Tumor Suppressor 101F6 and Ascorbate Synergistically and Selectively Inhibit Non–Small Cell Lung Cancer Growth by Caspase-Independent Apoptosis and Autophagy

101F6 is one of the candidate tumor suppressor genes identified in the human chromosome 3p21.3 region where allele loss and genomic alterations have been frequently found in lung cancer and many other cancers (1–3). In lung and breast cancers, the frequent and early loss of heterozygosity and the overlapping homozygous deletions observed in the 3p21.3 region suggest that one or more genes in this region may play a critical role in the molecular pathogenesis of these cancers (1–4). The cloned cDNA of 101F6 (GenBank accession no. AF040704) is 1117 bp long and encodes a protein consisting of 223 amino acid residues. We previously showed that the forced expression of wild-type (wt) 101F6 by adenoviral vector–mediated gene transfer significantly inhibited tumor cell growth in 3p21.3-deficient non–small cell lung cancer (NSCLC) cells in vitro and in vivo (1). However, the mechanism of 101F6-mediated tumor suppression remains unknown.

A computer-aided homologous structure analysis of the 101F6 protein predicted it to be a member of the di-heme cytochrome b561 (Cyt b₅₆₁) protein family, as shown by a more than 90% alignment of its consensus amino acid sequence within the transmembrane domain of the Cyt b₅₆₁ (5, 6). The Cyt b₅₆₁ family of proteins constitutes a class of intrinsic high-redox-potential membrane proteins containing two heme molecules that are involved in the regeneration and homeostasis of ascorbate (ascorbic acid, vitamin C) and has been shown to play an important role in a wide variety of physiologic processes, including iron uptake, cell defense, nitrate reduction, and signal transduction (5–8). Cyt b₅₆₁ has been identified in a large number of phylogenetically distant species, and most species contain three or four Cyt b₅₆₁ paralogous proteins. Cyt b₅₆₁ family proteins share well-conserved structural features characterized by multiple transmembrane helices, four histidine residues that may coordinate two heme molecules, and putative ascorbate and mono-dehydro-ascorbate substrate–binding sites. However, the protein sequence similarities are very low, both within a single and among the phylogenetically distant species (5–7). Three Cyt b₅₆₁ proteins encoded by the human genome have been identified thus far: Cyt b₅₆₁ (ProteinSeq ID: NP_001017916; gene locus: at chromosome 17q11) locates inside chromaffin granules and may be involved in ascorbate regeneration. Cyt b_561-2 (NP_079119; at 2q31.1) presents in the plasma membrane of duodenum cells and may function as a ferric reductase. Cyt b_561-3 (NP_872386; at 1p13.3) may be a “ubiquitous” Cyt b₅₆₁, and its physiologic function is unknown (6, 7). The finding of 101F6 (NP_008953; at 3p21.3) as a Cyt b₅₆₁ domain-containing protein may add a new member to the human Cyt b₅₆₁ protein family and predict a potential biological function. The homology with Cyt b₅₆₁ triggers us to study the interaction of 101F6 with ascorbate in cancer pathogenesis and tumor suppression.

Cyt b₅₆₁ family proteins probably evolved from a common ancestral protein and may ubiquitously use ascorbate as a primary antioxidant and are involved in ascorbate homeostasis and antioxidative defense (7). Ascorbate, a well-known antioxidant, has been used as a supplemental therapeutic agent for human cancer prevention and therapy, although controversy surrounds regarding its effectiveness in treating some cancers and in improving patients' well-being (6, 9–16). Recent evidence, however, has shown that ascorbate at high concentrations (>1000 μmol/L) is selectively toxic to cancer cells but not normal cells in vitro, and preclinical and early clinical studies have also shown that high-dose ascorbate, given by i.v. and p.o. routes, may inhibit tumor growth in animal models and improve symptoms and prolong survival in patients with advanced cancer (6, 9, 10, 13, 17, 18). These studies suggest that the role of high-dose ascorbate in cancer treatment should be reevaluated. On the basis of the structural association of 101F6 with the Cyt b₅₆₁ protein family and its potent tumor-suppressing activity shown in vitro and in vivo, we hypothesized that 101F6 may function as a novel class of Cyt b₅₆₁ protein. Under normal physiologic conditions, 101F6 may be activated in response to environmental oxidative stress to protect cells from oxidative damage by regenerating ascorbate and maintaining its homeostasis, a function that would be aberrant in 101F6-deficient tumor cells. By contrast, restoration of 101F6 activity in 101F6-deficient tumor cells should facilitate ascorbate-mediated cytotoxic H₂O₂ formation and suppress tumor cell growth by inducing apoptosis.

To test these hypotheses, we examined endogenous expression of 101F6 protein in normal human bronchial epithelial cells, lung fibroblasts, and NSCLC cell lines and tissue samples to assess its role in lung cancer pathogenesis. Next, we investigated the combined effects of 101F6 and ascorbate on normal lung cells and 3p21.3-deficient NSCLC cells. Finally, we evaluated the therapeutic efficacy of a combination of systemic injection of 101F6 nanoparticles and i.p. injection of ascorbate at a subpharmacologic dose in a human NSCLC H322 orthotopic lung tumor mouse xenograft model to explore the translational applications of this novel treatment strategy. We found that 101F6 nanoparticles and ascorbate had a synergistic and selective antitumor effect through unique caspase-independent apoptotic and autophagic pathways in vitro and in vivo, setting the stage for clinical application of this novel approach.

Cell lines and cell culture. Normal human bronchial epithelial (HBE) cells, WI-38 lung fibroblasts and NSCLC cell lines A549, H1299, and H322 with varied 3p21.3 and p53 expression status were used for both in vitro and in vivo experiments. The A549 cell line was maintained in F12 nutrient mixture supplemented with 10% FCS. The H1299 and H322 cell lines were maintained in RPMI 1640 supplemented with 10% FCS and 5% glutamine. WI-38 cells were cultured in minimal essential medium supplemented with 10% FCS and 5% glutamine. Normal HBE cells were cultured in keratinocyte serum-free medium containing epidermal growth factor and bovine pituitary extract.

Reagents and antibodies. Ascorbate (Cenolate) was purchased from the pharmacy at the University of Texas M. D. Anderson Cancer Center. DC reagent (20 mmol/L) in D5W was produced in our laboratory, and preparation of DC:plasmid-DNA nanoparticles were described previously (19, 20). Mouse, rabbit, and goat polyclonal antibodies were purchased from various commercial sources: anti-101F6 and FITC-conjugated anti-101F6 (Bethyl Laboratories); anti-Myc and anti–β-actin (A1978; Sigma); anti–caspase-2 and antiapoptosis-inducing factor (Santa Cruz Biotechnology); anti–caspase-3, anti–caspase-6, anti–caspase-7, and anti–caspase-9 (Cell Signaling Technology), and anti–caspase-8 (BD Biosciences). Anti–microtubule-associated protein 1 light chain 3 antibody was kindly provided by Dr. Seiji Kondo.

Immunofluorescence and immunocytochemical analysis. Immunofluorescence staining was done in cells cultured in chamber slides. Briefly, 48 h after transfection, the cells were fixed in 10% formalin and then incubated with 5.0 μg/mL FITC-conjugated rabbit anti-101F6 antibodies in 5% PBS-bovine serum albumin buffer for 1 h at room temperature. Subsequently, the stained samples were mounted with 4′,6-diamidino-2-phenylindole (DAPI)–containing Vectashield solution (Vector Laboratories Inc.) to counterstain the nuclei. The slides were then examined under a fluorescence microscope equipped with ImagePro image analysis software (Media Cybermetrics) to determine the subcellular localization of 101F6 protein. Expression of the 101F6 protein in cell lines was analyzed by immunocytochemical staining with anti-101F6 rabbit polyclonal antibodies and the Vectastain Elite ABC kit (Vector Laboratories Inc.). The 10% formalin-fixed cell lines and the paraffin-embedded tissue sections were incubated with anti-101F6 antibodies (2.5 μg/mL in PBS), and immunostaining was done with the Vectastain kit according to the manufacturer's instructions. Subsequently, the samples were counterstained with Harries hematoxylin and examined under a microscope equipped with a digital camera.

Tumor cell-induced clonogenicity assay. To analyze the effect of 101F6 protein expression on tumor cell–induced clonogenicity in vitro, we transfected H1299 cells (2 × 10⁵ cells) on six-well plates with 101F6- or Myc-101F6–expressing vector or with empty vector using the DC reagent. H1299 cells were transfected with 2 μg of each test plasmid DNA together with 1 μg of neomycin resistance gene containing the pcDNA3.1 vector (Invitrogen). The vectors (1 μg) alone and with the wt p53 plasmid were used as negative and positive controls, respectively. After 24 h, cells were harvested, stained with trypan blue, and counted. Cells (1 × 10⁴) were replated onto a 100-mm dish in triplicate and grown in 10% FCS and RPMI 1640 containing 400 μg/mL of G418 for 2 weeks. The numbers of G418-resistant colonies were counted after staining with crystal violet.

Cell viability assay. Cells counts were done to determine IC₅₀ and IC₂₀ values of ascorbate in NSCLC cell lines, WI-38 cells, and HBE cells. For this, 2 × 10⁵ cells were plated and precultured for 24 h in six-well plates and then incubated in the presence of 10 different concentrations of ascorbate for 72 h. The IC₅₀ and IC₂₀ values of ascorbate were calculated using a curve-fitting system as described previously (21). Inhibition of tumor cell growth by 101F6 nanoparticles and ascorbate was also analyzed by cell counts. Cells were seeded as above and 24 h later were transfected with 3 μg of 101F6 or LacZ nanoparticles in serum-free medium for 2 h. Cell media were replenished with or without various concentrations of ascorbate. Cells were harvested 24, 48, and 72 h after treatment, and cell viability was determined by trypan blue exclusion assay.

Analysis of apoptosis induction in vitro. Induction of cell death in tumor cells transfected with the 101F6 nanoparticles with or without ascorbate at an IC₂₀ level was analyzed by fluorescence-activated cell sorting (FACS) using the Apo-BrdU kit (BD Biosciences PharMingen) and the FITC-labeled annexin V apoptosis detection kit (BD Biosciences PharMingen). Briefly, cells were plated in 60-mm dishes at a density of 4 × 10⁵ cells per dish and transfected the following day with 6 μg of LacZ vector or 101F6-expressing vector with or without ascorbate at an IC₂₀ level. After 72 h, cells were harvested and fixed in 1% paraformaldehyde. After incorporation of BrdUTP, cells were visualized using FITC-labeled anti-bromodeoxyuridine antibody. The amount of total cellular DNA was determined by staining cells with propidium iodide (PI)/RNase buffer. Cells were then processed for FACS analysis to determine the degree of apoptosis and cell cycle kinetics, as described previously (20). For the annexin V analysis, cells were collected 24 h after transfection and then resuspended at a density of 1 × 10⁶ cells/mL in 1× binding buffer [10 mmol/L HEPES buffer (pH 7.4), 150 mmol/L NaCl, 5 mmol/L KCl, 1 mmol/L MgCl₂, and 1.8 mmol/L CaCl₂] and stained simultaneously with FITC-labeled annexin V (25 ng/mL; green fluorescence) and PI (50 ng/mL). PI was provided as a 50 μg/mL stock solution (BD Biosciences PharMingen) and was used as a cell viability marker. Cells were analyzed using FACS, and the data were analyzed with CellQuest software (BD Biosciences).

Measurement of intracellular uptake of ascorbate. Cells (1 × 10⁷) were treated and then incubated with ascorbate for 2 h, washed thrice with PBS, and then collected. Cell pellets were resuspended in 0.5% formic acid and centrifuged at 500 × g for 5 min. The supernatant was analyzed with the Agilent 1200 Series high-performance liquid chromatography (HPLC) coupled with electrospray ionization tandem mass spectrometry (LC/MS/MS; refs. 22, 23). The HPLC analytic column used was a PrimesPhere (5-μm C18 HC 250 × 2.00 mm; Phenomenex). The precolumn was a 0.5-μm replacement frit from Upchurch Scientific Inc. (Oak Harbor, WA). The mobile phase was a gradient over 10 min with 0.2% formic acid in water and methanol mixtures. The flow rate was 300 μL/min, and the column was maintained at 25°C. Ascorbate's retention time was 2.08 min. The concentrations of the standard solution of ascorbate were in the 5.0–1,000.0 ng/mL range; the ascorbate was dissolved in a mixture of methanol and 0.2% formic acid (50:50).

Determination of intracellular H₂O₂ production. Intracellular H₂O₂ production was measured using H₂DCFDA (Molecular Probes), a cell-permeable H₂O₂-sensitive fluorescent dye indicator. The cell-permeable H₂DCFDA is not fluorescent, and once it is in the cytosol, the esterase activity renders the indicator impermeable by forming the fluorescent product dichlorofluorescein. The fluorescence intensity of the dye is proportional to the rate of oxidation by reactive oxygen species (predominantly H₂O₂) generated by agents with a high redox potential, such as ascorbate. Briefly, cells (3 × 10⁵) were treated with various experimental agents and then incubated with or without ascorbate for 72 h. H₂DCFDA (0.5 μmol/L) was added to the cell cultures during the last 30 min of incubation in the dark. Cells were washed in PBS, resuspended in PBS buffer, and analyzed by FACS.

Caspase inhibition assay. To determine whether caspase is involved in the effects of exogenous 101F6 and ascorbate, a caspase inhibition assay was done with a pan-caspase inhibitor Z-VAD-fmk (R&D Systems, Inc.) dissolved in DMSO according to the manufacturer's instructions. Ad-CMV-p53 was used as a positive control. Briefly, H1299 cells (3 × 10⁶ cells per well) were seeded in a 60-mm dish and incubated overnight. Next, 50 or 100 μmol/L Z-VAD-fmk was added 2 h before treatment with 101F6 nanoparticles and ascorbate or recombinant adenoviral infection with Ad-p53 at a multiplicity of infection (MOI) of 50. After 24 h, cell viability was determined by trypan blue staining, as described above. The viability of untreated H1299 cells in the presence of diluted DMSO was regarded as 100%.

Characterization and quantification of autophagy. To quantify acidic vesicular organelles (AVO) in cells treated with 101F6 nanoparticles and ascorbate, we did vital staining with an acridine orange fluorescent dye according to the protocol described elsewhere (24–26). Briefly, treated tumor cells were stained with acridine orange, which was added at a final concentration of 1 μg/mL for 15 min. To quantify the AVOs that developed, cells were detached by trypsinization and collected for FACS analysis.

The green fluorescent protein (GFP)–tagged LC3-expressing cells were used to show the induction of autophagy (24–26). GFP-LC3 cells showed a diffuse distribution under control conditions, but showed a pundit pattern (GFP-LC3 dots) and increased in both number and fluorescence intensity in the presence of autophagy. Therefore, we determined the involvement of LC3 in tumor cells treated with exogenous 101F6 and ascorbate using the GFP-LC3 expression vector. H1299 cells were transfected with the GFP-LC3 expression vector using the DC reagent. After an overnight culture, cells were treated according to the protocol cited above, and then 24 h after treatment, cells were fixed with 4% paraformaldehyde and examined under a Nikon TC200 fluorescence microscope equipped with a digital camera and imaging analysis software.

An autophagy-specific inhibitor 3-methyladenine (3-MA) was used to verify whether the 101F6 and ascorbate-induced autophagy in these NSCLC cells was a death-promoting or protective process. H1299 and H322 cells were treated with 101F6 nanoparticle and ascorbate at an IC₂₀ level alone or in combination in the presence and absence of 5 mmol/L of 3-MA. Forty-eight hours after treatment, cell viability was determined as described in Cell Viability Assay.

Animal studies. All animals were maintained, and animal experiments were done in accordance with NIH and institutional guidelines established for the Animal Core Facility at the University of Texas M. D. Anderson Cancer Center. The animals used in this study were female nu/nu mice (4–6 weeks old) that were purchased from Charles River Laboratories. Before tumor cell inoculation, mice were subjected to 3.5 Gy of total body irradiation from a cesium-137 radiation source. To evaluate the therapeutic efficacy of the systemic injection of 101F6 nanoparticles and i.p. injection of ascorbate in a human H322 orthotopic lung cancer mouse model, mice were inoculated with H322 cells (2 × 10⁶ cells per mouse) in 100 μL of PBS by intrathoracic injection using a 27-gauge needle. The tumor cell–inoculated mice were randomly divided into six treatment groups (six mice per group): group A was treated with PBS; group B with ascorbate; group C with LacZ nanoparticles; group D with LacZ nanoparticles and ascorbate; group E with 101F6 nanoparticles; and group F with 101F6 nanoparticles and ascorbate. On days 6, 9, and 12 after tumor cell inoculation, we administered LacZ or 101F6 nanopartcles i.v. via the tail veins at a concentration of 25 μg of plasmid and 10 nmol of DC each in 100 μL of D5W per mouse. In groups B, D, and F, ascorbate was injected i.p. thrice, each at a dose of 250 mg/kg of body weight, together with the nanoparticles in groups D and F. On day 28 after tumor cell inoculation, all mice were killed, and the total weight of the pleural tumors in each mouse was measured.

In addition, to estimate 101F6 protein expression and apoptosis in tumors, on day 26, two mice from each group were repeatedly treated with the respective agents for that group (101F6 nanoparticles were replaced by Myc-tagged 101F6 nanoparticles for immunohistochemical analysis). After 48 h (on day 28), these mice were killed, and pleural tumors larger than 5 mm in diameter were randomly harvested and freshly frozen. 101F6 protein expression was then analyzed by anti-Myc antibody (1:100) using the Vectastain Elite ABC kit and examined under a microscope. Induction of apoptosis was analyzed using an in situ cell death detection kit with fluorescein (Roche Biochemicals) according to the manufacturer's instructions. Tissues were then examined under a fluorescence microscope and analyzed using the equipped software.

Statistical analysis. All in vitro experiments were done at least thrice with duplicate or triplicate samples. ANOVA and Fisher's tests were used to compare the values for the test and control samples, and P < 0.05 was considered statistically significant. StatView 5.0 (Abacus Concepts, Inc.), and SAS software were used for all of the statistical analyses. To analyze the interaction between 101F6 and ascorbate, we assumed that treatment with ascorbate would multiply the number of cells by a certain amount (p1) and that treatment with 101F6 would multiply the number of cells by another amount (p2). If there were no interaction, the combination of ascorbate and 101F6 should multiply the cell count by p1 × p2. For example, if ascorbate multiplies the count by 0.80 and 101F6 multiplies it by 0.70, then, in the absence of an interaction, the combined treatments should multiply the count by 0.8 × 0.7 = 0.56. Because we assumed a multiplicative effect, the log of the cell count was taken as the dependent variable to allow the use of a linear model. To account for the fact that the data were paired (i.e., if there is a different baseline for each experiment), the experiment was modeled as a random effect. Thus, we fit a linear model with the log of the cell count as the dependent variable, the experiment as a random effect, and the actions of ascorbate, 101F6, and the two combined as fixed effects. The estimation was made with the restricted maximum likelihood method using JMP software from SAS. The significance of the interaction was evaluated by a paired t test.

Expression and tumor-suppressing activity of 101F6 protein. To determine the expression and subcellular localization of exogenous 101F6, we did immunocytochemical (Fig. 1A) and immunofluorescence (Fig. 1B) staining assays in 101F6-negative human NSCLC H1299 cells transiently transfected with the 101F6-expressing plasmid vectors complexed with N-[1-(2,3-dioleoyloxyl)propyl]-N,N,N-trimethylammoniummethyl sulfate:cholesterol (DC) nanoparticles (101F6 nanoparticles). Expression of 101F6 showed a typical subcellular localization on the cytoplasm and organelle or vesicular (e.g., endoplasmic reticulum) membrane. No endogenous 101F6 protein expression was detected in H1299 cells. In addition, no 101F6 was detected in H1299 cells transfected with the empty vector. On the other hand, 101F6 was clearly detected in the cytoplasm of H1299 cells transfected with 101F6-expressing plasmid vector. We also checked endogenous 101F6 protein expression in 101F6-negative human NSCLC A549, H322, and H460 cells; HBE and lung fibroblast WI-38 cells. Endogenous 101F6 protein expressions could not be detected in the A549, H322, and H460 cells but was strongly expressed in the cytoplasm of normal HBE and WI-38 cells. We also analyzed endogenous 101F6 expression in a panel of 19 lung cancer cell lines and in normal HBE and WI-38 cells by Western blot, using ectopically expressed 101F6 protein in H1299 cell as a positive control (Fig. 1C). A low and moderate level of 101F6 protein expression was detected in normal HBE and WI-38 cells, respectively, consisting with the above immunofluorescence imaging analysis. No expression of 101F6 protein was detected in most (∼90%) of these NSCLC cells, and a low level of expression was detected in two cell lines H69 and H128, suggesting that the inactivation of 101F6 may be predepository for tumorigenesis.

Clonogenic assays in soft agar confirmed the tumor-suppressing activity of exogenously expressed 101F6 in vitro (Fig. 1D and E). Tumor cell-induced clonogenicity was significantly inhibited in 101F6- and Myc-101F6–expressing H1299 cells compared with vector control cells (P < 0.0001), and this tumor-suppressing function was significant but not as great as exogenously expressed p53 that was used as a positive control. These results show that the restoration of 101F6 protein function in 3p21.3-deficient NSCLC cells efficiently suppresses tumor cell-induced clonogenicity.

Synergistic and selective inhibition of 3p21.3-deficient NSCLC cell growth by 101F6 nanoparticles and ascorbate in vitro. To determine the role of 101F6 as a potential member of the Cyt b₅₆₁ protein family in ascorbate-associated biological activities, we first analyzed the effect of forced expression of 101F6 protein on cell growth in various normal lung and NSCLC cells by 101F6 nanoparticle–mediated gene transfer in the presence and absence of ascorbate at varied doses. We determined the IC₅₀ value for ascorbic acid in two NSCLC cell lines, H1299 and H322, and two normal cell lines, HBE and lung fibroblast WI-38. There was about a 250- to 500-fold difference in the IC₅₀ value between normal and cancer cell lines (IC₅₀ values of ascorbate for NSCLC H322 and H1299 were 0.54 and 1.17 mmol/L and for normal HBE and WI-38 cells were 252 and 150.2 mmol/L, respectively). To evaluate the combined effects of exogenous 101F6 and ascorbate on cell growth, we determined the viability of cells transfected with 101F6 nanoparticles and treated with ascorbate at a subpharmacologic concentration (IC₂₀; Fig. 2A). Untransfected or LacZ nanoparticle–transfected cells were used as negative controls. A significant reduction (70–80%) in cell viability was seen in H1299 and H322 cells treated with 101F6 nanoparticles and ascorbate at an IC₂₀ level compared with those in cells treated with ascorbate at an IC₂₀ level in combination with DC (40%) or LacZ nanoparticles (50%), or in cells treated with 101F6 nanoparticles alone (30–40% reduction). Furthermore, statistical analysis of the interaction between two therapeutic agents indicated that combination treatment with 101F6 nanoparticles and an IC₂₀ dose of ascorbate were synergistic in tumor cell growth inhibition (P < 0.0001). By comparison, transfection with the 101F6 nanoparticles alone or in combination with an IC₂₀ dose of ascorbate had no effect on the growth of normal HBE or WI-38 cells. These results suggest that the inhibitory effects mediated by the agents alone or in combination are highly tumor selective.

Enhanced induction of apoptosis by 101F6 nanoparticles and ascorbate. We next examined the ability of exogenously expressed 101F6 protein and an IC₂₀ dose of ascorbate to induce apoptosis in H1299 and H322 cells by FACS analysis using annexin V and terminal nucleotidyl transferase–mediated nick end labeling (TUNEL) assays (Fig. 2B and C). The earlier apoptotic events were quantified by FACS with an annexin V assay in H1299 cells 24 h after treatment with 101F6 nanoparticles and ascorbate. In cells treated with either 101F6 nanoparticles or ascorbate at an IC₂₀ level alone, 12% and 14% of cells, respectively, were positive for annexin V staining, compared with the basal levels of staining in control cells treated with PBS (10%) or DC nanoparticles alone (12%). In cells treated with DC and LacZ nanoparticles and ascorbate at an IC₂₀ level, 25% and 25% of cells, respectively, were positive for annexin V staining, whereas 83% of cells treated with 101F6 nanoparticles and ascorbate stained positively for annexin V. Similar results were obtained for H322 cells (data not shown). The induction of apoptosis was determined by FACS with a FITC-dUTP–labeled TUNEL reaction and PI staining in H1299 and H322 cells. In cells treated with 101F6 nanoparticles or ascorbate alone, no apoptosis was induced. Of the cells treated with LacZ nanoparticles and ascorbate at an IC₂₀ level, only 12% of H1299 and 3% of H322 cells showed apoptosis (Fig. 2C). By contrast, the apoptotic populations in cells treated with 101F6 nanoparticles and ascorbate at an IC₂₀ level were dramatically increased with ∼10-fold higher induction than in untreated cells (PBS) or cells treated with LacZ nanoparticles and ascorbate. Treatment with 101F6 nanoparticles and ascorbate dramatically enhanced the induction of apoptosis and was synergistic in both cell lines (P < 0.0001) compared with either agent alone or combination controls. These results suggest that the observed synergistic inhibition of NSCLC cell growth may be mediated through the induction of apoptosis promoted by the mutual activities of exogenous 101F6 expression and ascorbate treatment.

Enhanced intracellular uptake of ascorbate by exogenous 101F6 expression. To elucidate the mechanism of the induction of tumor cell killing by 101F6 and ascorbate, we examined the effect of exogenous expression of 101F6 protein on ascorbate-mediated cytotoxic H₂O₂ formation and accumulation in NSCLC cells. We first measured the intracellular uptake of supplemental ascorbate in H1299 and H322 cells by a quantitative HPLC-LC/MS/MS analysis 2 h after transfection with 101F6 nanoparticles (refs. 22, 23; Fig. 3A and B). Exogenous expression of 101F6 protein significantly enhanced the intracellular uptake of extracellularly supplied ascorbate, as shown by a more than 50% greater intracellular ascorbate concentration in both H1299 (P < 0.0011) and H322 (P < 0.0003) cells treated with 101F6 nanoparticles and ascorbate than in cells treated with DC reagent and ascorbate or with LacZ nanoparticles and ascorbate. Interestingly, exogenous expression of 101F6 also seemed to increase intracellular accumulation of intrinsic ascorbate, suggesting that 101F6 may function as a typical Cyt b₅₆₁ in facilitating ascorbate regeneration and maintenance.

Enhanced intracellular H₂O₂ accumulation by 101F6 and ascorbate. Next, we analyzed the effect of exogenous expression of 101F6 protein on ascorbate-mediated H₂O₂ formation and accumulation in NSCLC cells using a fluorescent dichlorodihydrofluorescein diacetate (H₂DCFDA) assay and quantified by flow cytometry (Fig. 3C and D). The oxidation of H₂DCFDA in H1299 cells treated with 100 μmol/L H₂O₂ for 1 h served as a positive assay control. No increase in the level of intracellular H₂O₂ was seen in either H1299 cells treated with ascorbate at an IC₂₀ level or with LacZ- or 101F6-expressing vector alone. In contrast, a significant increase in the H₂O₂ level was observed in response to 101F6 nanoparticle and ascorbate at an IC₂₀ level (P < 0.0033). Similar results were also obtained in H322 cells. A significant increase in the intracellular accumulation of H₂O₂ was detected in both cell lines only in response to exogenous 101F6 and ascorbate. Taken together, our results showed that exogenous 101F6 promoted the intracellular uptake of ascorbate, which subsequently facilitated the formation and accumulation of cytotoxic H₂O₂ mediated by bioactivity of the ascorbate inside tumor cells and, thus, accelerated tumor cell killing.

Caspase-independent cell death mediated by activities of 101F6 and ascorbate. To identify which cell death pathway is responsible for mediating the effect of exogenous 101F6 and ascorbate, we did Western blotting to analyze the activity of caspases, a family of proteases that are involved in the signal transduction of apoptotic stimuli (Fig. 4A). For this purpose, H1299 cells were transfected with 101F6 nanoparticles alone or with an IC₂₀ level of ascorbate. Untreated (PBS), DC, and LacZ vector-treated cells served as negative controls. Cells treated with cisplatin, a well-known apoptosis-inducing and caspase-activating anticancer drug, were used as a positive control, and the predicted cleavage bands (p17, p19) of procaspase-3 were clearly detected on the blot. In contrast, no activated caspase-3 was found in the H1299 cells treated with 101F6 nanoparticles and ascorbate alone or in combination, as shown by the lack of cleavage products of precaspase-3 protein on the Western blot. No activation of caspase-2, caspase-6, caspase-7, or caspase-9 was detected in response to this treatment (data not shown). We also did a caspase activity assay with and without caspase-specific inhibitors to further verify the involvement of caspases in 101F6 and ascorbate-mediated apoptosis induction (Fig. 4B). As shown in Fig. 5B, the proportions of viable H1299 cells transduced with a recombinant adenoviral vector Ad-p53 at a MOI of 50 significantly increased from 17% to 24% (P < 0.012) in the presence of 50 μmol/L Z-VAD-fmk, a pan-caspase inhibitor, and from 17% to 39% (P < 0.001) in the presence of 100 μmol/L Z-VAD-fmk. On the other hand, addition of 50 or 100 μmol/L Z-VAD-fmk did not significantly affect the viability of H1299 cells treated with exogenous 101F6 and ascorbate. Because these high concentrations of Z-VAD-fmk inhibitor efficiently suppress caspase activities in most cellular systems, these results confirmed that the cell death induced by 101F6 and ascorbate is independent of caspase activation.

101F6 and ascorbate-mediated tumor cell killing by induction of autophagy. To identify the molecular pathway leading to this caspase-independent cell death, we analyzed whether autophagy, which is often seen in caspase-independent cell death, might be induced. Autophagy is the process of sequestering cytoplasmic proteins in the lytic compartment and is characterized by the development of AVOs (27). We first quantified AVO development by FACS using vital staining with an acridine orange fluorescence dye (Fig. 5A). The percentage of AVO-positive cells showing prominent red fluorescence dramatically increased 4–5-fold from 8% in cells treated with ascorbate at an IC₂₀ level that was determined in tumor cells or 101F6 nanoparticles alone to 38% in cells treated with both. These results indicate that nanoparticle-mediated wt 101F6 gene transfer and an IC₂₀ dose of ascorbate induced the development of AVOs in H1299 cells, suggesting that autophagy had occurred.

To further characterize the autophagy induced by 101F6 and ascorbate, we analyzed the expression status of the microtubule-associated protein 1 light chain 3 (LC3) proteins in H1299 cells. LC3 is a mammalian homologue of yeast Apg8p and specifically associated with the autophagosome membrane (25, 27–29). We used a novel GFP-tagged LC3 expression vector (GFP-LC3; refs. 25, 27–29) to specifically detect and quantify the autophagy induced by exogenous 101F6 and ascorbate in tumor cells by fluorescence imaging analysis (Fig. 5B and C). When autophagy is induced, the extensive formation of GFP-LC3–labeled structures (GFP-LC3 dots), which are pre-autophagosomes and autophagosomes, are detected. Formation of GFP-LC3 dots were detected in H1299 cells treated by 101F6 nanoparticles or ascorbate alone or in combination, but no GFP-LC3 dot formation was observed in untreated (PBS) or LacZ vector-treated controls. The percentage of autophagic cells was determined by counting and averaging the number of cells containing GFP-LC3 dots in at least five view fields for each treatment under a fluorescence microscope (Fig. 5C). The populations of GFP-LC3–positive cells were significantly greater in cells treated with 101F6 nanoparticles and ascorbate than in cells treated with either agent alone. These results collectively suggest that treatment with exogenous 101F6 and an IC₂₀ dose of ascorbate specifically and efficiently induces autophagy.

There are mainly two types of autophagic effects observed in cancer therapy. One is the apoptosis and cell death–inducing autophagy and the other is a protective one (27). To verify whether the 101F6- and ascorbate-induced autophagy in these NSCLC cells was a cell death–promoting or protective process, we analyzed the effect of the autophagy-specific inhibitor 3-MA on 101F6- and ascorbate-induced cell death. 3-MA was shown to act specifically upon the autophagic/lysosomal pathway of endogenous protein degradation (30). In the presence of 3-MA, an increased cell viability was observed in both H1299 and H322 cells treated with 101F6 nanoparticles and ascorbate at an IC₂₀ level alone or in combination compared with that in the absence of 3-MA (Fig. 5D), indicating that the 101F6- and ascorbate-induced autophagy mainly attribute to cell death. The autophagic cell death possessed about 30–40% of the total tumor cell killing induced by either agent alone or in combination (Fig. 5D), suggesting an important role of autophagy in 101F6- and ascorbate-mediated tumor cell growth inhibition.

Inhibition of tumor formation and growth by systemic treatment with 101F6 nanoparticles and ascorbate in orthotopic lung cancer mouse models. We next evaluated the therapeutic efficacy of systemic (i.v.) administration of 101F6 nanoparticles and i.p. injections of ascorbate in a human H322 orthotopic lung cancer mouse model (Fig. 6A and B). The effect of a combination of 101F6 nanoparticles and ascorbate was compared with that of PBS, PBS and ascorbate, LacZ nanoparticles, LacZ nanoparticles and ascorbate, and 101F6 nanoparticles alone. Each treatment group consisted of six mice, and the in vivo experiment was done twice. After 7 days of tumor-cell inoculation, animals were given i.v. injections of nanoparticles every 2 days thrice at a dose of 25 μg of plasmid DNA and 10 nmol of DC in 100 μL of 5% dextrose in water (D5W) per mouse. For the combination treatment groups, ascorbate was injected i.p. at 250 mg/kg of body weight together with nanoparticles. Fourteen days after the last treatment, all mice were killed, and the total weight of the pleural tumors and the body weight of each mouse were recorded. There were fewer and smaller tumors in mice treated with PBS and ascorbate, LacZ and ascorbate, and 101F6 alone or with ascorbate than in those treated with either PBS or LacZ alone. In particular, mice treated with 101F6 and ascorbate had no large tumors and had only a few small tumors in the thoracic cavity. The mice treated with 101F6 and ascorbate also had almost no pleural micrometastases (smaller than 0.5 mm). The total tumor weight in mice treated with 101F6 and ascorbate was significantly less in any other treatment group (P < 0.005). A statistical analysis of changes in total tumor weights among treatment groups using a mathematical model developed previously (21) showed that 101F6 nanoparticles and ascorbate were synergistic (P < 0.0002) in this orthotopic model.

Furthermore, we analyzed the expression of 101F6 protein by immunohistochemical staining (Fig. 6C) and the induction of apoptosis with an in situ cell death detection kit with a TUNEL reaction (Fig. 6D and E) in tumors from each treatment group. Animals were killed 48 h after treatment, and freshly frozen tumor tissue samples were prepared for these assays. Exogenous expression of 101F6 protein was detected in tumor cells from the mice treated with 101F6 alone or with ascorbate but not in those from other groups. Furthermore, a marked induction of apoptosis was detected in tumors cells from animals treated with ascorbate, 101F6, or both. The intensity of staining of apoptotic cells was much stronger for animals treated with 101F6 and ascorbate than for animals treated with ascorbate or 101F6 alone. Few apoptosis-positive cells were observed in tumors from animals treated with PBS or LacZ. Apoptosis induction was significantly greater in tumor cells in mice treated with 101F6 nanoparticles and ascorbate than in mice treated with either agent alone (P < 0.0014; Fig. 6E). These results indicated that the 101F6 nanoparticles were delivered to the tumor sites by the systemic administration and expressed and functioned successfully inside the tumor cells. The in vivo mediation of tumor suppression by 101F6 and ascorbate is consistent with that observed in vitro.

We found that exogenous expression of 101F6 enhanced intracellular uptake of supplemental ascorbate and thus facilitated the ascorbate-mediated formation and accumulation of cytotoxic H₂O₂ in tumor cells. We also found that the activation of 101F6 protein expression stimulated intracellular accumulation of intrinsic ascorbate. These results suggest that 101F6 protein may function as a novel class of Cyt b₅₆₁ protein in ascorbate regeneration and homeostasis. More extensive biochemical studies are needed to characterize the potential role of 101F6 protein as a transmembrane electron transfer protein. In addition, we also observed that both the exogenous and endogenous expression of 101F6 proteins in cell cultures were up-regulated by the addition of ascorbate, although the exact mechanism was not clear (data not shown). These mutually facilitated biological activities of 101F6 protein and ascorbate may partially contribute to their observed synergism in selective inhibition of NSCLC cell growth and the induction of apoptosis.

We have shown that tumor cell killing induced by 101F6 and ascorbate is mediated through unique caspase-independent apoptotic and autophagic pathways. Autophagy, a process characterized by dynamic changes in the morphology of subcellular membranes that lead to the degradation of cellular proteins and cytoplasmic organelles, has recently been found to play an important role in human cancer development and response to chemotherapy and radiation therapy (27). It has also been shown that tumor suppressors such as Beclin 1, DAP kinase, and PTEN may mediate autophagy, and that deficiencies in or suppression of autophagy might be associated with malignant transformation (31). Autophagy has also been shown to be involved in multiple physiologic processes in multicellular organisms, including protein degradation and organelle turnover, and caspase-independent cell death has been suggested to be a hallmark of autophagy (27, 31). Tumor cells may undergo both apoptosis (programmed cell death type I) and autophagy (programmed cell death type II) in response to some anticancer drugs, and the two processes may occur separately (32, 33) or simultaneously (34). In this study, we found strong evidence that 101F6 and ascorbate induced autophagy, as shown by a significant induction of AVO development and GFP-LC3 dot formation on the autophagosome membrane. In addition, we found that apoptosis and autophagy induced by 101F6 and ascorbate did not occur synchronously, as indicated by the evidence that autophagy was maximally induced 24 h after treatment, whereas apoptosis induction reached a maximum of 72 h after treatment. This asynchronous occurrence of apoptosis and autophagy may contribute to the synergistic antitumor effect of 101F6 and ascorbate by efficiently activating multiple cell death pathways at different times to achieve maximal killing of tumor cells. Such caspase-independent cell death pathways are also important in protecting the organism against unwanted and potentially harmful cells when caspase-mediated routes fail (35, 36), suggesting that this 101F6-induced cell death may be a defense mechanism initiated in response to oxidative stress to remove damaged cells and prevent tumorigenesis.

We also found that the cytotoxicity induced by the overexpression of 101F6 protein and a high dose of ascorbate was highly selective for tumor cells and had no effect on normal cell growth under the same treatment conditions. This tumor-selective cytotoxicity has been observed in ascorbate-mediated cancer treatment at a high pharmacologic dose (>1,000 μmol/L) in vitro and in preclinical and clinical trials in various human cancers (6, 9, 10, 13, 17, 18, 37, 38). It has been speculated that ascorbate-mediated cell death is due to protein-dependent H₂O₂ generation via ascorbate radical formation, with ascorbate as the electron donor (10). Our evidence of the 101F6-enhanced intracellular uptake of ascorbate and accumulation of H₂O₂ suggests that the 101F6 protein functions as the right electron transfer protein in mediating this process. The resulting H₂O₂ might target membrane lipids and form hydroperoxides or reactive intermediates that are quenched or reduced in normal cells but not in cancer cells and, in addition, the intracellular H₂O₂ could target DNA and DNA repair proteins more efficiently in the tumor cells because of the generally diminished superoxide dismutase activity in cancer cells (8, 10, 37, 39).

Ascorbate has a controversial history as a cancer treatment, and recent findings indicate that the route of ascorbate administration may produce huge differences in plasma concentrations (6, 9, 10, 13, 17, 18, 37, 38). In particular, recent pharmacokinetic studies in humans showed that 10 g of ascorbate given i.v. was estimated to produce a plasma concentration of nearly 6 mmol/L, which was more than 25-fold higher than the concentrations achieved by the same oral dose (9, 17, 40). Other findings have shown as much as a 70-fold difference in the plasma concentration between p.o. and i.v. administration, depending on the dose (10). It was for this reason that we administered ascorbate i.p. rather than p.o. in our preclinical study. Total doses of 250 and 500 mg/kg ascorbate were administered i.p., and the 500 mg/kg dose was regarded as the maximum tolerated dose in mice. The tumor-selective and synergistic therapeutic efficacy of 101F6 nanoparticles and ascorbate at a subpharmacologic dose may offer a safe and efficient anticancer treatment strategy and improve cancer patient care.

In summary, we found that nanoparticle-mediated wt 101F6 gene transfer and a subpharmacologic concentration of ascorbate synergistically and selectively inhibited NSCLC cell growth by caspase-independent apoptosis and autophagy in vitro and in vivo. These findings provide new insight into the molecular mechanism of tumor suppression by 101F6 and ascorbate and suggest that combining 101F6 and ascorbate could be effective in overcoming tumor resistance to therapeutic agents that depend on caspase activation and may underline the potential utility of synergistic induction of autophagy as a new cancer treatment modality. The synergistic therapeutic efficacy shown by the systemic administration of 101F6 nanoparticles and ascorbate at a subpharmacologic dose may offer an attractive strategy and rationale for future clinical trials in the reevaluation of ascorbate-mediated cancer prevention and treatment.

Grant support: National Cancer Institute, the NIH (Specialized Programs of Research Excellence CA70970, CA71618, MMHCC U01CA10535201, and RO1CA116322); grants from the Department of Defense TARGET (DAMD17-02-1-0706) Lung Cancer Programs; a W.M. Keck Gene Therapy Career Development grant; the M. D. Anderson Cancer Center Support Core Grant (CA16672); and a grant from the Tobacco Settlement Funds as appropriated by the Texas State Legislature.

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.

We thank Drs. Kamensky and Palmer at Rice University, Houston, Texas, for anti-bovine Cyt b₅₆₁ antibodies. We thank David Galloway for manuscript editing.