Fig. 1A, Ca2+ transients were initiated under EMF stimulation with various waveforms in irradiated osteoblasts; nonetheless, no Ca2+ response was observed in irradiated osteoclasts or osteocytes treated by any EMF waveform. Furthermore, in contrast to sin-EMF and sPEMF, PEMF triggered a more robust intracellular Ca2+ signaling response characterized by unique Ca2+ oscillations with multiple robust Ca2+ spikes in irradiated osteoblasts. PEMF at 2 mT induced more notable Ca2+ oscillations in irradiated osteoblasts than other PEMF intensities with the frequency of 15 or 50 Hz [two most commonly used EMF frequencies (40, 41)]. Furthermore, PEMF at 15 Hz and 2 mT induced the most notable Ca2+ oscillations in irradiated osteoblasts among various frequencies (including 5, 10, 15, 20, 30, 50, 75, and 100 Hz; figs. S1 to S3 and movie S1) in the range of 5 to 100 Hz that is reported to elicit positive physiological effects (40). Moreover, quantitative analyses showed that irradiated osteoblasts under PEMF stimulation at 15 Hz and 2 mT had higher Ca2+ spike numbers and magnitudes than any other parameters (fig. S4). No significant difference in Ca2+ spike numbers or magnitudes was observed in irradiated osteoclasts or osteocytes among all EMF parameters (figs. S5 and S6). We then investigated the effects of PEMF at 15 Hz and 2 mT on the viability and function of osteoblasts, osteoclasts, and osteocytes. PEMF at 15 Hz and 2 mT significantly enhanced the differentiation and mineralization of both normal and irradiated osteoblasts according to alkaline phosphatase (ALP) assays and Alizarin red staining (Fig. 1, B and C) and also stimulated the expression of osteogenic genes and proteins in osteoblasts, including Col1a1, Cbfα1, and Osx (Fig. 1, D and E). However, PEMF at 15 Hz and 2 mT had no observable effect on the survival and function of either normal or irradiated MLO-Y4 osteocytic cells (fig. S7, A to F). PEMF also had no impact on osteoclast differentiation based on tartrate-resistant acid phosphatase (TRAP) staining (fig. S7G).
Fig. 1. EMF induces specific signaling response in irradiated bone cells in a waveform- and dose-dependent manner.
(A) Primary osteoblasts, RAW264.7 cells (7 days following RANKL incubation), and MLO-Y4 osteocytic cells were exposed to 8-Gy radiation twice (with 1-day interval) at a rate of 1.65 Gy/min. Intracellular Ca2+ dynamics was analyzed under sin-EMF, sPEMF, or PEMF waveforms with various frequencies/intensities. The sPEMF waveform has a duty cycle of 7.5% with a frequency of 15 or 50 Hz. The PEMF waveform is generated on the basis of the low-frequency sPEMF wave (15 or 50 Hz) that is modulated by another carrier wave with a frequency of 4.55 kHz (pulse width, 0.2 ms; pulse wait, 0.02 ms). The pulse intensity of the carrier wave is the same as the low-frequency modulating wave (15 or 50 Hz). The PEMF waveform at 15 Hz and 2 mT induced the most notable Ca2+ oscillations with repetitive robust Ca2+ spikes in irradiated osteoblasts. (B to E) Osteoblast function analyses under 15 Hz, 2mT PEMF stimulation (n = 6). (B) ALP staining and ALP activity assays following osteogenic medium incubation and PEMF stimulation for 7 days. (C) Alizarin red staining following osteogenic medium incubation and PEMF stimulation for 21 days. (D and E) Quantitative real-time polymerase chain reaction (qRT-PCR) and Western blotting for the Col1a1, Cbfα1, and Osx expression following osteogenic medium incubation and PEMF stimulation for 3 days. R, radiated group; NR, nonradiated group. Graphs represent means ± SD. P < 0.05; *P < 0.01. Analyses were done using two-way analysis of variance (ANOVA) with Bonferroni’s posttest. Scale bars, 50 μm (B and C).
PEMF mitigates focal radiation–induced bone loss by modulating osteoblasts
We next investigated the changes in the bone phenotype of rats with focal irradiation in response to EMF with the optimal parameter initiating signaling response in irradiated osteoblasts (i.e., PEMF at 15 Hz and 2 mT). We established an animal model of unilateral focal hindlimb radiation (Fig. 2A). The rats with focal hindlimb radiation following 45 days exhibited significant femoral trabecular bone loss according to micro–computed tomography (micro-CT) results, as characterized by an approximate 50% decrease in bone volume fraction (BV/TV), BMD, and trabecular number (Tb.N), a 65% decrease in trabecular thickness (Tb.Th), an approximate 30% increase in bone surface per bone volume (BS/BV), and a 50% increase in trabecular separation (Tb.Sp) compared with the contralateral side (Fig. 2B). PEMF significantly improved the trabecular bone microstructure in nonirradiated femora and inhibited irradiation-induced bone loss to a level similar to that of the nonradiated (NR) side in the sham group (Fig. 2B). Furthermore, PEMF induced a significant increase in intrinsic material properties in both irradiated and nonirradiated femora, according to nanoindentation assays (fig. S8). Hematoxylin and eosin (H&E) staining revealed that PEMF significantly increased the trabecular bone mass in irradiated tibiae (Fig. 2C). Immunostaining and calcein/Alizarin double labeling showed decreased osteoblast numbers, increased osteoblast apoptosis, and reduced bone formation rates on the NR side (Fig. 2, D to F). PEMF significantly improved osteoblast survival and bone formation rates in irradiated hindlimbs (Fig. 2, D to F). Moreover, H&E and terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) staining revealed that PEMF had no effect on osteocyte viability and apoptosis in either irradiated or nonirradiated hindlimbs (Fig. 2, G to J). PEMF also had no impact on osteoclast numbers on the bone surface according to TRAP staining (Fig. 2K). Serum enzyme-linked immunosorbent assay (ELISA) assays showed that PEMF caused a significant increase in the concentrations of bone formation biomarkers OCN and P1NP without any change in bone resorption biomarkers CTX-1 and TRACP-5b (Fig. 2, L and M). PEMF also had no observable effect on body weight or food intake in rats with focal radiation (fig. S9A).
Fig. 2. PEMF ameliorates focal radiation–induced bone loss by modulating osteoblasts.
(A) Schematic representation of focal radiation and PEMF administration. (B) Representative micro–computed tomography (micro-CT) images and quantitative analyses of rat distal femora (n = 8 femora per group). (C) H&E staining showing the trabecular area of rat proximal tibiae (n = 8 tibiae per group). (D) Dynamic bone histomorphometric analyses based on calcein/Alizarin red labeling (n = 8 femora per group). (E) Runx2 immunohistochemical staining (n = 8 tibiae per group). (F) TUNEL staining for the detection of apoptotic osteoblasts on bone surface (n = 8 tibiae per group). (G) H&E staining showing the empty osteocyte lacunae (n = 8 tibiae per group). (H) TUNEL staining for the detection of apoptotic osteocytes (n = 8 tibiae per group). (I and J) Scanning electron microscope observation for the osteocyte number and osteocyte dendrite number in rats with focal irradiation in the presence or absence of the PEMF stimulation (n = 8 femora per group). (K) TRAP staining for labeling osteoclasts on bone surface (n = 8 tibiae per group). (L and M) Serum ELISA assays for bone formation markers (OCN and P1NP) and bone resorption markers (CTX-1 and TRACP-5b) (n = 8 rats per group). Graphs represent means ± SD. (B to K) P < 0.05; *P < 0.01. Statistical analyses were done using two-way ANOVA with Bonferroni’s posttest. (L and M) **P < 0.01. Statistical analyses were done using Student’s t test. Scale bars, 200 μm (C and K), 100 μm (I), 50 μm (E to H), 20 μm (D), and 5 μm (J).
Ku70 mediates the PEMF-induced enhancement of cell function of irradiated osteoblasts
Next, we examined whether PEMF repairs radiation-induced DNA damage. The expression of γ-H2AX, a sensitive marker for DNA DSBs, was significantly increased in osteoblasts in irradiated rat hindlimbs (Fig. 3A). Furthermore, γ-H2AX foci induction was enhanced in irradiated primary osteoblasts (Fig. 3B). PEMF significantly suppressed the radiation-induced increase in osteoblast γ-H2AX expression both in vivo and in vitro (Fig. 3, A and B). To elucidate the mechanism by which PEMF repairs DNA DSBs, we examined the change in the Ku70 expression (the core protein of NHEJ repair). PEMF induced a significant increase in osteoblast Ku70 expression in irradiated rat hindlimbs (Fig. 3C). PEMF was also found to stimulate nuclear Ku70 expression in irradiated primary osteoblasts in vitro, according to fluorescence staining assays (Fig. 3D). Moreover, Western blotting analysis revealed that Ku70 protein levels increased significantly from 6 hours of PEMF stimulation in irradiated primary osteoblasts and remained persistently high within 72 hours (Fig. 3E). Then, a lentivirus expression system with stable knockdown of Ku70 in primary osteoblasts was constructed, and approximately 70% knockdown efficiency for Ku70 in shKu70-#3 was observed according to Western blotting analysis (fig. S10A). PEMF-induced inhibition of cellular apoptosis in irradiated osteoblasts was almost completely abolished when the Ku70 expression was blocked (Fig. 3F). Moreover, PEMF-mediated augmentation of differentiation and mineralization of irradiated osteoblasts was completely blunted after Ku70 knockdown (Fig. 3, G to K). Thus, our study demonstrates that Ku70 is the key nuclear molecule responsible for PEMF-mediated cytoprotection in irradiated osteoblasts.
Fig. 3. The DNA repair protein Ku70 mediates the PEMF-induced enhancement of cell function of irradiated osteoblasts.
(A) Osteoblast γ-H2AX immunofluorescence staining in proximal tibiae of irradiated rats (n = 8). (B) γ-H2AX immunofluorescence staining in irradiated primary osteoblasts following PEMF stimulation for 6 hours (n = 6). (C) Osteoblast Ku70 immunofluorescence staining in proximal tibiae of irradiated rats (n = 8). (D) Ku70 immunofluorescence staining in irradiated primary osteoblasts following PEMF stimulation for 6 hours (n = 6). (E) Western blotting analyses of the Ku70 expression in irradiated primary osteoblasts after PEMF stimulation at various time points (n = 4). (F to J) Survival and function assays of primary osteoblasts infected with shCtrl and shKu70 lentivirus after PEMF stimulation (n = 6). (F) Annexin V–FITC/propidium iodide (PI) apoptosis assays following PEMF stimulation for 3 days. (G and H) ALP staining and ALP activity analyses following osteogenic medium incubation and PEMF stimulation for 7 days. (I) Alizarin red staining following osteogenic medium incubation and PEMF stimulation for 21 days. (J and K) qRT-PCR and Western blotting analyses of the expression of Col1a1, Cbfα1, and Osx following osteogenic medium incubation and PEMF stimulation for 3 days. Graphs represent means ± SD. (A to D and F to K) P < 0.05; *P < 0.01. Statistical analyses were done using two-way ANOVA with Bonferroni’s posttest. (E) **P < 0.01. Statistical analyses were done using one-way ANOVA with Bonferroni’s posttest. Scale bars, 50 μm (A, C, G, and I) and 20 μm (B and D).
Differentiated expression of primary cilia dominates the highly distinct sensitivity of osteoblasts and tumor cells to PEMF
We next compared the sensitivity of osteoblasts and tumor cells to PEMF in the presence or absence of radiation. Ion radiation reduced cell viability and promoted apoptosis in primary osteoblasts, MCF-7 breast cancer cells, colon cancer SW620 cells, malignant melanoma A375 cells, and 143B osteosarcoma cells (Fig. 4, A and B). PEMF markedly improved osteoblast viability and inhibited osteoblast apoptosis in the presence and absence of radiation; nonetheless, PEMF had no effect on cell viability or apoptosis in any tumor cell type at any time point (Fig. 4, A and B). Furthermore, the viability and apoptosis of any irradiated tumor cell type did not change under sin-EMF, sPEMF, or PEMF stimulation at various intensities and frequencies (fig. S11). To identify the mechanism of distinct sensitivity between osteoblasts and tumor cells to PEMF, we examined the expression of primary cilia on the plasma membrane. Primary cilia were highly abundant in primary osteoblasts and absent in most tumor cells (Fig. 4C). Then, we constructed a shIFT88 lentivirus to block ciliogenesis in primary osteoblasts and confirmed shIFT88-#1 to be efficiently silenced (fig. S10B). PEMF-mediated apoptosis inhibition and differentiation acceleration in irradiated osteoblasts were almost completely abolished when primary cilia were depleted (Fig. 4, D to I). The PEMF-induced increase in Ku70 expression in irradiated osteoblasts was also completely blunted after ciliogenesis was inhibited (Fig. 4, J and K).
Fig. 4. The differentiated expression of primary cilia between osteoblasts and tumor cells mediates their highly distinct sensitivity to the PEMF stimulation.
(A and B) MTT proliferation assays and annexin V–FITC/PI apoptosis assays in irradiated primary osteoblasts, human breast cancer MCF-7 cells, human colon cancer SW620 cells, human malignant melanoma A375 cells, and osteosarcoma 143B cells after the PEMF stimulation at various time points (n = 6). (C) Acetylated-α-tubulin (AC-α-tubulin) immunofluorescence staining for detecting the expression of primary cilia in primary osteoblasts, MCF-7, SW620, A375, and 143B cells. (D to I) Survival and function assays of primary osteoblasts infected with shIFT88 lentivirus (blockade of ciliogenesis) after PEMF stimulation (n = 6). (D) Annexin V–FITC/PI apoptosis assays following PEMF stimulation for 3 days. (E and F) ALP staining and ALP activity analyses following osteogenic medium incubation and PEMF stimulation for 7 days. (G) Alizarin red staining following osteogenic medium incubation and PEMF stimulation for 21 days. (H and I) qRT-PCR and Western blotting analyses of the expression of Col1a1, Cbfα1, and Osx following osteogenic medium incubation and PEMF stimulation for 3 days. (J and K) Ku70 expression assays of primary osteoblasts infected with shIFT88 lentivirus based on immunofluorescence staining and Western blotting (n = 6). Graphs represent means ± SD. (C) **P < 0.01 versus primary osteoblasts by a chi-square test. (D to K) P < 0.05; *P < 0.01. All analyses were done using two-way ANOVA with Bonferroni’s posttest. Scale bars, 20 μm (C and J) and 50 μm (E and G).
PEMF-induced osteoblast Ca2+ oscillations depend on extracellular Ca2+ influx via ciliary PC-1/2 and ER Ca2+ release
We next investigated the mechanisms underlying the Ca2+ response of irradiated osteoblasts to PEMF stimulation. PEMF-induced cytosolic Ca2+ oscillations were weakened (characterized by decreased Ca2+ spike number and magnitude) in irradiated osteoblasts infected with shIFT88 lentivirus (inhibition of ciliogenesis) as compared with cells infected with shCtrl (Fig. 5A and fig. S12A). Then, we examined the change in PEMF-mediated Ca2+ signaling after a treatment with antagonists for various potential Ca2+ channels in primary cilia, including TRPV4, Piezo1, and the PC-1/2 complex. PEMF-induced Ca2+ oscillations in irradiated osteoblasts were not significantly altered after the inhibition of TRPV4 or Piezo1 (Fig. 5, B and C). However, the blockade of PC-1 and PC-2 with short hairpin RNAs (shRNAs) (shPC-1-#3 and shPC-2-#2; fig. S10, C and D) alone or in combination almost completely abolished PEMF-induced Ca2+ oscillations in irradiated osteoblasts (Fig. 5D). Both PC-1 and PC-2 not only were abundant in irradiated primary osteoblasts but also were mainly expressed on primary cilia rather than plasma membrane according to the immunofluorescence colocalization analysis (Fig. 5E). To determine whether the N-terminal G protein–coupled receptor proteolytic site (GPS) cleavage of PC-1 is necessary for the channel activity of the PC-1/2 complex in response to PEMF stimulation, we constructed the PC-1 GPS cleavage mutant PC-1-L3040H (Fig. 5F) and found that PEMF-induced Ca2+ oscillations in irradiated osteoblasts were blocked after PC-1 GPS mutation (Fig. 5G). The PEMF-induced enhancement of cell differentiation and mineralization of irradiated osteoblasts was almost completely abolished after infected with shPC-1 and shPC-2 lentivirus alone or in combination (fig. S13). Furthermore, the PC-1 and PC-2 protein expression levels in irradiated MLO-Y4 osteocytes were significantly lower than those in irradiated primary osteoblasts (fig. S14). In particular, the PC-1 protein was rarely expressed in irradiated osteocytes but abundantly expressed in irradiated osteoblasts. Thus, the differentiated expression of the PC-1/2 proteins mediated distinct PEMF sensitivity between irradiated osteoblasts and irradiated osteocytes. We then studied the role of intracellular and extracellular Ca2+ pools in PEMF-mediated osteoblast cytosolic Ca2+ oscillations. After being incubated with Ca2+-free medium, irradiated osteoblasts no longer exhibited any Ca2+ response under PEMF stimulation (Fig. 5H), indicating that the PEMF-induced osteoblast Ca2+ response is initiated by the extracellular Ca2+ influx. In addition, after the depletion of the Ca2+ stores in the ER using thapsigargin, an approximately 80% decrease in the Ca2+ spike number was observed in irradiated osteoblasts under PEMF stimulation (Fig. 5I). Moreover, PEMF-induced Ca2+ oscillations were partially inhibited by a single blockade of inositol 1,4,5-trisphosphate (IP3R) or ryanodine receptor (RyR) and almost completely abolished under a combined treatment with inhibitors of IP3R and RyR (Fig. 5J). Our results reveal that the maintenance of the Ca2+ oscillation profile depends on mobilization from the ER Ca2+. The corresponding quantitative parameters of Ca2+ dynamics following various treatments with inhibitors/lentivirus, including Ca2+ spike magnitude and percentage of cells exhibiting Ca2+ response, are shown in fig. S12.
Fig. 5. PEMF-induced intracellular Ca2+ oscillations in irradiated osteoblasts depend on extracellular Ca2+ influx via ciliary PC-1/2 and subsequent Ca2+ mobilization from the ER.
(A) Intracellular Ca2+ signaling in irradiated osteoblasts infected with shIFT88 lentivirus (inhibition of ciliogenesis) under PEMF stimulation (n = 80 to 120 cells per group). (B to D) Intracellular Ca2+ signaling in irradiated osteoblasts pretreated with the inhibitor of (B) TRPV4 (HC067047, 10 μM) or (C) Piezo1 (GsMTx4, 5 μM) or (D) infected with shPC-1/2 lentivirus and exposed to PEMF (n = 80 to 120 cells per group). (E) Immunofluorescence colocalization of PC-1/2 with AC-α-tubulin (a primary cilium maker) in irradiated osteoblasts. Scale bar, 20 μm. (F) Schematic representation of PC-1 G protein–coupled receptor proteolytic site (GPS) cleavage mutant PC-1-L3040H. The cleavage site is marked by an arrow. (G) Intracellular Ca2+ signaling in irradiated osteoblasts after PC-1 GPS mutation under PEMF stimulation (n = 80 to 120 cells per group). (H and I) Intracellular Ca2+ signaling in irradiated osteoblasts pretreated with Ca2+-free medium or thapsigargin [the SERCA (Sarco/endoplasmic reticulum Ca2+-ATPase) inhibitor to deplete Ca2+ store in the ER, 1 μM) and exposed to PEMF (n = 80 to 120 cells per group]. (J) Intracellular Ca2+ signaling in irradiated osteoblasts pretreated with the inhibitor of IP3R (2-APB, 100 μM), RyR (ryanodine, 50 μM), or their combination and exposed to PEMF (n = 80 to 120 cells per group). Graphs represent means ± SD. (A to C and G to I) P < 0.05; *P < 0.01. Statistical analyses were done using two-way ANOVA with Bonferroni’s posttest. (D and J) P < 0.05; *P < 0.01. Statistical analyses were done using three-way ANOVA with Bonferroni’s posttest.
PEMF-induced osteoblast Ca2+ oscillations activate the Ras/MAPK/AP-1 axis, thereby regulating Ku70 transcription
Next, we identified Ca2+ oscillation–mediated downstream events that regulate Ku70 transcription in irradiated osteoblasts under PEMF stimulation. Gene set enrichment analysis (GSEA) based on RNA sequencing (RNA-seq) revealed the significant enrichment of KRAS signaling genes in irradiated primary osteoblasts in response to PEMF stimulation (Fig. 6A). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses showed that the activation of the KRAS signaling pathway is one of the events strongly affected under the PEMF stimulation (Fig. 6B). Western blotting showed that PEMF resulted in the up-regulation of phosphorylated MAPK kinase 1/2 (p-MEK1/2), phosphorylated extracellular signal–regulated kinase 1/2 (p-ERK1/2), and phosphorylated c-Jun N-terminal kinase (p-JNK) protein expression in irradiated osteoblasts (Fig. 6C and fig. S15A), which confirmed the PEMF-induced activation of Ras/MAPK signaling. After blocking ciliogenesis using shIFT88 lentivirus, the PEMF-induced increase in p-MEK1/2, p-ERK1/2, and p-JNK expression in irradiated osteoblasts was blunted (Fig. 6D and fig. S15B), suggesting that the PEMF-induced activation of Ras/MAPK signaling highly depends on the involvement of primary cilia. Similarly, the PEMF-mediated up-regulation of the p-MEK1/2, p-ERK1/2, and p-JNK expression was mitigated after the depletion of the Ca2+ stores in the ER using thapsigargin (Fig. 6E and fig. S16A). AP-1 is known as one of the key downstream transcription factors of Ras/MAPK signaling; our GSEA analyses showed a significantly elevated expression of AP-1–regulated genes (Fig. 6F). Western blotting confirmed the AP-1 activation, as evidenced by a significant increase in p-c-Fos and p-c-Jun expression in irradiated osteoblasts under PEMF stimulation (Fig. 6G and fig. S17A). After the depletion of primary cilia, PEMF-mediated increases in p-c-Fos and p-c-Jun were also abolished (Fig. 6H and fig. S17B). We also found that AP-1 had potential binding sites that could recognize the promoter sequence of Ku70 according to transcription factor prediction (Fig. 6I). Then, we truncated the predicted binding sites (with promotor sequence −1100 to +1) of AP-1 recognizing Ku70 into three successively smaller fragments (Fig. 6I), thereby creating the plasmid constructs TRUNC0 [~1100 base pairs (bp)], TRUNC1 (~600 bp), and TRUNC2 (~500 bp). The promoter activities of all four shaved fragments are shown in Fig. 6J. Luciferase reporter assays revealed that the TRUNC1 promoter region (−1022 to −1011) instead of the TRUNC2 promoter region (−568 to −560) was the binding site of AP-1, which recognizes Ku70. Furthermore, the luciferase activity was significantly inhibited in radiated osteoblasts under PEMF stimulation after the blockade of AP-1 (Fig. 6K). The chromatin immunoprecipitation (ChIP) assays confirmed the binding of AP-1 to the predicted Ku70 promoter (Fig. 6L). The PEMF-induced increase in Ku70 expression was significantly inhibited after the treatment with the AP-1 antagonist SR11302 (Fig. 6M and fig. S17C). Similarly, PEMF-mediated up-regulation of the Ku70, p-c-Fos, and p-c-Jun expression was also blocked after the incubation with farnesyl thiosalicylic acid (FTS; the Ras inhibitor; Fig. 6N and fig. S18A) and thapsigargin (depletion of ER Ca2+ stores; Fig. 6O and fig. S16B). In addition, the PEMF-induced enhancement of cell differentiation and mineralization of irradiated osteoblasts was blunted after the depletion of the ER Ca2+ stores (fig. S16, C to E), blockade of KRAS (fig. S18, B to D), and inhibition of AP-1 (fig. S17, D to F).
Fig. 6. PEMF-induced Ca2+ oscillations activate the Ras/MAPK/AP-1 signaling axis and thereby regulate Ku70 transcription in irradiated osteoblasts.
(A and B) RNA-seq showing the significant enrichment of KRAS signaling in irradiated osteoblasts under PEMF stimulation for 2 hours. (C) Western blotting for the expression of kinase proteins in the Ras/MAPK pathway (n = 6). (D and E) Expression of kinase proteins in the Ras/MAPK pathway in irradiated osteoblasts infected with shIFT88 lentivirus or pretreated with thapsigargin (n = 6). (F) RNA-seq showing increased AP-1 signaling expression. (G) Western blotting for the phosphorylated c-Fos and c-Jun expression (n = 6). (H) AP-1 protein expression in irradiated osteoblasts infected with shIFT88 lentivirus (n = 6). (I) Schematic representation of the AP-1 promoter region (~1100 bp). Green box indicates the luciferase gene location. Bar graph indicates the relative luciferase activity of TRUNC0, TRUNC1, and TRUNC2 promoter regions. PGL3-basic is the negative control. (J) Relative luciferase activity analysis of TRUNC0, TRUNC1, and TRUNC2 (n = 6). (K and L) Relative luciferase activity and ChIP analyses in irradiated osteoblasts pretreated with the AP-1 inhibitor (SR11302, 10 μΜ) (n = 6). (M to O) Western blotting for the AP-1 and Ku70 expression in irradiated osteoblasts pretreated with the inhibitor of AP-1 (SR11302, 10 μΜ), Ras (FTS, 50 μM), or SERCA (thapsigargin, 1 μM) (n = 6). Graphs represent means ± SD. (I) P < 0.01 versus the TRUNC0 group by one-way ANOVA with Bonferroni’s posttest. (J to L) P < 0.01. Statistical analyses were done using two-way ANOVA with Bonferroni’s posttest.
Ku70 knockdown in osteoblasts blunts PEMF-induced bone gain in irradiated mice
We next generated a Ku70 conditional knockout (cKO) mouse model by crossing Ku70flox/flox mice with Col1-Cre mice to further elucidate the role of Ku70 in PEMF-induced bone protection in irradiated mice (Fig. 7A). The generation and confirmation based on quantitative real-time polymerase chain reaction (qRT-PCR) and immunofluorescence of this animal model are shown in fig. S19. The Ku70flox/flox; Col1-Cre (Ku70 cKO) mice showed a significant decrease in trabecular bone mass as compared with the wild-type (WT) mice based on micro-CT and H&E staining assays (Fig. 7, B and C). The Ku70 cKO mice exhibited much more trabecular bone decrease on the radiation side than on the contralateral side compared with the WT mice (Fig. 7, B and C), suggesting that mice with osteoblast Ku70 deletion are highly vulnerable to irradiation. Furthermore, no significant difference in trabecular bone mass was observed on the radiation side between the Ku70 cKO mice and Ku70 cKO + PEMF mice (Fig. 7, B and C), revealing that mice with osteoblast Ku70 knockdown are resistant to the PEMF treatment. Serum ELISA assays showed decreased OCN and P1NP secretion in Ku70 cKO mice compared with WT mice, while PEMF increased serum OCN and P1NP concentrations (Fig. 7D). In addition, no significant difference in serum biomarkers of bone resorption (CTX-1 and TRACP-5b) was observed among the WT, Ku70 cKO, and Ku70 cKO + PEMF mice (Fig. 7E). The osteoblast number/survival and bone formation rate on the trabecular bone surface in the Ku70 cKO mice were significantly lower than those in the WT mice according to calcein/Alizarin labeling and Runx2 and TUNEL immunostaining (Fig. 7, F to H). The Ku70 cKO mice exhibited higher decrease in osteoblast number/survival and bone formation on the radiation side than on the contralateral side compared with WT mice (Fig. 7, F to H). The Ku70 cKO + PEMF mice exhibited no difference in osteoblast number/survival and bone formation on the radiation side, in contrast to the Ku70 cKO mice (Fig. 7, F to H). PEMF also exhibited no observable change in the body weight or food intake of mice subjected to focal radiation (fig. S9B).
Fig. 7. Deletion of Ku70 in osteoblasts blunts PEMF-induced bone gain in irradiated mice.
(A) Schematic representation of focal radiation and PEMF administration in Ku70flox/flox; Col1-Cre (Ku70 cKO) mice. (B) Representative micro-CT images and quantitative analyses of distal femora in mice (n = 6). (C) H&E staining showing the trabecular area of proximal tibiae in mice (n = 6). (D and E) Serum ELISA assays for bone formation markers (OCN and P1NP) and bone resorption markers (CTX-1 and TRACP-5b) (n = 6). (F) Dynamic bone histomorphometric analyses based on calcein/Alizarin red labeling (n = 6). (G) Runx2 immunohistochemical staining for labeling osteoblasts on bone surface. (H) TUNEL staining for detection of apoptotic osteoblasts on bone surface. Graphs represent means ± SD. (B, C, and F to H) P < 0.05; *P < 0.01. Statistical analyses were done using three-way ANOVA with Bonferroni’s posttest. (D and E) P < 0.05; *P < 0.01. Statistical analyses were done using one-way ANOVA with Bonferroni’s posttest. Scale bars, 100 μm (C) and 50 μm (F to H). WT, wild-type mice; Ku70 cKO, Ku70flox/flox; Col1-Cre+ mice.
DISCUSSION
Radiotherapy-induced bone damage represents an increasingly important clinical challenge. However, effective countermeasures are still lacking, especially considering the risk of tumor recurrence. Following radiotherapy, many patients fail bone treatment, which leaves them at a substantially higher risk of fragility fractures and osteonecrosis (42). Here, on the basis of a systemic parameter optimization through Ca2+ dynamics analyses, we identified a previously unidentified PEMF waveform at 15 Hz and 2 mT that effectively resisted radiation-induced bone loss by inducing the specific activation of osteoblasts with no impact on tumor cells. Mechanistic evidence showed that primary cilia were major PEMF detectors in osteoblasts, and the differentiated ciliary expression dominated the distinct PEMF sensitivity between osteoblasts and tumor cells. Moreover, PEMF-induced unique Ca2+ oscillations depended on extracellular Ca2+ influx via ciliary PC-1/2 and ER Ca2+ release, thereby activating the Ras/MAPK/AP-1 axis, and eventually regulating Ku70 transcription and DNA repair. These findings provide an effective method to reduce radiotherapy-induced bone damage in a noninvasive, cost-effective, and highly specific manner.
Cytosolic Ca2+ oscillations encode the sensitivity of nonexcitable cells to external stimuli and then induce much more efficient activation of downstream target effectors than constant Ca2+ signals (38, 43). We observed oscillatory Ca2+ signaling events in irradiated osteoblasts instead of osteoclasts or osteocytes in response to various EMF waveform parameters, revealing that osteoblasts act as the major EMF sensors in bone. In contrast to sin-EMF and sPEMF, the unique PEMF waveform triggers much stronger cytosolic Ca2+ oscillations with robust and repetitive Ca2+ spikes. Furthermore, our in vivo results demonstrate that this PEMF parameter restored bone mass and mechanical properties in irradiated hindlimbs to the control level by rescuing osteoblasts. Considering that, among all bone cell types, osteoblasts are particularly sensitive to ion radiation, this PEMF regimen, which induces the specific activation of osteoblasts, seems to be a promising and highly efficient approach against radiation-induced bone damage. In addition, although PEMF at 15 Hz and 2 mT, the parameter inducing most notable osteoblastic Ca2+ oscillations, was found to be effective against radiation-induced bone damage, the possibility of the variation in the optimal PEMF exposure parameter from in vitro to in vivo transition cannot be excluded due to the thick soft tissues and complicated in vivo scenario, which may need to be validated by more experimental and clinical studies.
The mechanism by which PEMF resists radiotherapy-induced bone loss is schematically summarized in Fig. 8. Primary cilia are sensory organelles owing to their unique lipid and receptor composition that detect and translate various extracellular mechanical cues (e.g., fluid flow, pressure, and vibration), especially in kidney and bone cells (31, 44). Primary cilia are also considered tumor suppressor organelles, while cilia loss increases the incidence of tumorigenesis (37, 45). Similar to previous findings (32–37), our results confirm that various types of cancer cells, including breast cancer, colon cancer, melanoma, and osteosarcoma cells, were devoid of primary cilia. We also found that osteoblasts were rich in the expression of primary cilia, and the PEMF-induced increase in osteoblast survival and differentiation was blocked after the depletion of primary cilia. Thus, our findings reveal that the primary cilia of osteoblasts act as critical EMF sensory organelles in the skeleton. The PEMF treatment is endowed with the ability to resist radiation-induced bone loss with a highly specific nature owing to the differentiated expression of primary cilia between osteoblasts and tumor cells.
Fig. 8. Schematic drawing of the mechanism by which PEMF specifically resists radiotherapy-induced bone loss.
(A) Osteoblasts are major PEMF sensors in bone and specifically sensitive to the PEMF waveform. The differential expression of primary cilia between osteoblasts and tumor cells mediates their highly distinct sensitivity to PEMF stimulation. (B) Primary cilia are the major sensory organelles in irradiated osteoblasts detecting external PEMF stimulation. Osteoblasts transform external biophysical PEMF signals into intracellular unique oscillatory Ca2+ signaling with repetitive robust Ca2+ spikes, which depends on extracellular Ca2+ influx via ciliary polycystins PC-1/2 (by cleaving the GPS site in PC-1) and Ca2+ release from the ER. PEMF-induced Ca2+ oscillations thereby activate the Ras/MAPK/AP-1 axis, enhance the nuclear transcription of Ku70 (a core repair protein of DNA DSBs) in irradiated osteoblasts, and eventually protect against radiotherapy-induced bone loss.
PC-1 is a large transmembrane protein that forms a heterodimer via its intracellular C-terminal tail with PC-2 (a Ca2+-permeable selective channel) (46). The extracellular region of PC-1 has a GPS site, and the GPS cleavage results in an N-terminal and C-terminal fragment (47). In kidney cells, PC-1 functions as a mechanosensor depending on its long extracellular N terminus, thereby regulating Ca2+ influx through PC-2 (47, 48). We found a high percentage of PC-1 and PC-2 (>80%) colocated at the primary cilia rather than the plasma membrane in osteoblasts. In addition to PC-1/2, TRPV4 and Piezo1 are important Ca2+ channels on the primary cilia of osteoblasts, whereas they also have extensive expression on the cell membrane (49, 50). Furthermore, our findings revealed that cytosolic Ca2+ oscillations were almost completely abolished after the blockade of PC-1/2 instead of TRPV4 and Piezo1. The cytosolic Ca2+ response to PEMF was significantly weakened but not completely abolished after the depletion of primary cilia in irradiated osteoblasts, which might be due to the existence of a small amount of PC-1/2 on the plasma membrane. Several studies also suggest that mechanically induced cytosolic Ca2+ signaling originates from primary cilia in cholangiocytes and kidney cells, and the blockade of primary cilia weakens cytosolic Ca2+ signaling (51–53). Coupled with the colocalization results, our findings reveal that the PC-1/2 complex is a unique channel at the primary cilia mediating PEMF-induced Ca2+ influx. Moreover, we found that PEMF induced the GPS cleavage of PC-1 to open the PC-1/2 channels. It has been shown that PC-2 physically interacts with both IP3R and RyR through its C terminus, thereby affecting the Ca2+ dynamics in the ER and amplifying cytosolic Ca2+ signaling (54, 55). Our results demonstrate that the ER is critical for maintaining a unique Ca2+ oscillation profile with multiple robust Ca2+ spikes in radiation osteoblasts in response to PEMF.
The PEMF waveform used in this study is generated on the basis of a low-frequency sPEMF (5 to 100 Hz) that is modulated by another carrier wave with the frequency of 4.55 kHz. It has been shown that electromagnetic stimulation with the frequencies of 1 to 5 kHz (in the range of the intermediate frequencies) can induce higher probability of the conformational changes of membrane receptors/channels (56, 57). We found that PEMF (but not sPEMF or sin-EMF) triggered unique intracellular Ca2+ oscillations in irradiated osteoblasts. The intermediate frequency component (i.e., the 4.55-kHz carrier wave) might be the primary factor mediating these Ca2+ oscillations. When osteoblasts are exposed to PEMF, the intermediate-frequency component (4.55 kHz) of PEMF that facilitates the conformational changes of ciliary PC-1 leads to the GPS cleavage by cis-autoproteolysis (58). Furthermore, studies have shown that the cyclotron frequency of Ca2+ is typically around 15 Hz (59, 60). According to the ion cyclotron resonance theory, the motion of Ca2+ through the receptors/channels in cell membranes is enhanced when the exogenous EMF is tuned to the cyclotron resonance frequency, and thus, biological effects appear (59, 60). Thus, the low-frequency component of PEMF might facilitate the cyclotron resonance of Ca2+ in osteoblasts. Collectively, the coupled low-frequency and intermediate-frequency components might be necessary and sufficient for the PC-1/2 channel activation and thus induce the unique intracellular Ca2+ oscillations in irradiated osteoblasts.
Unlike ion radiation, which has a high penetration capacity to directly target DNA, PEMF mainly acts on the cell membrane owing to its much lower frequency. How does PEMF promote DNA damage repair in osteoblasts by initiating the signaling cascade from the cell surface to the nucleus? Ku70 is a “double-sided tape” protein responsible for the recognition and binding of DNA DSB ends and subsequent stability maintenance of the NHEJ complex (19). In addition, Ku70 is able to suppress cell apoptosis by interacting with the proapoptotic protein Bax and antiapoptotic Mcl-1, which is independent of DSB repair (20). Our in vitro and in vivo data revealed that PEMF significantly increased Ku70 expression in irradiated osteoblasts. Moreover, the PEMF-mediated recovery of osteoblasts and bone damage was blunted after Ku70 knockdown, suggesting that Ku70 is the core nuclear protein for the protection of PEMF against irradiation-induced bone damage. In particular, we established osteoblast-specific Ku70 knockout mice and found that these mice exhibited a mild osteopenia phenotype, which was associated with decreased osteoblast survival and differentiation owing to increased genomic instability. As expected, the skeleton of this mouse model was more vulnerable to ion radiation. The irradiated skeletons of these mice were also found to be resistant to the PEMF treatment. Thus, our results provide strong evidence that Ku70 is the target effector molecule of irradiated osteoblasts in response to PEMF.
We then explored the downstream signaling pathway of Ca2+ oscillations that mediate the Ku70 transcription. Initially using RNA-seq and further expression verification, we found that PEMF caused a specific high expression of AP-1 in irradiated osteoblasts. We also found that AP-1 bound to the Ku70 promoter and regulated Ku70 transcription. AP-1, a heterodimeric transcription factor consisting of the Jun and Fos families, regulates substantial cellular processes including proliferation, apoptosis, and DNA repair (61). Ku70 was found to bind to AP-1 sites in the liver and brain (62). Next, our GSEA and protein quantitative analyses showed that PEMF promoted the specific activation of the molecular expression of the Ras/MAPK pathway, an important upstream pathway of AP-1 (63). Ras also acts as a downstream decoder of Ca2+ signaling, and Ca2+ oscillations optimize the activation of Ras/MAPK signaling (64). We found that the blockade of PEMF-induced Ca2+ oscillations suppressed the expression of Ras/MAPK signaling, and the inhibition of Ras/MAPK decreased the AP-1 and Ku70 expression. Thus, Ras/MAPK signaling is the pathway that bridges Ca2+ signaling and the transcription factor AP-1.
In conclusion, our study identified a unique physiological effect of exogenous PEMF stimulation in the context of ion radiation exposure, from the PEMF parameter optimization to cellular PEMF detection, intracellular signaling response, and transduction, until eventual nuclear DNA damage repair. These findings provide strong evidence for the therapeutic potential of PEMF as a noninvasive approach against radiotherapy-induced bone loss.
Club Recharge - 14490 Pearl Road - Strongsville - OH 44136.
Hours: Monday-Friday 10AM-7PM - Saturday 10AM-3PM
(Phone: 440-567-1146)
Comments