Dictyostelium Differentiation-Inducing Factor-1 Promotes Glucose Uptake, at Least in Part, via an AMPK-Dependent Pathway in Mouse 3T3-L1 Cells

Differentiation-inducing factor-1 (DIF-1) is a chlorinated alkylphenone (a polyketide) found in the cellular slime mold Dictyostelium discoideum. DIF-1 and its derivative, DIF-1(3M) promote glucose consumption in vitro in mammalian cells and in vivo in diabetic rats; they are expected to be the leading antiobesity and antidiabetes compounds. In this study, we investigated the mechanisms underlying the actions of DIF-1 and DIF-1(3M). In isolated mouse liver mitochondria, these compounds at 2–20 μM promoted oxygen consumption in a dose-dependent manner, suggesting that they act as mitochondrial uncouplers, whereas CP-DIF-1 (another derivative of DIF-1) at 10–20 μM had no effect. In confluent mouse 3T3-L1 fibroblasts, DIF-1 and DIF-1(3M) but not CP-DIF-1 induced phosphorylation (and therefore activation) of AMP kinase (AMPK) and promoted glucose consumption and metabolism. The DIF-induced glucose consumption was reduced by compound C (an AMPK inhibitor) or AMPK knock down. These data suggest that DIF-1 and DIF-1(3M) promote glucose uptake, at least in part, via an AMPK-dependent pathway in 3T3-L1 cells, whereas cellular metabolome analysis revealed that DIF-1 and DIF-1(3M) may act differently at least in part.


Introduction
The cellular slime mold Dictyostelium discoideum is an excellent model organism for cell and developmental biology; at the end of its development, it forms fruiting bodies, each consisting of spores and a multicellular stalk [1,2]. Differentiation-inducing factor-1 (DIF-1) ( Figure 1A), a chlorinated alkylphenone (a polyketide), functions as an inducer of stalk cell differentiation and also as a modulator of chemotactic cell movement in the development of D. discoideum [3][4][5][6]. DIF-3 ( Figure 1B) is the first metabolite produced during the degradation of DIF-1 and has virtually no function in D. discoideum [4][5][6][7].
In this study, to elucidate the mechanism(s) of the glucose uptake-promoting activity of DIF-1 and DIF-1(3M), we first analyzed the effects of some DIF derivatives on mitochondrial oxygen consumption (MOC) and AMPK phosphorylation in confluent 3T3-L1 fibroblasts, and then assessed the involvement of AMPK in DIF-promoted glucose consumption by manipulating AMPK activity. We show that DIF-1 and DIF-1(3M) may promote glucose uptake, at least in part, via an AMPK-dependent pathway in 3T3-L1 cells.

Effects of DIF Derivatives, DNP, and AICAR on Glucose Consumption in 3T3-L1 Cells
To assess the relationship between AMPK activation and glucose uptake (consumption) induced by the compounds, we analyzed the effects of DIF derivatives, DNP, and AICAR on glucose consumption ( Figure 4). As expected, DIF-1 and DIF-1(3M) at 20 µM promoted glucose consumption by more than 2-fold [21,22], while CP-DIF-1 at 20 µM did not ( Figure 4A). DNP at 0.05 or 0.1 mM promoted glucose consumption by 1.3-1.7-fold ( Figure 4A). DIF derivatives at 20 µM and DNP at 0.1 mM were not toxic to the cells ( Figure 4C). AICAR at 0.05-0.2 mM did not significantly promote glucose consumption ( Figure 4A,B) and was toxic to the cells at 0.5 mM ( Figure 4C), although it strongly activated AMPK at 0.2-2 mM ( Figure 3D). Since AICAR can activate many other AMP-dependent en-zymes [41], long-term stimulation (15-20 h) with AICAR might disturb some cell function and thus cause toxicity to 3T3-L1 cells. AMPK activation by DIF derivatives or DNP was much weaker than that by AICAR ( Figure 3). Taken together, these results suggest that the involvement of AMPK activation, if any, in the actions of DIF-1 and DIF-1(3M) may be partial.

Effects of AMPK Inhibition on DIF-Promoted Glucose Uptake in 3T3-L1 Cells
We then examined the effects of compound C, an inhibitor of AMPK, on glucose uptake in the presence of AICAR, DNP, DIF-1, or DIF-1(3M) ( Figure 5). In DMSO control cells, 15 µM compound C reduced the basal rate of glucose consumption by 20%. AICAR at 0.1 mM slightly but significantly promoted glucose consumption, and this effect was completely inhibited in the presence of compound C. Glucose consumption was significantly increased by DNP at 50 µM (1.5-fold) and by DIF-1 or DIF-1 (3M) at 15 µM (2-fold).
We then examined the effects of AMPKα knockdown on glucose consumption in the presence of DIF-1 or DIF-1(3M) ( Figure 6). Transfection with siRNA against AMPKα performed twice decreased the AMPKα protein level by 80% in confluent 3T3-L1 cells ( Figure 6A) and significantly reduced the rate of glucose consumption in the presence of 0.2% DMSO or 20 µM DIF-1 or DIF-1(3M) ( Figure 6B). Importantly, however, both DIF-1 and DIF-1(3M) significantly increased the rate of glucose consumption (1.8-fold) under the conditions ( Figure 6B) despite AMPKα knockdown throughout the assay ( Figure 6C).
These results suggest that DIFs promote glucose uptake via an AMPK-dependent pathway, at least in part. On the other hand, DIFs may function via an AMP-independent pathway in parallel with the AMPK-dependent pathway, although we cannot exclude that DIFs promoted glucose consumption via a very small amount of remaining AMPK activity in the presence of compound C ( Figure 5) or AMPK protein under the AMPK knockdown conditions ( Figure 6).
calculated. The mean values and SD of the three independent experiments are presented. ** p < 0.01, † p < 0.05, # p < 0.05 versus control; n.s., not significant. (B) Cells were incubated for 15-20 h in the presence of the indicated additives and the combined effects of DIF-1(3M) and AICAR or DNP on glucose consumption were also assessed as in (A). The mean values and SD of the four independent experiments are presented. * p < 0.05, ** p < 0.01; n.s., not significant. (C) Cells were incubated for 20 h in the presence or absence of the indicated concentrations of the additives and observed by a phase-contrast microscope.

Effects of AMPK Inhibition on DIF-Promoted Glucose Uptake in 3T3-L1 Cells
We then examined the effects of compound C, an inhibitor of AMPK, on glucose uptake in the presence of AICAR, DNP, DIF-1, or DIF-1(3M) ( Figure 5). In DMSO control cells, 15 μM compound C reduced the basal rate of glucose consumption by 20%. AICAR at 0.1 mM slightly but significantly promoted glucose consumption, and this effect was completely inhibited in the presence of compound C. Glucose consumption was significantly increased by DNP at 50 μM (1.5-fold) and by DIF-1 or DIF-1(3M) at 15 μM (2-fold).  We then examined the effects of AMPKα knockdown on glucose consumption in the presence of DIF-1 or DIF-1(3M) ( Figure 6). Transfection with siRNA against AMPKα performed twice decreased the AMPKα protein level by 80% in confluent 3T3-L1 cells ( Figure  6A) and significantly reduced the rate of glucose consumption in the presence of 0.2% DMSO or 20 μM DIF-1 or DIF-1(3M) ( Figure 6B). Importantly, however, both DIF-1 and DIF-1(3M) significantly increased the rate of glucose consumption (1.8-fold) under the conditions ( Figure 6B) despite AMPKα knockdown throughout the assay ( Figure 6C). These results suggest that DIFs promote glucose uptake via an AMPK-dependent pathway, at least in part. On the other hand, DIFs may function via an AMP-independent pathway in parallel with the AMPK-dependent pathway, although we cannot exclude that DIFs promoted glucose consumption via a very small amount of remaining AMPK activity in the presence of compound C ( Figure 5) or AMPK protein under the AMPK

Effects of DIF Derivatives on Glucose Metabolism in 3T3-L1 Cells
Using CE-TOFMS (capillary electrophoresis time-of-flight mass spectrometry), we performed a metabolome analysis and showed that DIF-1 and DIF-1(3M) promoted glucose metabolism but did not significantly affect cellular ATP level in 3T3-L1 cells [23]. In this study, to further assess the differences, if any, in the effects of the DIF derivatives, we used the same approach to analyze the effects of 20 µM DIF-1, DIF-1(3M), and CP-DIF-1 on glucose metabolism and the AMP/ATP ratio ( Figure 7A). DIF-1 and DIF-1(3M) tended to increase the glucose metabolite levels but did not significantly affect cellular ATP level ( Figure 7A). DIF-1(3M) increased the AMP/ATP ratio slightly but significantly, while DIF-1 tended to increase it ( Figure 7A). At the same concentration, CP-DIF-1 did not significantly affect glucose metabolite levels, ATP level, or the AMP/ATP ratio ( Figure 7A). A heat map of cellular metabolites showed considerable differences between cells incubated with DIF-1 or DIF-1(3M) and DMSO control cells, while CP-DIF-1 slightly affected the metabolites in comparison with the DMSO control ( Figure 7B). The effects of DIF-1 and DIF-1(3M) on cellular metabolites differed considerably from each other (Figures 3A and 4). These results suggest that the mechanisms underlying the actions of the two DIFs differ from each other, at least in part. , or CP-DIF-1, and metabolite levels per 10 6 cells were determined by use of CE-TOFMS to construct a metabolome pathway map. The metabolites levels in the control cells were set to 1, and the relative amounts in the DIF-treated cells are presented as the mean and SD of the duplicate samples. * p < 0.05 versus Control. N.D., not detected; G6P, glucose 6-phosphate; F6P, fructose 6-phosphate; F1,6P, fructose 1,6-diphosphate; G3P, glyceraldehyde 3-phosphate; 3PG, 3-phosphoglyceric acid; 2PG, 2-phosphoglyceric acid; PEP, phosphoenolpyruvic acid; AcCoA, acetyl CoA_divalent; MalCoA, malonyl CoA_divalent; cis-Aco, cis-aconitic acid; IsCit, isocitric acid; 2OG, 2-oxoglutaric acid. (B) Heat map of cellular metabolites. Of the 205 metabolite peaks identified in this study, metabolites showing similar relative abundance throughout the 4 duplicated samples were clustered into a metabolome heat map. The horizontal axis shows the sample names, and the vertical axis shows the metabolites. HCA (hierarchical cluster analysis) was performed, and the distance between the peaks is shown in the dendrogram. Green, small average abundance; red, large average abundance.
In the light of this special issue "Cancer Biology in Diabetes", it should be noted that A heat map of cellular metabolites showed considerable differences between cells incubated with DIF-1 or DIF-1(3M) and DMSO control cells, while CP-DIF-1 slightly affected the metabolites in comparison with the DMSO control ( Figure 7B). The effects of DIF-1 and DIF-1(3M) on cellular metabolites differed considerably from each other ( Figures 3A and 4). These results suggest that the mechanisms underlying the actions of the two DIFs differ from each other, at least in part.
In the light of this special issue "Cancer Biology in Diabetes", it should be noted that AMPK activators such as AICAR and metformin can inhibit tumor cell growth [40,[44][45][46][47], and metformin has been used in some clinical trials [45,48]. Anticancer and antidiabetes agents such as DIF derivatives may have some common mechanisms of action, which we intend to investigate in the future.

Involvement of AMPK in the Actions of DIF-1 and DIF-1(3M)
DIF-1 triggers GLUT1 translocation from an intracellular pool to the plasma membrane via a PI3K/Akt-independent pathway, thus promoting glucose uptake in both 3T3-L1 fibroblasts and 3T3-L1 adipocytes [21]. DIF-1 and DIF-1(3M) also promote the metabolism of glucose taken up by the cells [23].
Since mitochondrial uncouplers have been shown to promote glucose uptake by activating AMPK [32,[34][35][36][37], in the present study we assessed the involvement of AMPK in DIF-induced glucose uptake in 3T3-L1 cells, comparing the effects of DIF-1, DIF-1(3M), and CP-DIF-1. We showed here that (1) DIF-1 and DIF-1(3M) but not CP-DIF-1 promoted MOC (Figure 2), (2) DIF-1 and DIF-1(3M) but not CP-DIF-1 induced the phosphorylation (and therefore activation) of AMPKα (Figure 3), (3) DIF-1 and DIF-1(3M) but not CP-DIF-1 promoted glucose uptake (Figure 4), and (4) suppression of AMPK activity significantly reduced the glucose uptake induced by DIF-1 and DIF-1(3M) (Figures 5 and 6). These results suggest that DIF-1 and DIF-1(3M) promote glucose uptake by mitochondrial uncoupling and subsequent activation of AMPK, at least in part ( Figure 8). However, since neither compound C nor AMPK knockdown completely inhibited DIF-induced glucose uptake ( Figures 5 and 6) and also because the AMPK activator, AICAR (Figure 3), did not promote ( Figure 4A) or only slightly promoted glucose uptake ( Figure 5A), DIF-1 and DIF-1(3M) may also promote glucose uptake via an AMPK-independent pathway ( Figure 8B). duced the glucose uptake induced by DIF-1 and DIF-1(3M) (Figures 5 and 6). These results suggest that DIF-1 and DIF-1(3M) promote glucose uptake by mitochondrial uncoupling and subsequent activation of AMPK, at least in part ( Figure 8). However, since neither compound C nor AMPK knockdown completely inhibited DIF-induced glucose uptake (Figures 5 and 6) and also because the AMPK activator, AICAR (Figure 3), did not promote ( Figure 4A) or only slightly promoted glucose uptake ( Figure 5A), DIF-1 and DIF-1(3M) may also promote glucose uptake via an AMPK-independent pathway ( Figure 8B). These compounds reduce ATP production and thus increase the AMP/ATP ratio, activating AMP kinase and promoting glucose uptake. AICAR is an activator and compound C is an inhibitor of AMP kinase. (B) Proposed scheme for the actions of DIF. By uncoupling mitochondrial activities, DIF may disturb ATP production and activate AMPK, which may then induce GLUT1 translocation to the plasma membrane and glucose uptake. It is generally unknown how AMPK activation induces GLUT1 translocation, whereas AMPK activation may inhibit GLUT1 internalization (endocytosis) to promote glucose uptake by triggering the degradation of TXNIP (thioredoxin-interacting protein), a stimulator of GLUT1 endocytosis [49,50]. Note that DIF may induce GLUT1 translocation partly via an AMPK-independent pathway, but it remains to be elucidated how DIF induces GLUT1 translocation. Glucose may be metabolized immediately via glycolysis and via the TCA cycle.
In this study, we analyzed the metabolic pathway of glucose in the presence of three DIF derivatives and found that the mitochondrial uncouplers DIF-1 and DIF-1(3M) but not the non-uncoupler CP-DIF-1 promoted glucose metabolism without affecting the cellular ATP level ( Figure 7A); the DIF-1 and DIF-1(3M) data agree well with our previous results [23]. A slight increase in the AMP/ATP ratio by DIF-1 and DIF-1(3M) ( Figure 7A) might activate AMPK (Figure 3).
We also revealed that the metabolomes of cells treated with DIF-1 or DIF-1(3M) differed from each other ( Figure 7B), suggesting that the two compounds promote glucose consumption via different pathways, at least in part. We will further elucidate the precise mechanisms underlying the actions of DIF-1 and DIF-1(3M) (i.e., the blank part of the These compounds reduce ATP production and thus increase the AMP/ATP ratio, activating AMP kinase and promoting glucose uptake. AICAR is an activator and compound C is an inhibitor of AMP kinase. (B) Proposed scheme for the actions of DIF. By uncoupling mitochondrial activities, DIF may disturb ATP production and activate AMPK, which may then induce GLUT1 translocation to the plasma membrane and glucose uptake. It is generally unknown how AMPK activation induces GLUT1 translocation, whereas AMPK activation may inhibit GLUT1 internalization (endocytosis) to promote glucose uptake by triggering the degradation of TXNIP (thioredoxin-interacting protein), a stimulator of GLUT1 endocytosis [49,50]. Note that DIF may induce GLUT1 translocation partly via an AMPK-independent pathway, but it remains to be elucidated how DIF induces GLUT1 translocation. Glucose may be metabolized immediately via glycolysis and via the TCA cycle.
In this study, we analyzed the metabolic pathway of glucose in the presence of three DIF derivatives and found that the mitochondrial uncouplers DIF-1 and DIF-1(3M) but not the non-uncoupler CP-DIF-1 promoted glucose metabolism without affecting the cellular ATP level ( Figure 7A); the DIF-1 and DIF-1(3M) data agree well with our previous results [23]. A slight increase in the AMP/ATP ratio by DIF-1 and DIF-1(3M) ( Figure 7A) might activate AMPK (Figure 3).
We also revealed that the metabolomes of cells treated with DIF-1 or DIF-1(3M) differed from each other ( Figure 7B), suggesting that the two compounds promote glucose consumption via different pathways, at least in part. We will further elucidate the precise mechanisms underlying the actions of DIF-1 and DIF-1(3M) (i.e., the blank part of the scheme in Figure 8B) and try to develop novel antiobesity and antidiabetes agents on the basis of these compounds.
(and AMPKβ and GAPDH expression for comparison). For glucose consumption assay, the media used for RNAi were replaced with DMEM-MG (1 mL/well), and the cells were incubated for 2 h. Then, the cells were incubated for 8-15 h with fresh DMEM-MG (1 mL/well) containing 0.2% (v/v) DMSO, 20 µM DIF-1, or 20 µM DIF-1(3M), and glucose consumption was assessed as described in Section 4.3. After the glucose consumption assay, the cells were used for Western blotting to check AMPKα expression (and AMPKβ and GAPDH expression for comparison) again.

Metabolome Analysis
Confluent 3T3-L1 cells in 90-mm tissue culture dishes were incubated for 3 h with 10 mL of DMEM-MG containing 0.1% (v/v) DMSO or 20 µM DIF-1, DIF-1(3M), or CP-DIF-1; the assay was performed in duplicate. The culture media were removed, and the cells were washed with 10 mL per dish of 5% (w/v) mannitol solution and then 2 mL of the same solution. The cells were collected by scraping in methanol (1.3 mL/well) containing 10 µM internal standard solution (Human Metabolome Technologies, Tokyo, Japan) and transferred into eight centrifugation tubes. Ionic metabolites were analyzed by capillary electrophoresis time-of-flight mass spectrometry (Agilent CE-TOFMS system; Agilent Technologies, Waldbronn, Germany) as described previously [23,[54][55][56][57]. Relative quantification data for the identified metabolites were used for hierarchical cluster analysis (HCA) and principal component analysis (PCA) performed with the proprietary software, PeakStat and SampleStat (Human Metabolome Technologies), respectively, to produce a metabolome heat map and a metabolome pathway-map.

Statistical Analysis
Welch's t-test was used for the statistical analyses. Values were considered to be significantly different when the p value was less than 0.05.

Patents
The following authors hold a patent related to this article: Kubohara