MiR-335 Inhibits Small Cell Lung Cancer Bone Metastases via IGF-IR and RANKL Pathways

This article is featured in Highlights of This Issue, p. 1

Small cell lung cancer (SCLC), a highly aggressive form of lung cancer associated with tobacco smoking (1, 2), represents 10% to 15% of all lung cancers (3, 4). Treatment is challenging because SCLC grows rapidly, often becoming well established in lung before becoming symptomatic; metastatic spread occurs early and rapidly (5, 6). For patients with localized or regionalized SCLC, 5-year survival in the United States in 2008 was 52% and 25%, respectively (7). For those who suffer metastatic SCLC or relapse, the prognosis is worse. Although there have been advances in therapy for localized SCLC with the introduction of positron emission tomography, SCLC mortality has remained unchanged over the past 30 years (6).

Compared with our knowledge of skeletal metastases in breast and prostate cancers, we know very little about cell and molecular mechanisms of skeletal metastases in lung cancer, especially in SCLC. In osteolytic lesions, such as those in breast cancer, bone loss predominates, while increased bone turnover in which both bone formation and bone resorption are deregulated, is characteristic of skeletal metastases in prostate cancer (8, 9). More than one third of patients with SCLC develop osteolytic bone metastasis, resulting in severe pain, pathologic fractures, spinal cord compression, and loss of mobility to greatly reduce the quality of life (10, 11). Because of the lack of therapeutic options in SCLC, there is an urgent need to better understand how skeletal progression in SCLC may be controlled and treated.

microRNAs (miRNA) critically regulate tumorigenesis and progression by targeting oncogenes, tumor suppressor genes, or genes related to proliferation, angiogenesis, and apoptosis. miRNAs are a class of small noncoding RNAs of about 19 to 25 nucleotides (nt) that function as negative posttranscriptional gene regulators (12, 13). By hybridizing to the 3′ untranslated region (UTR) of target mRNAs, miRNAs can serve as mediators of mRNA cleavage and cause translational repression, or act on transcription to reduce protein output by destabilizing mRNAs (14). Different tumor types and tumors at various differentiation stages may exhibit unique miRNA profiles (15). In non–small cell lung cancer (NSCLC) cells, miRNA (miR)-494, miR-30a, miR-193b, miR-101, miR-7, and miR-206 have been reported as tumor inhibitors (16–21), whereas miR-212 is believed to promote carcinogenesis in vitro (22). In patients with NSCLC, high miR-155 and low let-7a-2 expression in tumor tissue have been reported to correlate with poor survival (23). Unfortunately, miRNAs, which have been associated with clinical outcomes in metastatic NSCLC, have not been found relevant to SCLC (24). More research is needed to investigate the selectivity and specificity of miRNA pathways in SCLC and its metastatic spread to bone, and their possible value as therapeutic targets to improve survival.

Human SBC-5 and SBC-3 cell lines were originally established from human SCLC, and found to differ in their predilection for bone when tested as xenografts in immunodeficient mice (25, 26). Xenografts of SBC-5 or SBC-3 cells formed multiple tumor foci in liver, pancreas, ovary/uterus, and kidney in natural killer (NK)-depleted immunodeficient mouse models. SBC-5 cell xenografts promoted osteolytic bone lesions, whereas SBC-3 cell xenografts did not (25, 26). We previously reported a more consistent immunodeficient mouse host model for xenografts of bone metastases, using human SCLC SBC-5 cells injected into the tail vein of nonobese diabetic/severe combined immunodeficient (NOD/SCID) IL2R_γ^null mice, deficient in T cells, B cells, and NK cells (27). SBC-5 and SBC-3 xenografts homed to multiple nonskeletal tissues, such as liver, pancreas, uterus, ovary, and kidney, and formed lesions that were indistinguishable in location and morphology (27). Osteolytic bone lesions were observed throughout the skeleton of mice with SBC-5 cell xenografts, whereas no bone lesions were observed in mice with SBC-3 cell xenografts (27). We applied this model to study skeletal progression of SCLC in the preclinical studies reported here.

Our goal was to investigate miRNAs involved in the regulation of skeletal metastatic SCLC lesions. A comparison of miRNA profiles of SBC-5 with SBC-3 cell lines revealed selective downregulation in miR-335 and miR-29a in SBC-5 cells. Overexpression of miR-335, but not miR-29a, in SBC-5 cells decreased cell proliferation, colony formation, migration and invasion, and osteoclast induction in vitro, and prevented or reduced osteolytic metastases in vivo.

This study was approved by the Ethics Committees of West China Hospital of Sichuan University (Sichuan, People's Republic of China). Experiments in the United States involving mice were reviewed and approved by The Jackson Laboratory Institutional Animal Care and Use Committee.

Human SCLC SBC-5 and SBC-3 cell lines (SBC-5 and SBC-3, respectively) were obtained from the Japan Health Sciences Foundation, Health Science Resources Bank (HSRB; JCRB0819 and JCRB0818), and authenticated by DNA short tandem repeats profile assay. SBC-5 and SBC-3 were maintained in advanced Dulbecco's modified Eagle medium (DMEM) medium (Invitrogen) supplemented with 10% FBS (Thermo-Hyclone). Human kidney cell line, 293TN, was purchased from SBI (System Biosciences) and maintained in advanced DMEM medium with 10% FBS.

RNA was extracted from cells using an RNeasy miRNA kit (Qiagen Inc.,). miRNA microarrays were conducted by LC Sciences. The miRNA microarrays included 833 human miRNAs, representing miRNA transcripts listed in Sanger miRBase Release 11.0. The microarrays included four independent RNA samples from each cell line. To ensure accuracy of the hybridizations, each RNA sample was hybridized with three membranes. Hybridization signals for each spot of the array and background values at 15 empty spots were measured. Hybridization signals that failed to exceed the average background value by more than three SD were excluded from analysis. Signal intensities for each spot were calculated by subtracting the background values from the total intensities. Data normalization was conducted using positive control RNA spots [tRNA(G), tRNA(L), tRNA(T), tRNA(H), and 5S rRNA] to allow comparisons among chips. The remaining data were averaged among triplicate arrays, and the resulting four datasets, each corresponding to an RNA sample, were considered independent measurements for the purposes of paired, two-sample t test when comparing miRNA profiles in SBC-5 with SBC-3.

Reverse transcriptase (RT) and quantitative real-time PCR were conducted in a two-step reaction using Taqman miRNA assays according to the protocol provided by the manufacturer (Applied Biosystems). U6 was used as the internal control. The 2^−ΔΔCt method described by Livak and Schmittgen (28) was used to analyze the data.

A lentiviral expression system was used to establish stable SBC-5 cell lines with high miR-335 or miR-29a expression. Lenti-miR-335 or Lenti-miR-29a miRNA Precursor Expression Construct (System Biosciences) was used to prepare lentivirus with the LentiSuite (System Biosciences) according to the manufacturer's protocol. The pGreenPuro Scramble Hairpin Control Construct (System Biosciences) was used to prepare control lentivirus. Lentiviral infection was conducted according to the manufacturer's protocol. Transfected cells were trypsinized, diluted in culture medium, and seeded in 96-well plate at one cell/well in average. After 7-day culture, single-cell colonies with high fluorescence were chosen out and cultured to be stable cell lines. Quantitative reverse transcriptase PCR (qRT-PCR) was used to assay the expression of miR-335 or miR-29a in these stable cell lines. The SBC-5 cell lines with highest miR-335 expression (named as SBC-5 miR-335⁺) or miR-29a expression (named as SBC-5 miR-29a⁺) were selected for the assays described in the following sections. SBC-5 cells transfected with control lentivirus were used as control cell line (named as SBC-5 Vector^Ctrl) for in vitro and in vivo experiments.

SBC-5 miR-335⁺, SBC-5 miR-29a⁺, and SBC-5 Vector^Ctrl cell lines were serum starved for 24 hours, trypsinized and resuspended in 0.1% FBS-supplemented medium with no additional growth factors. SBC-5 cells were plated at a density of 1 × 10⁴ cells/well in a Transwell insert (3 μm pore size, BD Biosciences) for the migration assay, and in a Matrigel-coated, growth factor-reduced invasion chamber (8 μm pore size, BD Biosciences) for the invasion assay. Ten percent FBS-containing medium was added into 24-well plates to provide a chemoattractant. After 6-hour incubation for the migration assay or after 22-hour incubation for the invasion assay, cells were fixed with 4% paraformaldehyde for 1 hour. Cells on the apical side of each insert were removed by mechanically scraping. Cells located on the basal side of the membrane were stained with 0.1% crystal violet, and visualized under a Zeiss Axiovert 200M microscope. Cell numbers were quantified using Metamorph analysis software.

SBC-5 miR-335⁺, SBC-5 miR-29a⁺, and SBC-5 Vector^Ctrl cell lines were cultured to 70% to 80% confluence, serum starved for 24 hours, and then cultured at a density of 1,500 cells/well in 96-well plates with advanced DMEM medium supplemented with 2% FBS at 5% CO₂, 37ºC. At selected time points, MTT was added at a final concentration of 0.5 mg/mL. After 4-hour incubation at 37ºC, medium was removed, and purple blue sediment was dissolved in 150 μL of dimethyl sulfoxide. The relative optical density for each well was determined using a Wellscan MK3 ELISA kit (Labsystems) as a measure of proliferation.

SBC-5 miR-335⁺, SBC-5 miR-29a⁺, and SBC-5 Vector^Ctrl cell lines were seeded at a density of 100 cells/dish in 60 mm dishes, and cultured for 14 days. Cells were fixed with 4% paraformaldehyde for 20 minutes, and stained with 0.1% crystal violet for 30 minutes. Colony numbers were counted using bright-field microscopy.

SBC-5 miR-335⁺, SBC-5 miR-29a⁺, and SBC-5 Vector^Ctrl cell lines were cultured to 90% confluence, washed with PBS, and incubated at 37°C for 24 hours in advanced DMEM with 0.5% FBS. Incubation supernatants from each SBC-5 cell line were harvested, and conditioned media (CM) were prepared as 10% incubation supernatant, 10% FBS, 80% α-MEM, 10 ng/mL macrophage colony-stimulating factor (M-CSF), and 10 ng/mL RANK ligand (RANKL) for each cell line. Cell-free unconditioned medium was constituted as 90% α-MEM with 10% FBS, 10 ng/mL M-CSF, and 10 ng/mL RANKL. For osteoclast induction assays, spleen cells from 4-week-old C57BL mice (strain: C57BL/6J, Jackson Laboratory) were prepared, using the method described by Granholm and colleagues (29). The cells were seeded in 24-well plate at 1 × 10⁷/well, and cultured in conditioned or unconditioned media. Cells were cultured for 7 days, with media replaced every 2 days. On the final day, cells were fixed in 4% paraformaldehyde and stained for TRAP (tartrate-resistant acid phosphatase) using Acid Phosphatase, Leukocyte (TRAP) Kit (Sigma-Aldrich). TRAP⁺ cells with 3 or more nuclei were counted under brigh-tfield microscopy, and expressed as number of osteoclasts/well.

Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer containing Halt Protease Inhibitor Cocktail (Pierce). Protein concentration was determined with Quick Start Bradford Protein Assay kit (Bio-Rad). The blotting membrane was incubated overnight at 4°C with different primary antibodies: anti-IGF-IR (1:1000; Abcam), anti-RANKL (1:500; Epitomics), and anti-β-actin (1:1000; Sigma). The blots were incubated for 1 hour at room temperature with either horseradish peroxidase-conjugated secondary antibody: anti-mouse or anti-rabbit immunoglobulin (Chemicon). Signals were visualized using enhanced chemiluminescence plus chemiluminescence substrate (Amersham).

The potential miR-335 target site predicted by miRanda on RANKL mRNA 3′UTR was amplified from human genomic DNA using the primer pair(forward: 5′- GTCTGGAGAGGAAATCAGCATCGA-3′; reverse: 5′-TTCAGATGATCCTTCAATTGCGCT-3′)and subcloned into the pMIR-REPORT miRNA reporter vector (Ambion). Recombined plasmids were confirmed by DNA sequencing. Point mutation within the target sequence for miR-335 in 3′UTR (3′UTRm) was generated by the QuickChange II XL site-directed mutagenesis kit (Stratagene) using primers (forward: 5′-TATCCATAAGGTTGACCTTGTAGAGAACACGCGTAT-3′ and 5′-AAGGTCAACCTTATGGATACTGAGTCGTGTACCGT-3′). The plasmid containing 3′ UTRm was sequenced to confirm replacement of the targeted residues. Lentivirus was used to stably transfect miR-335 precursor gene or control sequence into 293TN cells. Transfected cells were further cotransfected via a pMIR reporter vector with RANKL 3′UTR fragment, or corresponding mutation fragment together with the pMIR-REPORT β-galactosidase reporter control vector (Ambion). Cells were collected at 24 hours after transfection. The ratio of β-galactosidase to firefly luciferase was measured with Dual Luciferase Assay kit (Promega).

Immunodeficient mice, NOD/SCID IL2R_γ^null (strain: NOD.Cg-Prkdc^scid Il2r^gtm1Wjl/SzJ; Jackson Laboratory), were housed as 2 to 5 same sex mice in polycarbonate cages (324 cm²) at The Jackson Laboratory under barrier conditions. Mice were maintained under 14:10-hour light: dark cycles; provided sterilized White Pine shavings for bedding, and fed with NIH 31 irradiated diet (6% fat, 19% protein, Ca:P of 1.15:0.85), with vitamin and mineral fortification (Purina Mills International), and sterilized water ad libitum.

SBC-5 miR-335⁺, SBC-5 miR-29a⁺, and SBC-5 Vector^Ctrl were cultured under equivalent conditions. Cells were harvested at about 80% confluence, washed with ice-cold PBS twice, and resuspended in cold PBS at a density of 5 × 10⁶ cells/mL. On day 0, the cell suspension was injected at 1 × 10⁶ cells per mouse via tail vein injection into 8-week-old NOD/SCID IL2R_γ^null male mice (10 mice per group). Mice were euthanized on day 28, and their bodies were fixed in 10% neutralized formalin for at least 2 days. Radiographs of the skeleton were taken with a Faxitron MX20 cabinet X-ray (Faxitron X-Ray Corp.,) and Kodak Min-R 2000 mammography film (Eastman Kodak Co.,). Skeletal osteolytic lesions observed in radiographs were counted in the spine, right and left femurs, and tibias and ulnar bones.

Analyses were conducted with JMP 8.0 software (SAS). We used ANOVA to compare multiple groups, followed by pairwise comparisons if significant differences were detected. Tukey–Kramer test was used for comparisons with a control group. Dunnett test was used to compare all groups. Unpaired t tests were used to compare data when only two groups were used. Differences were considered statistically significant at P < 0.05 on a two-tailed test. Data were expressed as mean ± SEM.

Prior work reports that SBC-5 cells colonize skeletal and nonskeletal tissues, whereas SBC-3 cells only colonize nonskeletal tissues. This fact indicates that SBC-5 cells may produce specific factors to communicate with osteoclasts and/or osteoblasts to colonize into skeleton. Indeed a few factors, such as PTHrP and CCR4, have been reported to be upregulated in SBC-5 cells (30, 31). The molecular mechanisms under upregulation of these factors are not clear. As miRNAs function as negative posttranscriptional gene regulators, we hypothesize that selective and specific miRNAs are expressed reductively in SBC-5 cells, which induce upregulation of the bone-communicating factors. In the primary screen, 14 miRNAs, including miR-9, miR-10a, miR-17–92 family, miR-29 family, and miR-335, were expressed at lower levels in SBC-5 than in SBC-3 (fold change >2, P < 0.001; Table 1; Supplementary Materials and Methods and Supplementary Fig. S1). Because microarray data reflect relative differences in miRNA expression patterns of SBC-5 and SBC-3, we determined whether candidate miRNAs were also present in human normal lung tissues. Compared with normal lung tissue, miR-335 and miR-29a expression was lower in SBC-5 cells, whereas no changes were observed for the other miRNAs. On the basis of these observations, we focused on miR-335 and miR-29a and excluded the other miRNAs (Fig. 1).

Using a lentiviral transfection strategy, miR-335 or miR-29a genes were stably transfected into SBC-5, respectively. The lentivirus system enabled external miRNA gene expression from the constitutive cytomegalovirus promoter and contained copGFP as a reporter. Lentiviral transfection of SBC-5 was successful, exhibiting a robust transfection rate (>90%). To obtain stable cell lines with high expression of the selected miRNA, we screened GFP expression in single colonies in 96-well plates, and verified miRNA expression by qRT-PCR. Transfection of miR-335 gene in SBC-5 (SBC-5 miR-335⁺) increased miR-335 expression by more than 2,600-fold, compared with control lentivirus-transfected SBC-5 (SBC-5 Vector^Ctrl), that expressed very little miR-335 (Fig. 2A; P < 0.001). SBC-5 expressed miR-29a at relatively high levels. Transfection of miR-29a gene in SBC-5 increased miR-29a expression only by approximately 4-fold compared with controls (Fig. 2B; P < 0.001).

In migration and invasion assays in vitro, miR-335 overexpression in SBC-5 miR-335⁺ reduced cell migration by approximately 52% (average cell number per well, SBC-5 miR-335⁺: 687 ± 77 vs. SBC-5 Vector^Ctrl: 1416 ± 107; P < 0.01; Fig. 2C) and cell invasion by approximately 46% (average invasive cell number per well, SBC-5 miR-335⁺: 86 ± 9 vs. SBC-5 Vector^Ctrl: 159 ± 21; P < 0.05; Fig. 2D). In colony formation assay, because of metastatic nature, SBC-5 cells adhere poorly to plastic and proliferating cells and often drift away from their original colonies to form additional colonies, so SBC-5 colony number usually exceeds the number of cells plated after 2-week culture. MiR-335 overexpression in SBC-5 miR-335⁺ reduced cell colony formation by approximately 48% (average colony number per dish, SBC-5 miR-335⁺: 168 ± 10 vs. SBC-5 Vector^Ctrl: 322 ± 11; P < 0.001; Fig. 2E). In contrast, cell migration, cell invasion, and colony formation of SBC-5 miR-29⁺ did not differ significantly from SBC-5 Vector^Ctrl (Fig. 2C–E). Compared with SBC-5 Vector^Ctrl, SBC-5 miR-335⁺ exhibited significantly lower proliferation from culture on day 3, whereas proliferation of SBC-5 miR-29a⁺ did not differ from SBC-5 Vector^Ctrl, except for a slight increase on days 3 and 4 of culture (Fig. 2F). Collectively, these data suggested that miR-335 overexpression reduced the potential for metastatic cancer progression by inhibiting cell migration, invasion, proliferation, and colony formation. In contrast, miR-29a overexpression did not modify key aspects of SBC-5 functional phenotype.

“Metastatic” spread to skeletal and nonskeletal tissues was observed following intravenous tail injections of SBC-5 Vector^Ctrl xenograft. Radiographs showed SBC-5 Vector^Ctrl xenografts induced osteolytic bone lesions in the spine and long bones. In mice with SBC-5 miR-335⁺ xenografts, skeletal osteolytic lesions were absent in 3 mice and reduced overall (average number of lesions/mouse: 1.3 ± 0.36 vs. SBC-5 Vector^Ctrl: 3.3 ± 0.42; P < 0.01; Fig. 3). Four mice with either SBC-5 miR-335⁺ or SBC-5 miR-29a⁺ xenografts, each exhibited 1 osteolytic lesion, whereas the remaining 3 mice with miR-335⁺ xenografts exhibited 3 osteolytic lesions each. Mice with SBC-5 miR-29a⁺ or Vector^Ctrl xenografts exhibited 2 or more (3–5) lesions each, in 6 and 9 mice, respectively (Fig. 3C). The number of osteolytic lesions developed by SBC-5 miR-29a⁺ xenografts did not differ significantly from SBC-5 Vector^Ctrl xenografts (average number of lesions/mouse: 2.3 ± 0.45 vs. SBC-5 Vector^Ctrl: 3.3 ± 0.42; P > 0.05; Fig. 3).

We used informatics prediction software (miRanda) to identify insulin-like growth factor-I receptor (IGF-IR) as a potential target of miR-335. Published literature confirmed that IGF-IR was a direct target of miR-335 (32), and its expression correlated with enhanced proliferation, invasion, and migration ability in tumor cells (33, 34) and bone metastases in breast and prostate cancers (35, 36). Western blotting result showed that SBC-5 expressed much higher IGF-IR than SBC-3 cells. In SBC-5 miR-335⁺, but not in SBC-5 miR-29a⁺, IGF-IR expression was significantly less than that in SBC-5 Vector^Ctrl (Fig. 4).

First, we assessed the effect of conditioned media from genetically modified SBC-5 cell lines on osteoclast induction in vitro. SBC-5 Vector^Ctrl CM, increased osteoclast numbers significantly compared with cell-free unconditioned medium (average osteoclast number per well, SBC-5 Vector^Ctrl CM: 67.2 ± 5.6 vs. cell-free unconditioned medium: 20.7 ± 3.8 in, P < 0.001; Fig. 5A and B). Compared with SBC-5 Vector^Ctrl CM, osteoclast induction was reduced by approximately 57% in SBC-5 miR-335⁺ CM (average 28.7 ± 1.8 osteoclasts per well; P < 0.001) and by 24% in miR-29a⁺ CM (average 49.5 ± 4.8 osteoclasts per well; P < 0.05).

Using miRanda software, we identified RANKL, a key cytokine regulating osteoclastogenesis, as a candidate target for miR-335. We found that RANKL expression was higher in cell lysates and culture medium of SBC-5, than that in SBC-3. We confirmed that RANKL expression was significantly reduced in SBC-5 after overexpression of miR-335 (SBC-5 miR-335⁺), whereas RANKL expression was not different from controls in SBC-5 after overexpression of miR-29a (SBC-5 miR-29a⁺; Fig. 5C). To show the selectivity of RANKL by miR-335, RANKL 3′UTR was cloned and placed within the 3′UTR of a luciferase reporter expression cassette. Cotransfection in 293TN cells with either the miR-335 expression vector or a control vector showed that miR-335 overexpression significantly reduced luciferase activity of the construct containing the target site of RANKL 3′UTR. When the target sequence was mutated, miR-335 failed to reduce luciferase activity (Fig. 5D and E). These data support the hypothesis that the abrogation or reduction of skeletal osteolysis by SBC-5 miR-335⁺ xenografts may be attributed, in part, to inhibition of RANKL expression. Collectively, our data support a role for miR-335, but not miR-29a, in the homing of human SCLC SBC-5 cells to bone.

Very little is known of the cell and molecular mechanisms underlying the metastatic spread to the skeleton in SCLC. Preclinical studies suggest a role for deregulation of bone turnover because mouse xenograft models of SCLC-induced bone metastases are responsive to bisphosphonate drugs, such as zolendronate, which blunt bone resorption and slow bone turnover (37). Despite their efficacy, bisphosphonates in patients with cancer have been linked to complications such as osteonecrosis of the jaw and bone fractures (38), creating a need for alternate therapies. To investigate alternate molecular regulators of the metastatic spread of SCLC cells to bone, we studied miRNA profiles using human SCLC SBC-5 and SBC-3 cell lines and xenografts in immunodeficient NOD/SCID IL2R_γ^null mice, and identified miR-335 as putative candidate regulator of RANKL, a key cytokine regulating resorption, and skeletal osteolysis. Confirming the importance of miR-335 as a candidate regulator in metastatic progression, in vitro work by others showed that miR-335 overexpression in malignant breast cancer cell xenografts suppressed spread of breast cancer cells to lung and bone (39). Taken together with data in the present study, we hypothesize that miR-335 may be an important regulator of bone metastasis. The work reported here is the first to associate miR-335 with human SCLC, and to link miR-335 with bone metastases of SCLC.

Deregulation of miRNAs has been implicated in carcinogenesis, and miRNAs are being investigated as candidate oncology therapeutic targets for several different types of tumors (40, 41). Multiple loci on chromosome 7 have been previously identified in genetic research on human lung cancer (42), so it is interesting that miR-335 is located in chromosome 7q32.2. In various metastatic breast cancer cell lines, gene deletion and epigenetic promoter hypermethylation of the miR-335 locus on 7q32.2 seemed to be a common feature, resulting in the designation of miR-335 as a “selective metastasis suppressor and tumor initiation suppressor locus in human breast cancer” (43). In breast cancer, expression of miR-335 activated the tumor suppressor gene BRCA1 via effects on BRCA1 repressor ID4, resulting in increased apoptosis by downregulation of estrogen receptor-α and IGF-IR (32). In gastric cancer cell lines, miR-335 targeted Bcl-w and specificity protein 1, both of which have also been linked to metastatic progression (44). Our data strongly suggest that increased miR-335 may mitigate human metastatic SCLC, but clinical research is needed to validate this speculation, and to determine whether there are common mechanisms regulated by miRNA across cancer types.

We speculated that miR-335 might target cytokines linked to osteoclast induction and bone turnover. We report two cytokines, IGF-IR and RANKL, both of which have been linked to metastases of other cancers (33, 34), were expressed more highly in SBC-5 cells, and were regulated by miR-335. Deregulated IGF-IR may interact with ligand IGFs to stimulate proliferation, invasion, and migration and inhibit apoptosis of cancer cells (33). Breast cancer research has shown that bone-derived IGF-I regulates interactions between bone and breast cancer cells via activation of the IGF-IR/Akt/NF-κB pathway and that IGF-IR knockdown significantly reduced xenograft-induced bone metastases (35). Extending the literature on breast cancer to SCLC, we now show downregulation of IGF-IR in SBC-5 miR-335⁺ correlated with decreased cell proliferation, migration, invasion, and colony formation.

RANKL regulates osteoclast induction, differentiation, survival, and activation (36, 45) and has been implicated in cancer mechanisms (46). Breast cancer and melanoma cell xenografts migrating to bone may stimulate osteoclast differentiation by RANKL to deregulate bone turnover, resulting in bone resorption and osteolysis (36, 46). Blocking RANKL markedly reduced tumor burden in bones (47). We extended previous reports in breast cancer and lung cancer literature by showing that downregulation of RANKL in SBC-5 miR-335⁺ markedly reduced osteoclast induction, and that miR-335⁺ abrogated or reduced skeletal osteolysis induced by miR-335⁺ xenografts.

Development of NOD/SCID IL2R_γ^null mice, which lack mature T cells, B cells, and functional NK cells, as xenograft hosts greatly improved the predictability of human tumor xenograft phenotyping for cancer research (48). Using these mice, we reported that skeletal osteolysis was induced within one month of placing SCLC SBC-5 xenografts. SBC-5 miR-335⁺ abrogated osteolysis in 3 out of 10 mice, and markedly reduced the sites of osteolysis in remaining mice, whereas SBC-5 miR-29a⁺ had no such effect. These outcomes support our hypothesis that miR-335, but not miR-29a, seems a selective and specific suppressor of bone metastases in this model of SCLC. Our in vitro and in vivo experiments suggest that miR-335 may be a candidate therapeutic target to mitigate bone metastases in SCLC. Clinical research will be necessary to validate this exciting possibility.

No potential conflicts of interest were disclosed.

Conception and design: M. Gong, J.M. Hock, X. Yu

Development of methodology: J. Ma, Y. Yang, X. Yu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Gong, J. Ma, R. Guillemette, M. Zhou

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Gong, Y. Yang, J.M. Hock, X. Yu

Writing, review, and/or revision of the manuscript: M. Gong, R. Guillemette, Y. Yang, J.M. Hock, X. Yu

Study supervision: J.M. Hock, X. Yu

This work was support by grants from the National Natural Science Foundation of China (No. 81072190 to X. Yu, No. 30900244 to M. Gong, No. 81101920 to J. Ma), Science &Technology Department of Sichuan Province (2010SZ0168 to X. Yu), the Ministry of Education of the People's Republic of China [(2011)1139 to X. Yu], Sichuan University (2011SCU04B42 to X. Yu), and U.S. Army Medical Research and Materiel Command research contract USAMRMC (No. 0704400), PI: J.M. Hock.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.