Aerosol gemcitabine inhibits the growth of primary osteosarcoma and osteosarcoma lung metastases



Osteosacarcoma (OS) lung metastases are often resistant to chemotherapy. Most anticancer drugs are administered systemically. In many cases this is followed by dose-dependent toxicity, which may not allow the achievement of therapeutic levels in lungs to eradicate metastases. We determined the efficacy of gemcitabine (GCB) by administering it directly to the lungs via aerosol and studied the role of the Fas pathway in response to the therapy. We used 2 osteosarcoma lung metastases animal models: human LM7 cells that form lung metastases in mice following intravenous injection and murine LM8 cells, which grows subcutaneously in mice and spontaneously metastasize to the lung. Treatment was initiated when the presence of lung metastases had been established. Aerosol GCB inhibited the growth of lung metastases in mice. Intraperitoneal GCB administration at similar dosage had no effect on lung metastases. Besides its direct effect on lung metastases, aerosol GCB suppressed the growth of subcutaneous LM8 tumor. Histopathological examination of mice receiving aerosol GCB showed no evidence of toxicity. Lungs are distinguished from other tissues by the constitutive expression of FasL. Since exposure of tumor cells to GCB upregulated Fas expression, we hypothesized that the susceptibility of the tumor cells to ligand-induced cell death by resident lung cells may be increased. Therefore, the Fas pathway may contribute to the therapeutic effect of aerosol GCB. © 2005 Wiley-Liss, Inc.

The most common site for metastatic spread of OS is the lungs. Patients with OS lung metastases have a poor prognosis with limited therapeutic options. Metastatic and relapsed disease is often resistant to salvage chemotherapy. Our laboratory's goal is to identify new therapeutic approaches for these patients.

GCB has been shown to have limited efficacy in the treatment of advanced sarcomas, including OS,1 but this may be explained in part because systemic administration of GCB does not achieve therapeutic levels in the lungs. We hypothesized that targeted delivery of GCB to the lungs via aerosol will circumvent this problem, yielding a higher drug concentration in the area of the tumor and offer a unique therapeutic opportunity for patients with lung metastases. Indeed, we previously demonstrated that the intranasal administration of GCB at a dose 8-fold lower than the i.v. dose induced the regression of established OS lung metastases.2 However, the intranasal delivery of drugs to the lungs in humans is not ideal as the drugs will not reach the peripheral regions of the lungs where metastases often develop.

Aerosol delivery can bypass this limitation. The clinical feasibility of aerosol drug delivery has been demonstrated with both 9-nitrocamptothecin and GM-CSF.3, 4 In our study, we investigated the efficacy of aerosol GCB in treating OS lung metastases using 2 OS animal model. Aerosol GCB significantly inhibited the growth of primary tumors and of established lung metastases and also prevented metastatic spread, with no evidence of toxicity to normal tissues. By contrast, intraperitoneal administration of a similar dose of GCB inhibited primary tumor growth but failed to prevent metastatic spread to the lungs and affect the growth of lung metastases.

The Fas/FasL death receptor pathway has been identified as a key mediator of chemotherapy-induced apoptosis in leukemia and several solid tumors.5, 6 Our previous investigations have indicated that Fas expression in OS cells inversely correlates with their metastatic potential. Cells with high Fas expression do not form lung metastases following i.v. injection, but those with little or no Fas do.7 We also demonstrated that the upregulation of Fas induced tumor regression in the lungs.8 Chemotherapy-induced alterations in the expression of Fas may therefore contribute to the antitumor activity of the drug. Our data indicate that aerosol GCB therapy resulted in increased cell surface Fas expression, which may contribute to the therapeutic effect of aerosol GCB treatment of lung metastases.


GCB, gemcitabine; GSD, geometric standard deviation; HPLC, high pressure liquid chromatography; MMAD, mass median aerodynamic diameter; OS, osteosarcoma.

Material and Methods

Cell lines and animal models

The human LM7 OS lung metastatic cell line was derived from the SAOS-2 cell line by repeated i.v. recycling through the lungs of nude mice.9 Tumors reached microscopic and visible size at 4 and 6 weeks, respectively, after 106 LM7 cells were injected into the tail vein of nude mice. Murine primary OS Dunn cells and its metastatic subline LM8 were kindly provided by Dr. T. Kashima (University of Tokyo, Tokyo, Japan). Spontaneous microscopic lung metastases become evident 4 weeks after subcutaneous implantation of 5 × 106 LM8 cells into syngeneic immunocompetent C3H mice.10 Both cell lines were maintained in DMEM media supplemented with nonessential amino acids, sodium pyruvate, L-glutamine and 10% fetal bovine serum. Cells were verified to be negative for mycoplasma using the Mycoplasma Plus PCR Primer set (Stratagen, Inc., La Jolla, CA) and for pathogenic murine viruses (M.A. Bioproducts, Walkersville, MD).

Nu/nu mice and C3H mice were purchased from National Cancer Institute (Behtesda, MD) and housed in standard cages with food and water provided ad libitum. All animal experiments were performed with the approval of the Institutional Animal Care and Use Committee.

Reagents and drugs

All media supplements, except serum, were purchased form Whitaker Bioproducts (Walkersville, MD). Serum was purchased from Intergen (Purchase, NJ). Gemcitabine HCl was purchased from Eli Lilly (Indianapolis, IN).

Cytotoxicity assay

Cytotoxicity assays were performed as described previously.11 Briefly, cells were plated on a 96-well plate and treated with GCB at different doses for 24 hr (Dunn and LM8 cells) or 48 hr (SAOS-2 and LM7 cells). Cells that received no treatment were used as the positive control; wells with medium without cells were used as the negative control. Twenty percent (v/v) of 0.42 mg/ml 3-(4,5-dimethylthiazol-2-yl)2,5-diphenyltetrazolium bromide was added to each well, and the cells were incubated for 2–4 hr. Medium was then aspirated, and cells were lysed with 0.1 ml of dimethylsulfoxide. Cytotoxicity was quantified by using a 96-well microtiter plate reader at 570 nm.

Aerosol GCB characterization

The particle size of aerosol droplets containing GCB was measured with an Anderson nonviable ambient particle sizing sampler (Andersen Instruments, Atlanta, GA). The concentration of GCB (1 mg of GCB per 1 ml of saline) in the aerosol generated by the AeroTech II jet nebulizer (CIS-USA, Bedford, MA) was also measured. Samples were collected over a 3 min period operation. The mass median aerodynamic diameter (MMAD) and geometric standard deviation (GSD) were calculated using KaleidaGraph 2.0 software (Synergy Software, Reading, PA).

To determine the concentration of GCB in aerosol, the GCB aliquots were collected from Andersen sampler and quantified by HPLC analysis using a Waters 710B WISP automatic injector and Waters Nova-Pak C18 column (3.9 × 150 mm; Waters, Milford, MA) at room temperature. The mobile phase was composed of 50 mM ammonium acetate (pH 5.0) and acetonitrile (96.5:3.5 v/v) at a flow rate of 1 ml/min.12 GCB was detected using a Waters 440 absorbance detector with monitoring at 275 nm. The data were analyzed with Waters Millenium software.

Aerosol GCB treatment and dosage

Treatment of mice with aerosol was performed as described previously.3 Briefly, an AeroTech II nebulizer was used to generate aerosol particles at the air flow rate of 10 l/min. Mice were placed unrestrained in a sealed plastic cage and exposed to the aerosol. The aerosol particles were generated with 5% CO2-enriched air obtained by mixing normal air and CO2 with a blender (Bird3M, Palm Springs, CA). The CO2 concentrations were calibrated with a Fluid Fyrite (Bacharach, Inc., Pittsburgh, PA). The use of CO2 increased pulmonary deposition of drugs approximately 3-fold.13 The estimated total deposited amount (D) of inhaled GCB was calculated by the following formula: D = C × V × DI × CF × T, where C is the concentration of drug in aerosol volume (for 1.0 mg/ml initial concentration of drug in the nebulizer C = 17.7 mg/L), V is the volume of air inspired by the animal during 1 min (for mice, V = 1 l/min per kg14), DI is the estimated deposition index (for mice the fraction of inhaled dose deposited throughout the respiratory tract was 0.315), CF is the CO2 factor, which increases pulmonary deposition of drug approximately 3-fold13, and T is the duration of treatment (min). Therefore, during the 30 min inhalation of GCB aerosol generated from 1 mg of GCB/ml stock, the total deposited dose of the drug in mice was ∼0.5 mg/kg.

Therapy studies

Nude mice were injected with human LM7 cells into the tail veins. Aerosol treatment was initiated 4 weeks after tumor cell inoculation when micrometastases were established in lungs. Mice received aerosol GCB treatment twice weekly for 60 min at a total deposited dose of 1.0 mg/kg (GCB concentration in nebulization solution was 1.0 mg/ml), or for 30 min for a total deposited dose of 0.5 mg/kg (GCB concentration in nebulization solution was 1.0 mg/ml) or 0.1 mg/kg (GCB concentration in nebulization solution was 0.1 mg/ml with the duration of treatment 60 min) for 6.5 weeks. Control mice did not receive treatment or received aerosolized saline using the same schedule as described for GCB. At the end of treatment mice were sacrificed and their lungs were weighed and fixed in formalin.

LM8 cells were subcutaneously injected into C3H mice. Therapy with either aerosol or i.p. GCB (0.5 mg/kg of GCB 3 times weekly for 3.5 weeks). The size of the primary s.c. tumors was measured 2 or 3 times/week. All animals were sacrificed when the s.c. tumor volume in the control group reached 1.5 mm3.

Toxicity studies

Naïve, tumor-free immunocompetent mice were treated with 0.5 mg/kg of GCB by inhalation 3 times weekly for 4 weeks or 2.5 mg/kg twice weekly for 5 weeks. During treatment, the animals were weighed 2 or 3 times/week. Twenty-four hours after the last treatment, animals were sacrificed and organs were resected for histopathology analysis. Five micrometer sections of lungs, livers, kidneys, hearts, brains, spleens and sternums (for bone marrow evaluation) were stained with hematoxylin and eosin and analyzed by a veterinary pathologist.


Tissues were formalin-fixed and paraffin embedded. Tissue sections were deparafinized in xylene and rehydrated. To block exogenous peroxidases tissues were blocked with 3% H2O2 for 12 min. Nonspecific binding was then blocked by PBS containing 10% normal horse serum and 1% normal goat serum. The primary antibody, polyclonal rabbit anti-Fas antibody (Santa Cruz Biotechnology, Santa Cruz, CA) diluted to 4 mg/ml was applied and left overnight at 4°C. The secondary antibody labeled with horse-radish peroxidase was then applied for 2 hr at ambient temperature. The slides were finally developed with 3,3′-diaminobenzidine as a substrate and lightly counterstained with hematoxylin. Negative controls were prepared by omitting the primary antibodies.

Flow cytometric analysis

For Fas staining of LM7 cells, 1 × 106 cells were suspended in FACS buffer (PBS, containing 2% fetal calf serum and 0.1% NaN3) and incubated with either 1.0 mg/ml PE-conjugated mouse anti-human Fas antibody (clone DX2) or isotype-matched, PE-conjugated control mouse anti-human IgG1 antibody (Pharmingen, San Diego, CA). For Fas staining of LM8 cells, 0.5 × 106 cells were suspended in FACS buffer and incubated with either 1 mg/ml PE-conjugated hamster anti-mouse Fas monoclonal antibody or isotype-matched, PE-conjugated control hamster IgG antibody (Pharmingen, San Diego, CA). Samples were analyzed with a FACScan (Becton Dickinson, Mountain View, CA).

Statistical analysis

To determine the significance of differences in the number of metastatic nodules between groups we used the Mann-Whitney rank sum test. In studies of subcutaneous tumors, the tumor volume was measured and the significance of differences was estimated using the unpaired, 2-tailed Student's t-test. The difference was considered significant if the p value was less than 0.05.


Aerosol characteristics

Aerosol particle size was determined for GCB using Anderson Cascade Impactor and particle size distribution (Fig. 1). The MMAD and GSD values of aerosolized GCB solution were 0.8 μm and 2.1, respectively. The aerosol particles are well suited for drug delivery throughout the respiratory zone of the lungs.14 Quantitative HPLC analysis indicated that the GCB concentration in aerosol produced by AeroTechII nebulizer under the conditions described in Material and Methods was 17.7 μg/l.

Figure 1.

Aerosol particle size distribution. Gemcitabine solution was aerosolized by AeroTech II nebulizer (CIS-USA, Bredford, MA) at 10 l/min air-flow rate. Particle size was determined using Anderson Cascade Impactor. Most aerosol particles generated at this condition were <3 mm.

Effect of aerosol GCB on human OS LM7 lung metastases

The primary human SAOS-2 cell line and its metastatic subline LM7 demonstrated sensitivity to GCB in vitro (Table I). Aerosol GCB given for 6.5 weeks starting 4 weeks after tumor inoculation significantly inhibited the growth of LM7 lung metastases (Table II, Fig. 2). The wet-weight of lungs in control group and groups, receiving 0.1, 0.5 and 1.0 mg/kg doses of GCB were 370 ± 262 g, 208 ± 52 g, 188 ± 17 g and 200 ± 18 g, respectively. There was no significant difference between control group and mice receiving 0.1 mg/kg GCB. Mice receiving higher doses of GCB, 0.5 and 1.0 mg/kg, had normal lung weights, which were significantly different from control group (p = 0.04). The best therapeutic response was observed in mice, who received 1.0 mg/kg aerosol GCB twice weekly. In this group no visible lung metastases were evident at the end of therapy (10.5 weeks following tumor cell injection). In mice receiving 0.5 mg/kg aerosol GCB the incidence of visible lung metastases was not different from that of the control group (p > 0.05). However, the tumor area of lung metastases was significantly smaller than in the control group (Table II, p = 0.03). Aerosol GCB at 0.1 mg/kg was not effective in reducing either visible tumor incidence or tumor size (p > 0.05 compared to control group). Similarly, lung micrometastases were inhibited only in the group of mice treated with 1.0 mg/kg GCB (Table II, p = 0.01).

Figure 2.

Effect of aerosol GCB treatment on human LM7 OS lung metastases. Nude mice with established pulmonary metastases received GCB aerosol treatment twice weekly for 6.5 weeks. Representative lungs are shown from an untreated group of mice (top row) and mice treated with 0.5 mg/kg GCB (middle row) or 1.0 mg/kg GCB (bottom row).

Table I. Cytotoxic Effect of GCB on OS Cells
Cell lineGCB ID50 (μM)1
Table II. Effect of Aerosol GCB Treatment on Established LM7 Lung Metastases1
TreatmentVisible lung metastasesLung micrometastases incidence4 (%)
Incidence2 (%)Mean tumor diameter ±SD (mm)Mean number of metastases [range]Tumor area3 (mm2)
  • 1

    Aerosol GCB treatment was initiated 4 weeks after LM7 tumor cell injection. The number of mice in each treated group was 12; 10 mice were use in the control group. The treatment was given twice weekly for 6.5 weeks.

  • 2

    Number of mice with visible lung metastases/total number of mice;

  • 3

    Tumor area represents as the sum of the individual tumor areas on the lung of each mouse calculated as πd2/4, where d is diameter of each tumor in mm;

  • 4

    Number of mice positive for pulmonary micrometastases/total number of mice;

  • *

    p >0.05 vs. control;

  • **

    p = 0.03 vs. control, p > 0.05 vs. 0.1 mg/kg and 1.0 mg/kg group;

  • ***

    p = 0.01 vs. control.

Control803.2 ± 2.846 [0–200]101±109100
GCB 0.1 mg/kg75*2.5 ± 1.77 [0–14]24.6±27.6*100
GCB 0.5 mg/kg50*1.2 ± 1.69 [0–16]7.3±16.0**75*
GCB 1.0 mg/kg0000***37***

Effect of GCB on LM8 primary tumors and lung metastases

LM7 cells do not form tumors when injected s.c. or into the bone. Therefore, the mouse LM8 animal model was used to determine whether aerosol GCB was effective against primary tumors as well as lung metastases. In vitro studies showed that both parental Dunn cells and its metastatic subline LM8 cells were sensitive to GCB (Table I). The efficacy of aerosol vs. i.p. GCB was compared using the 0.5 mg/kg dose of GCB given 3 times weekly for 3.5 weeks. Because LM8 tumor growth is rapid, we increased the frequency of treatment in these studies. Treatment was initiated when the primary tumor volume reached 130 mm3. The number of micrometastases in the lung was significantly reduced only in mice receiving aerosol GCB compared to the control group (p = 0.02, Experiment 1, Table III). GCB administered i.p. was not effective (p > 0.05 vs. control). In this experiment we also determined the effect of GCB therapy on the primary tumor. Aerosol GCB was as effective as i.p. GCB in inhibiting tumor growth (Fig. 3). The difference in tumor size between untreated mice and mice receiving aerosol GCB treatment was first noticed on day 12 after treatment initiation (p=0.01) and continued until the end of the experiment.

Figure 3.

Effect of GCB administered by aerosol or intraperitoneally on LM8 subcutaneous tumor. C3H mice with established s.c. LM8 tumors (130 mm3) received GCB 3 times weekly by aerosol (triangles) or intraperitoneally (squares) at 0.5 mg/kg per treatment for 3.5 weeks or no treatment (circles). *p<0.05 for mice receiving aerosol GCB vs. no treatment.

Table III. Effect of Aerosol VS. Intraperitoneal GCB on LM8 OS Lung Metastases1
TreatmentIncidence of lung micrometastases (%)
Experiment 1Experiment 2
  • 1

    Aerosol GCB treatment was initiated when LM8 s.c. tumor was 130 mm3 (Experiment 1) or 20 mm3 (Experiment 2). The treatment was given at the dose of 0.5 mg GCB/kg 3 times weekly for 3.5 weeks.

  • *

    p = 0.02 vs. control;

  • **

    p=0.85 vs. control;

  • ***

    p = 0.002 vs. control.

GCB aerosol29*0***
GCB i.p.83**n/a

Initiation of aerosol GCB when the primary LM8 tumor size was smaller (20 mm3) completely inhibited lung metastasis formation (Experiment 2,Table III).

Toxicity of aerosol GCB

Toxicity studies were performed in immunocompetent naïve mice. Aerosol GCB was well tolerated. We found no difference in animal weights (p > 0.05, data not shown). Histologic examination of lungs, bone marrow, livers, kidneys, spleens and hearts from mice treated with aerosol GCB at 0.5 mg/kg 3 times weekly for 4 weeks demonstrated no abnormalities. In addition, there were no signs of inflammation or fibrosis in the lungs. Mice receiving 5-fold higher dose (2.5 mg/kg) of GCB by inhalation twice weekly for 5 weeks also did not have any signs of toxicity.

Effect of GCB on Fas expression

Fas expression was examined in LM7 and LM8 pulmonary metastases after they had been treated with aerosol GCB. Fas expression was negligible in untreated metastases. By contrast, Fas expression significantly increased after aerosol GCB treatment (Fig. 4). Upregulation of cell surface Fas expression was also observed in vitro when LM7 and LM8 cells were treated with GCB and analyzed by flow cytometry (Fig. 5).

Figure 4.

Fas expression in LM7 and LM8 OS lung metastases in mice following aerosol GCB treatment. Lungs were resected from mice at the end of treatment, fixed in 10% formalin buffer and paraffin-embedded. Sections were prepared and stained as described in Material and Methods. Anti-Fas immunoreactivity is shown as the brown-stained areas. Counterstaining was done by hematoxylin. In negative controls, the primary antibody treatment was omitted. Original magnification ×100.

Figure 5.

Fas expression in metastatic OS cell lines following GCB treatment in vitro. LM7 cells were treated with 0.01 mM of GCB for 36 hr, LM8 cells were treated with 1 mM GCB for 24 hr. Fas expression was measured by flow cytometry using PE-labeled anti-Fas monoclonal antibodies (dotted line for untreated samples and solid bold line for GCB treated samples) as described in Material and Methods. PE-conjugated IgG isotype control antibody histograms are indicated as solid thin lines. The percentage of Fas-positive cells increased from 49% to 65% in LM7 cells and from 43% to 95% in LM8 cells following GCB treatment.


The development of aerosol technology for administration of chemotherapy to the lungs may provide a novel approach for treatment of OS and other malignancies that metastasize to the lung. We have previously reported that aerosol delivery of chemical agents and genes was an effective method for targeting these agents to the pulmonary region.3, 16, 17, 18 This method of delivery significantly altered the agent's biodistribution and pharmacokinetics in favor of pulmonary deposition, which included the peripheral areas of the lung, an area where OS metastases are frequently found.19, 20 GCB is a nucleoside analogue, which has a different mechanism of anticancer activity than standard drugs used for OS treatment. However, GCB showed limited efficacy in treatment of sarcomas, including OS, at tolerable systemic doses.1, 21, 22 We assume that these systemic doses were not sufficient to achieve therapeutic levels of GCB in lungs to affect lung metastases growth. The data from our current study indicate that this drug may have therapeutic potential for OS. Both human and mouse OS cells were sensitive to GCB in vitro. We have also previously demonstrated that intranasal instillation of GCB at a dosage lower than the typical intravenous dosage was more effective at inhibiting OS lung metastases than systemic GCB administration.2 Therefore, topical administration of GCB to the lung allows attaining therapeutic doses, which affect lung metastases growth.

In our study, we evaluated the efficacy of aerosol GCB. The response of LM7 and LM8 lung metastases to aerosol GCB was dose dependent. Furthermore, using the LM8 model, we demonstrated that aerosol GCB was also effective against the primary tumor. The initiation of aerosol GCB (0.5 mg/kg 3 times/week) when the primary tumor was small (20 mm3) effectively inhibited the formation of lung metastases. At a more advanced stage of the disease, when the primary tumor volume reached 130 mm3 before initiation of therapy, aerosol GCB significantly suppressed the growth of lung metastases. By contrast, GCB given by i.p. injection at the same dosage and at the same schedule had no effect on lung metastases. We assume that aerosol administration changed the pharmacology of GCB in favor of pulmonary deposition when compared to intraperitoneal administration. In fact, in our previous studies with camptothecin, we found that biodistribution and pharmacokinetics of aerosolized drug were different than those for systemic administration.19 During aerosol treatment all inhaled drug is processed through the lungs before it gets into systemic circulation, whereas the drug administered systemically is first diluted in the blood before it reaches the lung. This dilution factor during systemic administration may decrease the pulmonary concentration of the drug. Also drug metabolism in the blood stream may be different from its metabolism in the lung environment, which may affect activity and the clearance pattern of the agent from the lungs. We are planning to perform pharmacological studies with aerosol GCB in our laboratory to investigate this. Aerosol GCB was well tolerated with no significant toxicity as determined by histopathologic examination.

We previously demonstrated that Fas expression correlates inversely with the metastatic potential of OS cells and that upregulation of Fas can induce tumor regression in the lung.7, 8 Upon binding to its ligand Fas will triggers the cell death. Since Fas ligand is constitutively expressed by the lung epithelium, tumor cells with low levels of Fas expression or tumor cells that have no Fas may evade this host defense mechanism. Agents that upregulate Fas may therefore assist and promote this natural mechanism. GCB, both in vitro and in vivo, resulted in increased expression of cell surface Fas in our study. These data suggest that in addition to its cytotoxic activity mediated by inhibiting DNA synthesis and cell proliferation in S phase, aerosol GCB may induce tumor regression of OS lung metastases by a mechanism involving the Fas/FasL pathway.

Despite aggressive chemotherapy and surgical resection of the primary tumor, 30–40% of OS patients will relapse with pulmonary metastasis.23 Incorporation of other effective therapeutic approaches into the initial adjuvant chemotherapy regimens may decrease the number of relapses. In our study, we found that aerosol GCB in addition to its effect on lung metastases, inhibited primary tumor growth. Thus, aerosol GCB may be useful in the treatment of both primary OS and the microscopic lung metastases that are usually present at the time of diagnosis.


We thank Dr. S.-F Jia for providing the human osteosarcoma animal model for these studies and Dr. C. Van Pelt for her assistance in the toxicology studies. This work was supported by Golfers against Cancer and Legends of Friendswood Awards to NVK.