Figures and data

Osteocytes inhibited the proliferation of NSCLC cells.
a. Representative immunohistochemical staining and quantification of Ki-67 in bone metastasis site of patients with NSCLC. The dashed line indicates the boundary between bone and tumor. Scale bars: 50 µm. Blue arrows, low expression or negative of Ki67. Red arrows, high expression of Ki67. Tumor cells ≤50μm from bone tissue were defined as adjacent cells, and those > 50μ m from bone tissue were defined as surrounding cells. n=6; Student’s two-sided unpaired t test.
b. An intraosseous model of bone metastasis was used via direct implantation of NSCLC cells A549 into tibia of nude mice. Three weeks after implantation, mice were sacrificed. Representative immunohistochemical staining and quantification of Ki-67 in tibia of tumor-bearing nude mice. The dashed line indicates the boundary between bone and tumor. Scale bars: 50 µm. Blue arrows, low expression or negative of Ki67. Red arrows, high expression of Ki67. Tumor cells ≤50μm from bone tissue were defined as adjacent cells, and those > 50μm from bone tissue were defined as surrounding cells. n=6; Student’s two-sided unpaired t test.
c. Schematic diagram of co-culture.
d. CCK-8 assays were performed to evaluate the effects of MLO-Y4 cells, MC3T3-E1 or NCM460 cells on the proliferation of A549 cells in the co-culture system. n=6; one-way analysis of variance with Turkey’s multiple comparisons test.
e. EdU flow cytometry were performed to evaluate the effect of MLO-Y4 cells on the proliferation of A549 cells in the co-culture system.
f. CCK-8 assays were performed to evaluate the effect of MLO-Y4 cells on the proliferation of MC3T3-E1 cells in the co-culture system. n=6; Student’s two-sided unpaired t test.
no significance (ns) P>0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Osteocyte sEVs inhibited the proliferation of NSCLC cells.
a. Schematic diagram of serial centrifugation and concentration of culture medium. Conditioned medium (CM) was separated into soluble factor (SF), large extracellular vesicles (lEVs), and small extracellular vesicles (sEVs) fractions, by serial ultracentrifugation and ultrafiltration.
b. CCK-8 assays were performed to evaluate the effect of sEVs on the proliferation of NSCLC cells. n=4, two-way analysis of variance with Sidak’s multiple comparisons test.
c. EdU flow cytometry were performed to evaluate the effect of sEVs on the proliferation of NSCLC cells.
d. Western blot analysis of the typical sEVs markers CD63, TSG101, and Alix. n=3; Student’s two-sided unpaired t test.
e. The concentration of the sEVs were detected by NanoSight Analysis. n=3; Student’s two-sided unpaired t test.
f. CCK-8 assays were performed to evaluate the rescue effect of GW4869 on the proliferation of NSCLC cells co-cultured with MLO-Y4 cells. n=4, two-way analysis of variance with Sidak’s multiple comparisons test.
g. EdU flow cytometry were performed to evaluate the rescue effect of GW4869 on the proliferation of NSCLC cells co-cultured with MLO-Y4 cells.
no significance(ns) P>0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Osteocyte sEVs miR-99b-3p inhibited the proliferation of NSCLC cells by directly targeting MDM2.
a. Comparison of relative miR-365-1-5p, miR-99b-3p, miR-193a-5p, miR-125a-3p, miR-690, miR-1195, miR-122-5p, miR-140-3p content between MLO-Y4-sEVs and MC3T3-E1-sEVs by qRT-PCR. n=3, student’s two-sided unpaired t test.
b. NSCLC cells were transfected with miR-365a-5p, miR-30c-1-3p, miR-193a-5p, miR-99b-3p mimics, or NC (negative control) in DMEM medium. 48 hours later, CCK-8 assays were performed. n=6, one-way analysis of variance with Turkey’s multiple comparisons test.
c, d. NSCLC cells were transfected with miR-99b-3p mimics, NC mimics, miR-99b-3p inhibitor and NC inhibitor. 48 hours later, EdU flow cytometry were performed. CCK-8 assays were performed at different points in time after transfection. n=6, two-way analysis of variance with Sidak’s multiple comparisons test.
e. Schematic diagram of putative miR-99b-3p binding sites in the MDM2 3′-UTR. Green letters denote mutation sites. Relative luciferase activities of wild-type (WT) or mutant (MUT) MDM2 3′-UTRs were determined in A549 cells, which were co-transfected with the miR-99b-3p mimics or negative control. Luciferase activity was normalized using Renilla. n=4, Student’s two-sided unpaired t test.
f. MDM2 protein in A549 cells were analyzed by Western blot 72 hours after transfection. n=3; one-way analysis of variance with Turkey’s multiple comparisons test.
no significance(ns) P>0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Mechanical stimulation increased the release of sEVs from osteocytes and enhanced inhibitory effect of osteocytes on the proliferation of NSCLC cells.
a. MLO-Y4 cells were mechanically stimulated by stretching at 12% strain at a 1.25 Hz frequency for 12 h, and sEVs were extracted from the conditioned medium. CCK-8 assays were performed to evaluate the effect of sEVs on the proliferation of A549 cells. n=4, two-way analysis of variance with Sidak’s multiple comparisons test.
b. EdU flow cytometry were performed to evaluate the effect of sEVs on the proliferation of A549 cells.
c. Western blot analysis of the typical sEVs markers CD63, TSG101, and Alix. n=3, student’s two-sided unpaired t test.
d. Concentration of sEVs was detected by NanoSight analysis. n=3, student’s two-sided unpaired t test.
e. Mechanical loading was applied to the tibiae with bone metastasis via direct implantation of GFP-LLC into tibia of mouse. qRT-PCR identification of miR-99b-3p in sEVs. n=3, student’s two-sided unpaired t test.
An intraosseous model of bone metastasis was used via direct implantation of murine Lewis lung carcinoma (LLC) cells expressing green fluorescent protein (GFP) into tibia of mouse. Mice were subsequently randomized into tumor-bearing only (control), tibial loading, tibial loading+antagomiR-NC and tibial loading+antagomiR-99b-3p groups. Four weeks later, mice were sacrificed.
f, i. Representative biophotonic images and quantification of fluorescence signal in hindlimb bones of mice. n=6, one-way analysis of variance with Turkey’s multiple comparisons test.
g, j. Representative immunofluorescence staining images of Ki-67(green), DMP1(red), and DAPI (blue) in the tibia of mice with bone metastases and quantification of the number of Ki-67-positive cells in the images. The dashed line indicates the boundary between bone and tumor. Scale bars: 50 µm. n=6, one-way analysis of variance with Turkey’s multiple comparisons test.
h, k. Representative micro-CT imaging and quantitation of the tibia (bone/total volume) in mice. n=6, one-way analysis of variance with Turkey’s multiple comparisons test.
no significance(ns) P>0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Moderate exercise combined with zoledronic acid effectively suppressed the progression of bone metastasis of NSCLC.
An intraosseous model of bone metastasis was used via direct implantation of murine Lewis lung carcinoma (LLC) cells expressing green fluorescent protein (GFP) into tibia of mouse. Mice were subsequently randomized into tumor-bearing only (control), zoledronic acid (ZA), treadmill running, and treadmill running combined zoledronic acid (treadmill running+ZA) groups. Four weeks later, mice were sacrificed.
a. Representative biophotonic images and quantification of fluorescence signal in hindlimb bones of mice. n=6, one-way analysis of variance with Turkey’s multiple comparisons test.
b. Representative immunofluorescence staining images of Ki-67(green), DMP1(red), and DAPI (blue) in the tibia of mice with bone metastases and quantification of the number of Ki-67-positive cells in the images. The dashed line indicates the boundary between bone and tumor. Scale bars: 50 µm. n=6, one-way analysis of variance with Turkey’s multiple comparisons test.
c-e Representative micro-CT imaging and quantitation of the tibia (bone/total volume, cortical bone thickness, trabecular bone thickness) in mice. n=6, one-way analysis of variance with Turkey’s multiple comparisons test.
no significance (ns) P>0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Exercise preconditioning effectively suppressed the progression of bone metastasis of NSCLC.
An intraosseous model of bone metastasis was used via direct implantation of NSCLC cells (LLC and A549) into tibia of mouse. Mice were subsequently randomized into tumor-bearing only (control), treadmill running before and after implantation of NSCLC cells into tibia (exercise preconditioning), and treadmill running after implantation of NSCLC cells into tibia (treadmill running) groups. For exercise preconditioning group, mice were subjected to 4 weeks of treadmill exercise before implantation of NSCLC cells into tibia of mouse. Four weeks after implantation, mice were sacrificed.
a. Representative biophotonic images and quantification of fluorescence signal in hindlimb bones of mice. n=6, one-way analysis of variance with Turkey’s multiple comparisons test.
b. Representative immunofluorescence staining images of Ki-67(green), DMP1(red), and DAPI (blue) in the tibia of mice with bone metastases and quantification of the number of Ki-67-positive cells in the images. The dashed line indicates the boundary between bone and tumor. Scale bars: 50 µm. n=6, one-way analysis of variance with Turkey’s multiple comparisons test.
c-e Representative micro-CT imaging and quantitation of the tibia (bone/total volume, cortical bone thickness, trabecular bone thickness) in mice. n=6, one-way analysis of variance with Turkey’s multiple comparisons test.
no significance (ns) P>0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

A hypothetic model depicting osteocytes induce and maintain tumor dormancy in bone metastasis of non-small cell lung cancer.
Osteocytes, sensing mechanical stimulation generated by exercise or mechanical loading, inhibit the proliferation and sustain the dormancy of NSCLC cells by releasing sEVs with tumor suppressor miRNAs, such as miR-99b-3p, which inhibits the proliferation of NSCLC cells by directly targeting MDM2.