Down-regulation of Phosphoglucose Isomerase/Autocrine Motility Factor Expression Sensitizes Human Fibrosarcoma Cells to Oxidative Stress Leading to Cellular Senescence*

Phosphoglucose isomerase/autocrine motility factor (PGI/AMF) is a housekeeping gene product present in all cells, is an essential enzyme of catabolic glycolysis and anabolic gluconeogenesis, and regulates tumor cell growth and metastasis. Because glycolytic enzyme up-regulation of expression contributes to glycolytic flux, leading to increased of cell growth and a resistance to cellular stress of normal fibroblasts whereas down-regulation of PGI/AMF leads to mesenchymal-to-epithelial transition in tumor cells, we examined the involvement of PGI/AMF in overcoming cellular senescence in cancer cells. PGI/AMF cellular expression in HT1080 human fibrosarcoma was down-regulated by small interfering RNA methodology, which resulted in an increased sensitivity to oxidative stress and oxidative stress-induced cellular senescence. Signaling analysis revealed that the senescence pathway involving p21 cyclin-dependent kinase inhibitor was up-regulated in PGI/AMF knockdown cells and that superoxide dismutase is the upstream regulator protein of p21-mediated cellular senescence. A specific inhibitor of PGI/AMF induced cellular senescence and p21 expression in tumor cells exposed to an oxidative stress environment. Taken together, the results presented here suggest that PGI/AMF is involved in oxidative stress-induced cellular senescence and should bring novel insights into the control of cellular growth leading to a new methodology for cancer treatment.

cancer suppressor phenomenon regulated in part by tumor suppressor genes. Because cancer cells have overcome the limits to proliferation imposed by the cellular senescence machinery, it was speculated to be an attractive target for cancer therapy (6,7). Phosphoglucose isomerase (PGI; EC 5.3.1.9) is a housekeeping cytosolic enzyme of sugar metabolism present in all cells and plays a key role in both glycolysis and gluconeogenesis pathways. It catalyzes the reversible aldose-ketose isomerization of D-glucose-6-phosphate to D-fructose-6-phosphate and in the recycling of hexose-6-phosphate in the pentose phosphate pathway (8). PGI is a secretable protein that extracellularly behaves as a cytokine following binding to its seven-transmembrane receptor (gp78/AMFR; Ref. 9). Molecular cloning and sequencing have identified PGI as an autocrine motility factor (AMF), which is originally identified as a major cell motility-stimulating factor associated with cancer progression and metastasis (8,10,11). PGI/AMF was independently identified as neuroleukin that promotes growth of embryonic spinal and sensory neurons (12), maturation factor mediating differentiation of human myeloid leukemia cells (13), sperm antigen-36 (14), and myofibril-bound serine proteinase inhibitor (15). Aberrations in expression or activities of PGI caused by mutations or deletions are of significant clinical importance because in humans they are associated with hereditary nonspherocytic hemolytic anemia diseases (16,17). In addition, PGI/AMF is an antigen of arthritis disease (18), and its presence in the serum and urine is of prognostic value associated with cancer progression (19 -21), and the levels of PGI/AMF and its cell surface receptor gp78/AMFR expressions are associated with the pathologic stage, grade, and degree of tumor penetration to surrounding tissues marking a poor prognosis (19 -23).
Recent data demonstrate that overexpression of PGI/AMF expression in murine fibroblasts leads to an acquisition tumorigenic property (24), whereas its down-regulation triggers premature senescence in mouse embryonic fibroblasts (25). In cancer cells the suppression of PGI/AMF expression leads to mesenchymal-to-epithelial transition and is manifested in the inhibition of cell proliferation and tumorigenicity (26), whereas the up-regulation of PGI/AMF expression is thought to contribute to resistance to various cellular stresses (24,27). All of the above implies that PGI/AMF is involved in cellular senescence; however, its signaling pathway remains elusive. Thus, we investigated the role of PGI/AMF in cellular in responses to oxidative stress-induced senescence.
Cell Lines and Culture-Human fibrosarcoma HT1080 cells (ATCC CCL-121) were purchased from the American Type Culture Collection (Manassas, VA) and were cultured in Dulbecco's modified Eagle's medium with 10% heat-inactivated fetal bovine serum and penicillin/ streptomycin. Human breast cell lines MCF10A, MCF10AT1, and DCIS.com cells were obtained from Karmanos Cancer Institute (originally called Michigan Cancer Foundation). MCF10A and MCF10AT1 cells were maintained in Dulbecco's modified Eagle's medium with Ham's F-12 medium supplemented with 0.1 g/ml cholera toxin, 0.02 g/ml epidermal growth factor, 10 g/ml insulin, 0.5 g/ml hydrocortisone, 100 units/ml penicillin, 100 g/ml streptomycin, and 5% horse serum. DCIS.com cells were maintained in Dulbecco's modified Eagle's medium with Ham's F-12 medium supplemented with 5% horse serum. The cultures were maintained at 37°C in an air/5% CO 2 incubator.
Western Analysis-Western analysis was performed as described previously (26). Equal amounts of the proteins were separated on SDS-PAGE gels and transferred to a 0.2-m polyvinylidene fluoride membrane (Osmonics Inc., Minnetonka, MN) at 15 V, 30 mA overnight at 4°C. Then the membrane was blocked with 0.1% casein solution in 0.2ϫ phosphate-buffered saline for 1 h at room temperature. The blocked membrane was incubated with primary antibody, and then secondary antibodies were conjugated with fluorophore (IRDye 800 antibodies (Rockland Immunochemical, Gilbertsville, PA) or Alexa-Fluor 680 antibodies (Molecular Probes, Eugene, OR)). The blots were visualized by using the LI-COR Bioscience Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln, NE).
Senescence-associated ␤-Galactosidase Assay-The SA ␤-Gal activity was monitored using a senescence detection kit (Calbiochem), and cells were observed under an Olympus fluorescence microscope.
MTT Assay-MTT assay was used to determine cell proliferation and viability. Briefly, 1 ϫ 10 3 cells/well were plated in 96-well plates and cultured for different time periods indicated. At the end of the assay time period, 10 l of MTT (5 mg/ml) was added to each well and then incubated at 37°C for 4 h. After removing the medium completely, solubilizing solution (acidic isopropanol) was added at 100 l/well. The plates were read at 570 nm on a spectrophotometric plate reader with a reference wavelength at 650 nm.
Detection of ROS Production-Cells seeded on glass culture slides (Nalge Nunc International Co., Naperville, IL) were washed with phosphate-buffered saline and loaded with 1 M CM-H 2 DCFDA in phosphate-buffered saline for determining the intracellular ROS level, for 15 min at 37°C in the dark. Dichlorofluorescein-dependent fluorescence was analyzed in an Olympus fluorescence microscope using a X400 lens.
Statistical Analysis-The data are expressed as the means Ϯ S.D. or S.E. Comparisons between the groups was determined using unpaired t test. p Ͻ 0.05 was considered statistically significant.

PGI/AMF Knockdown Induces
Cellular Senescence-To study the possible role of PGI/AMF in cellular senescence, we partially knocked down PGI/AMF in HT1080 fibrosarcoma cells by gene silencing using siRNA duplexes derived from the human PGI/AMF sequence ( Fig.  1A; complete silencing of PGI/AMF protein expression results in loss of viability). Transfection of anti-PGI/ AMF siRNA markedly reduced the protein level of PGI/AMF in HT1080 cells (Ͼ50%), whereas the control siRNA did not affect the PGI/AMF protein level of expression (Fig. 1A).
Cellular senescence is a cell-triggered program in response to a variety of stresses including oxidative stress such as exposure to H 2 O 2 , which induces premature cellular senescence (2)(3)(4)28). To assess the knockdown effect of PGI/AMF on senescence, the cell transfectants were exposed to H 2 O 2 , and the senescence biomarker, SA ␤-gal activity, was measured. The SA ␤-gal activity was increased in a dose-dependent manner in response to treatment with H 2 O 2 , and HT1080 cells transfected with anti-PGI/AMF siRNA exhibited a higher rate of SA ␤-gal activity as compared with the control cells (Fig. 1B). siPGI/AMF cells showed about a 2-fold higher SA ␤-gal-positive cell formation than the control. Time course experiments revealed that SA ␤-gal-positive cell formation was detected in HT1080 cells within 4 days after exposure to 150 M H 2 O 2 and was highly increased in siPGI/AMF cells at all of the time points (Fig. 1C). Fig. 1D shows an example of SA ␤-gal assay of 150 M H 2 O 2induced senescent cells at day 8. Furthermore, the cellular response to H 2 O 2 as measured by cell proliferation revealed that such a treatment lead to a dose-dependent retardation of cell growth, and the siPGI/AMF cells showed significantly lower growth rate and greater sensitivity to H 2 O 2 than the controls (Fig. 2). Similar results were obtained with another clone of PGI/AMF knockdown cells (siPGI/AMF-2; supplemental Figs. S1 and S2).
Next, we examined whether the above results were limited to HT1080 cells or are a general phenomenon. Thus, we analyzed the breast cancer progression cell model that provides a unique tool for the investigation of molecular changes associated with progression of human breast cancer as it depicts early and late stages of human breast cancer, i.e. MCF10A cell (nontumori- genic), MCF10AT1 cell (tumorigenic), and DCIS.com cell (tumorigenic; locally invasive) (29 -31). Initially, we have detected a significantly lower expression of secreted PGI/AMF from MCF10A cells than from MCF10AT1 and/or DCIS.com cells (Fig. 3A) and found that the MCF10A cells were ϳ2.5-fold more sensitive to oxidative stress that the counterpart cells (Fig. 3B). Next, using the PGI/AMF-siRNA technique, we found that both endogenous and exogenous expressions were suppressed in DCIS.com cells (Fig. 3C) associated with a 2-fold increase in SA ␤-gal-positive cells formation than the control, not only under oxidative stress condition but also under normal conditions (Fig.  3D). Furthermore, we checked the effect of PGI/AMF knockdown for cellular senescence in MCF10A cells (Fig. 3, E and F). In PGI/AMF knockdown MCF10A cells, a similar pattern was observed for oxidative stress-induced cellular senescence, but the differences were not statistically significant (Fig. 3F). These results suggest that PGI/AMF involvement in the cellular senescence process is not cell type-specific, and exogenous PGI/AMF is required for regulating cellular senescence.
The effect of PGI/AMF knockdown on cell cycle proteins expression under oxidative stress. Cellular senescence is the state of stable cell cycle arrest induced by oxidative stress, DNA damage, and oncogene activation and is thought to be a tumor suppressor mechanism (2)(3)(4). To explore the molecular mechanisms underlying cellular senescence in PGI/AMF knockdown tumor cells, we studied the expression of cell cycle regulatory proteins and antioxidant proteins. As compared with the HT1080 control cells, p53 expression was significantly elevated in siPGI/AMF cells followed by induction of a target of p53, e.g. p21 (Fig. 4A).
It is now well accepted that the enhanced activities of antioxidative enzymes, such as SOD, play an essential role in preventing the development of cancer (32)(33)(34). Thus, we examined their level in response to PGI/AMF down-regulation in HT1080 cells. The levels of both of CuZnSOD and MnSOD enzyme were increased about ϳ2-fold in these cells as compared with the control cells (Fig. 4B). Consistent with the increased level of SODs, siPGI/AMF tumor cells exhibited a marked decrease of ROS generation (Fig. 4C), a factor known to stimulate cell proliferation and induce genetic instability, are elevated in cancer cells, and are eliminated from the cells by SOD (35).
Akt protein is involved in several regulatory pathways controlling cell proliferation and survival and is thought to play an important role during cancer progression (36); thus, we have analyzed its expression. Although expression level was not changed in the siPGI/AMF cells as compared with control cells, its phosphorylation level was almost completely abolished (Fig. 4D). These results give credence to the prior results showing that overexpression of PGI/AMF in NIH-3T3 fibroblasts leads to accumulation of cellular phosphorylated Akt protein (24).
Next, we tested the effect of down-regulation of PGI/AMF expression on the expression level of MDM2 protein, which is phosphorylated by Akt, inhibits the transcriptional activity of p53, and induces p53 degradation as an E3 ubiquitin ligase (36,37). Unexpectedly, MDM2 protein was not decreased, but rather increased in siPGI/AMF cells (Fig. 4E), suggesting that up-regulated p53 by PGI/AMF knockdown may stimulate the expression of MDM2 (p53-MDM2 feedback loop; Ref. 37).
To verify our proposed signaling pathway, PGI/AMF knockdown cells were transfected with siRNA specific to the SODs, which is the most upstream factor in our proposed pathway. Western blot analysis revealed that knockdown of SODs by siRNA down-regulated p53 and p21 protein expressions and up-regulated ROS and phosphorylated Akt levels, although SOD knockdown did not affect PGI/ AMF and MDM2 expressions (Fig.  5, A and B). Moreover, suppression of MnSOD, not CuZnSOD, in PGI/ AMF knockdown cells showed a significantly lower rate of cellular senescence as compared with the control siPGI/AMF cells (Fig. 5C). These results suggest that SOD is upstream factor of p21, and SOD expression, especially MnSOD expression, has an important role in oxidative stress-induced cellular senescence. We proceeded to explore the expression of p21 protein, a critical mediator of cellular senescence, and its expression level was markedly increased following treatment with H 2 O 2 (Fig. 6). In addition, CuZn-SOD and MnSOD protein levels were clearly induced by H 2 O 2 treatment in both control and PGI/AMF knockdown cells (Fig. 6). The level of MnSOD was highly elevated by H 2 O 2 treatment in siPGI/AMF cells, whereas CuZnSOD showed a slight increase in the H 2 O 2 -treated siPGI/AMF cells, and the same level  of CuZnSOD expression was seen in both treatment groups (Fig. 6). Because the level of p21 expression was correlated with that of MnSOD level, we suggest that PGI/AMF knockdown increases sensitivity to oxidative stress.
Effect of PGI/AMF Inhibitor on Cellular Senescence-Because PGI/AMF can be neutralized by specific inhibitor carbohydrate phosphates that inhibit both enzymatic and cytokine activities (10,38), we examined whether carbohydrate phosphates could induce cellular senescence in HT1080 cells. In the presence of the carbohydrate phosphate erythrose 4-phosphate, an increase in the number of SA ␤-gal-positive cells was observed following exposure to the H 2 O 2 oxidative stress (Fig. 7A). Moreover, MnSOD and p21 expression levels were up-regulated by erythrose 4-phosphate treatment with H 2 O 2 exposure (Fig. 7B). These results also support the notion that PGI/AMF expression/activity is involved in cellular senescence response to oxidative stress, and MnSOD is a downstream factor of PGI/ AMF in oxidative stress-induced cellular senescence.

DISCUSSION
PGI/AMF has been reported to regulate proliferation and survival of cells and prevent stress-induced apoptosis in tumor cells (24,26,27). Here, we report that PGI/AMF knockdown induces cellular senescence in tumor cells in response to oxidative stress by investigating SA ␤-gal expression, a well known marker of cellular senescence. This is also supported by the observation that PGI/AMF inhibitor induces senescence in tumor cells under oxidative stress conditions. To study this phenomenon, we have used the human fibrosarcoma cell line HT1080, which harbors wild-type p53 protein (39). Wild-type p53 is a tumor suppressor protein, is a transcription factor that activates the transcription of downstream effectors genes and affects cell cycle arrest at the G 1 and G 2 checkpoints in response to DNA damage. Among the p53-inducible genes, the cyclindependent kinase inhibitor p21 is thought to be a critical mediator of cellular senescence (4,6,40). The data presented here show that down-regulation of PGI/AMF induces p21 expression in tumor cells and is markedly enhanced under oxidative stress conditions, suggesting that PGI/AMF is an important regulator in the senescence program through the p21 pathway.
What signaling molecules are upstream of p21? The data indicate that SOD expression, especially that of MnSOD, is upregulated by oxidative stress under conditions of insufficient expression of PGI/AMF, and therefore the generation of ROS is decreased because of the increase of its scavenger (35). Reduction of ROS leads to a decrease in Akt activation (41,42), resulting in up-regulation of p21 expression (36,43), which promotes cellular senescence. In addition, it may be suggested that p53 initiates senescent growth arrest at least in part by inducing p21, and elevated p53 may induce MDM2 expression, which in turn binds p53 and promotes p53 degradation via the ubiquitinproteasome pathway (36,37). Thus, no simple mechanism regulates cellular senescence, and based on our data we suggest the pathway in Fig. 8.
Many tumors exhibit strong suppressing function(s) on cellular antioxidant defense systems such as SOD, catalase, glutathione, etc., and one of the most effective intracellular enzymatic antioxidants is SOD. SOD exists in several isoforms in humans, and cytosolic CuZnSOD and mitochondrial MnSOD are two well known subtypes (44). The activity of total SOD (CuZnSOD and MnSOD) has also been found to be reduced for certain tumor cells (32). Cancers from a variety of cell types have been reported to exhibit decreased MnSOD activity (32), overexpression of MnSOD leads to tumor growth retardation (45), and elevated MnSOD activity is correlated with a loss of metastatic abilities of tumor cells (33). Although it is clear that MnSOD has tumor-suppressing activity, the mechanism of action is still unclear. At least one mechanism that MnSOD decreases tumor growth is by causing a delay at G 0 /G 1 in the cell cycle (46). We show here that SOD expression could be induced under oxidative stress condition and was controlled by PGI/ AMF. In addition the data show that ROS production was fol-  lowed by Akt stimulation, and Akt-p21 signaling axis could be under the control of PGI/AMF, signifying that SOD-ROS-Akt-p21 pathway is an important tumor suppressor mechanism through cellular senescence, and PGI/AMF could be involved in the protection of tumor cells from the onset of senescence via SOD pathway.
The development of cancer is initiated by a multiplicity of genetic and epigenetic alterations. Moreover, tumor cells need to escape normal cell barriers that limit proliferation. One of these barriers is cellular senescence considered to function as a powerful tumor suppressive mechanism, limiting the proliferative capacity of cells (3,4). Therefore, it is suggested that signal mediators in the cellular senescence programs might or should be a novel target for cancer therapy/prevention. The data presented here strongly suggest that PGI/AMF could be one of such therapeutic targets. Previously we reported that overex-pression of PGI/AMF in NIH-3T3 fibroblasts leads to a gain of tumorigenicity, whereas down-regulation of PGI/AMF in mesenchymal tumor cells leads to mesenchymal-to-epithelial transition as reflected by lower proliferating and loss of tumorigenic properties in vitro and in vivo (24,26).
In conclusion, the data signify that PGI/AMF plays a regulatory role in the control of cellular senescence and suggest that cellular senescence by oxidative stress is induced, in part, via the PGI/AMF pathway, and that the function of p21 cyclin-dependent kinase inhibitor during senescence is PGI/AMF-dependent. The data presented here provide a novel insight not previously described on the understanding of cellular senescence, suggest a novel therapeutic target, and underscore the role of glycolysis and gluconeogenesis enzymes in the pathobiology of the cell.