Báo cáo khoa học: Dictyostelium differentiation-inducing factor-1 induces glucose transporter 1 translocation and promotes glucose uptake in mammalian cells pdf

glucose transporter 1 translocation and promotes glucose uptake in mammalian cells Waka Omata1, Hiroshi Shibata1, Masahiro Nagasawa1, Itaru Kojima1, Haruhisa Kikuchi2, Yoshiteru Oshima2,

glucose transporter 1 translocation and promotes glucose uptake in mammalian cells

Waka Omata1, Hiroshi Shibata1, Masahiro Nagasawa1, Itaru Kojima1, Haruhisa Kikuchi2,

Yoshiteru Oshima2, Kohei Hosaka3 and Yuzuru Kubohara4

1 Department of Molecular Medicine, Institute for Molecular and Cellular Regulation, Gunma University, Maebashi, Japan

2 Graduate School of Pharmaceutical Sciences, Tohoku University, Sendai, Japan

3 Department of Basic Sciences for Medicine, Gunma University School of Health Sciences, Maebashi, Japan

4 Department of Molecular and Cellular Biology, Institute for Molecular and Cellular Regulation, Gunma University, Maebashi, Japan

Keywords

antitumor agent; Dictyostelium; DIF-1;

glucose uptake; GLUT1

Correspondence

Y Kubohara, Department of Molecular and

Cellular Biology, Institute for Molecular and

Cellular Regulation, Gunma University,

Maebashi 371-8512, Japan

Fax: +81 27 2208834

Tel: +81 27 2208831

E-mail: kubohara@showa.gunma-u.ac.jp

Website: http://imcr.showa.gunma-u.ac.jp/

index-e.htm

(Received 26 February 2007, revised 3 May

2007, accepted 8 May 2007)

doi:10.1111/j.1742-4658.2007.05872.x

The differentiation-inducing factor-1 (DIF-1) is a signal molecule that induces stalk cell formation in the cellular slime mold Dictyostelium dis-coideum, while DIF-1 and its analogs have been shown to possess anti-proliferative activity in vitro in mammalian tumor cells In the present study, we investigated the effects of DIF-1 and its analogs on normal (nontransformed) mammalian cells Without affecting the cell morphology and cell number, DIF-1 at micromolar levels dose-dependently promoted the glucose uptake in confluent 3T3-L1 fibroblasts, which was not inhib-ited with wortmannin or LY294002 (inhibitors for phosphatidylinositol 3-kinase) DIF-1 affected neither the expression level of glucose trans-porter 1 nor the activities of four key enzymes involved in glucose meta-bolism, such as hexokinase, fluctose 6-phosphate kinase, pyruvate kinase, and glucose 6-phosphate dehydrogenase Most importantly, stimulation with DIF-1 was found to induce the translocation of glucose trans-porter 1 from intracellular vesicles to the plasma membranes in the cells

In differentiated 3T3-L1 adipocytes, DIF-1 induced the translocation of glucose trasporter 1 (but not of glucose transporter 4) and promoted glucose uptake, which was not inhibited with wortmannin These results indicate that DIF-1 induces glucose transporter 1 translocation and thereby promotes glucose uptake, at least in part, via a inhibitors for phosphatidylinositol 3-kinase⁄ Akt-independent pathway in mammalian cells Furthermore, analogs of DIF-1 that possess stronger antitumor activity than DIF-1 were less effective in promoting glucose consumption, suggesting that the mechanism of the action of DIF-1 for stimulating glucose uptake should be different from that for suppressing tumor cell growth

Abbreviations

DIF-1, differentiation-inducing factor-1 [1-(3,5-dichloro-2,6-dihydroxy-4-methoxyphenyl)hexan-1-one]; DIF-3, differentiation-inducing factor-3 [1-(3-chloro-2,6-dihydroxy-4-methoxyphenyl)hexan-1-one]; DMEM-HG, DMEM containing a high concentration of glucose; DMEM-LG, DMEM containing a low concentration of glucose; EtOH, ethanol; GLUT, glucose transporter; LDM, low-density microsome; LY294002,

2-(4-morpholinyl)-8-phenyl-1(4H)-benzopyran-4-one; 8-MIBMX, 8-methoxymethyl-3-isobutyl-1-methylxanthine; 2-MIDIF-1, 2-methoxy isomer of DIF-1; PDE1, calmodulin-dependent cyclic nucleotide phosphodiesterase; PI3K, phosphatidylinositol 3-kinase; PM, plasma membrane;

Tg, thapsigargin; THPH, 1-(2,4,6-trihydroxyphenyl)hexan-1-one.

The cellular slime mold Dictyostelium discoideum is an

excellent model organism in the fields of cell and

devel-opmental biology because of its simple pattern of life

cycle; this organism forms fruiting bodies, each

consist-ing of spores and a cylindrical multicellular stalk, at the

end of its development The differentiation-inducing

fac-tor-1 [DIF-1;

1-(3,5-dichloro-2,6-dihydroxy-4-methoxy-phenyl)hexan-1-one] (Fig 1) is a lipophilic signal

molecule that was identified as a stalk-cell-inducing

fac-tor in D discoideum [1,2] DIF-3

[1-(3-chloro-2,6-dihyd-roxy-4-methoxyphenyl)hexan-1-one] (Fig 1), the initial product in the process of DIF-1 breakdown, is much less active in inducing stalk cell differentiation [2–4] DIF-1

is thought to function at least in part via an increase in the cytoplasmic calcium concentration ([Ca2+]c) [5–7], but the precise signaling system of DIF-1, including the target molecule(s) of DIF-1, is still unknown

On the other hand, it has been shown that DIF-1 and DIF-3 (designated DIFs) exhibit strong antiproliferative activity and occasionally induce cell differentiation

in vitro in mammalian cells [8–16] and that DIF-3 and some derivatives of DIF-3 are the most potent anti-tumor agents among the DIF analogs tested to date [12,16,17] As to the mechanism of the actions of DIFs,

it has been shown that: (a) DIFs increase [Ca2+]c in some tumor cells [9–12], (b) DIFs activate Akt⁄ protein kinase B in human leukemia K562 cells [13], (c) DIF-1 inactivates STAT3 in gastric cancer cells [15], and (d) DIFs inhibit the expression of cyclins D⁄ E and the phosphorylation of the retinoblastoma protein in vascu-lar smooth muscle cells [14] and K562 cells [18] Recently, we have found that calmodulin-dependent cyclic nucleotide phosphodiesterase (PDE1) is a phar-macological and specific target of DIFs in mammalian cells [19] Yet, the mechanisms underlying the actions of DIFs in mammalian cells remain to be elucidated Meanwhile, in the course of a study of the in vitro actions of DIFs in mammalian normal (nontrans-formed) cells, we noticed that DIF-1 promotes color conversion (red to yellow) of the incubation media of rat gastric mucosal RGM-1 and rat leptomeningeal cells [20], which suggests that DIF-1 might promote the cellular metabolism in some way

In the present study, in order to assess the newly found function of DIF-1, we examined the effects of DIF-1 and its analogs (Fig 1) on glucose consumption in vitro in RGM-1 cells, mouse 3T3-L1 fibroblasts, and 3T3-L1 adi-pocytes, and investigated the mechanism of the novel function of DIF-1 We show here that DIF-1 induces the translocation of glucose transporter (GLUT) 1 from intracellular vesicles to the plasma membrane and thereby promotes glucose uptake, at least in part, via a phosphatidylinositol 3-kinase (PI3K)⁄ Akt-independent pathway in the cells We also show that DIF-1 possesses the most potent activity among its analogs tested

Results Effect of DIF-1 on glucose consumption in 3T3-L1 fibroblasts and RGM-1 cells

We first assessed the cytotoxic effect of DIF-1 on con-fluent 3T3-L1 cells (Fig 2) Neither ethanol (EtOH)

Fig 1 Chemical structure of DIF-1, DIF-3 and their derivatives.

(A) DIF-1;

1-(3,5-dichloro-2,6-dihydroxy-4-methoxyphenyl)hexan-1-one DIF-3; 1-(3-chloro-2,6-dihydroxy-4-methoxyphenyl)hexan-1-1-(3,5-dichloro-2,6-dihydroxy-4-methoxyphenyl)hexan-1-one.

2-MIDIF-1; 2-methoxy isomer of DIF-1 DMPH;

1-(2,6-dihydroxy-4-methoxyphenyl)hexan-1-one THPH;

1-(2,4,6-trihydroxyphenyl)hex-an-1-one (B) Derivatives of DIF-1 that have alkyl chains of different

lengths (C) Derivatives of DIF-1 and DIF-3 that have a different

halogen.

(vehicle) at 0.2% (Fig 2A), nor

1-(2,4,6-trihydroxyphe-nyl)hexan-1-one (THPH) at 20 lm (Fig 2B) affected

cell morphology and viability but, as previously

repor-ted with RGM-1 cells [20], DIF-1 at 20 lm induced

cell death (Fig 2C) where medium acidification

(red-to-yellow color change) was promoted However, since

the apparent cytotoxic effect of DIF-1 was canceled

when the incubation medium was exchanged daily with

a fresh one containing DIF-1 (Fig 2F), it was likely

that DIF-1 would promote cellular metabolism in

some way, resulting in medium acidification and

subse-quent cell death

In order to assess what happens in the cells in the

presence of DIF-1, we examined the effect of DIF-1 on

glucose consumption by monitoring the glucose

concen-tration in the incubation media (Fig 3A) As expected,

DIF-1 at the micromolar levels promoted glucose

con-sumption in a dose-dependent manner (Fig 3A,B)

with-out affecting the cell number (Fig 3B) or cell

morphology (data not shown) Withdrawal of DIF-1 in

the media resulted in the restoration of the rate of glu-cose consumption to the normal level (Fig 3A), indica-ting that the promoindica-ting effect of DIF-1 is reversible It should be noted that similar effects of DIF-1 were observed using confluent RGM-1 cells (Fig 4A,B), sug-gesting that DIF-1 would be a broad promoter for glucose consumption in mammalian cells across species

Effect of DIF-1 on glucose uptake in 3T3-L1 fibroblasts

In order to confirm that DIF-1 should promote glucose consumption by the cells, we checked the effect of DIF-1 on glucose uptake in confluent 3T3-L1 cells (Fig 3C,D) The cellular activity of the 2-deoxy-glucose uptake was promoted in the presence of 20 lm DIF-1 and reached a saturated level after a 4-h incubation (Fig 3C) DIF-1 at 10–20 lm promoted the 2-deoxy-glucose uptake dose-dependently, while the 2-methoxy isomer of DIF-1 (2-MIDIF-1) at 20 lm did not result in such an activity (Fig 3D) The results agree quite well with those shown in Fig 3B, indicating that the glu-cose-uptake promoting activity of DIF-1 (and possibly its analogs) can be assessed approximately by monitor-ing the glucose concentration in the incubation media under the conditions In addition, the simple assay sys-tem for glucose consumption should be a convenient method for the broad screening of candidate molecules that promote the cellular metabolism

Effects of DIF analogs on glucose consumption

in 3T3-L1 fibroblasts

To assess the chemical structure–effect relationship and specificity of the action of DIF-1, we examined the effects of DIF analogs on glucose consumption in confluent 3T3-L1 cells (Fig 5) While DIF-1 at 20 lm greatly promoted glucose consumption, the same con-centration of 2-MIDIF-1, 1-(2,6-dihydroxy-4-methoxy-phenyl)hexan-1-one (DMPH), and THPH (Fig 1A) showed no effect or a very weak one (Fig 5A) Similar results were obtained with confluent RGM-1 cells (Fig 4C) These results indicate that two chlorines and

a methoxy group on the proper position of the ben-zene ring are important for the promoting effect of DIF-1 on glucose consumption and, in other words, that the action of DIF-1 should be specific to its chem-ical structure

Among DIF-1 derivatives that have different length

of alkyl chains (Fig 1B), DIF-1 exhibited the most potent activity for stimulating glucose consumption in 3T3-L1 cells (Fig 5B), indicating that the hexanone chain is most appropriate for the activity On the other

Medium exchange Continuous incubation

EtOH A

THPH B

DIF-1 C

EtOH D

THPH E

DIF-1 F

0.1 mm

Fig 2 Effect of DIF-1 on cell morphology in 3T3-L1 fibroblasts.

Confluent 3T3-L1 cells were incubated in vitro for 3 days

continu-ously with 0.2% EtOH (vehicle) (A), 20 l M THPH (B), or 20 l M

DIF-1 (C), or cells were incubated for 3 days by exchanging the

media every 24 h with 0.2% EtOH (vehicle) (D), 20 l M THPH (E),

or 20 l M DIF-1 (F) On day 3, the cells were observed by using a

phase-contrast microscope Most of the cells under continuous

incubation with DIF-1 were dead (C), but cells appeared to be

healthy when the incubation medium was exchanged (F)

Sche-matic color images of the culture wells are shown on the photos.

Scale bar, 0.1 mm.

hand, DIF-1 was the only agent that showed strong

activity among the halogen-substituted derivatives

(Figs 1C and 5C) in spite of the fact that all of them are

potent antiproliferative agents in K562 leukemia [17]

The results suggest that the mechanism of the action of

DIF-1 for stimulating glucose consumption should be

different from that for suppressing tumor cell growth

Effects of the [Ca2+]i-increasing agents and PDE1

inhibitor on glucose consumption in 3T3-L1

fibroblasts

It has been shown that DIF-1 increases [Ca2+]c in

mammalian cells [9–11,20] and that DIF-1 is a specific

inhibitor for the calmodulin-dependent cyclic

nucleo-tide phophodiesterase, PDE1 [19] We thus examined

the effects of two [Ca2+]c-increasing agents,

thapsi-gargin (an inhibitor of Ca2+-ATPase present in

endoplasmic reticula) and A23187 (a calcium iono-phore), and 8-methoxymethyl-3-isobutyl-1-methylxan-hine (8-MIBMX; an inhibitor of PDE1) on glucose consumption in confluent 3T3-L1 cells (Fig 5A) How-ever, these agents did not promote glucose consump-tion in the cells, suggesting that DIF-1 promotes glucose consumption in a way other than by increasing [Ca2+]cor inhibiting PDE1

To support our hypothesis that an increase in [Ca2+]c

is not involved in the DIF-1-promoted glucose con-sumption, we examined the effects of DIF analogs and thapsigargin (Tg) on [Ca2+]cin 3T3-L1 cells As shown

in Fig 6, DIF-1 at 20 lm indeed increased [Ca2+]c, while THPH showed no significant effect on [Ca2+]c Importantly, however, DIF-3 and Tg also increased [Ca2+]c considerably, regardless of the fact that both reagents scarcely promoted glucose consumption (Fig 5A) These results support our hypothesis

withdrawal of DIF-1

at 12 h

A

B

addition of DIF-1

0.5 0.6 0.7 0.8 0.9 1

None EtOH (0.2%)

Approx rate of glucose consumption (µg/ml/h) 0

10 20 30

DIF-1

Cell number (% of control)

0 20 40 60 80 100 120

DIF-1

0 5000

5000

10 000

15 000

20 000

25 000

2-MID-1

p < 0.05

p < 0.001 p < 0.001

0

10 000

20 000

30 000

Time of incubation (h)

Time of incubation (h)

C

D

DIF-1

15 000

25 000

n.s.

**

**

**

*

n.s.

**

**

Fig 3 Effect of DIF-1 on glucose consumption, glucose uptake, and cell number in 3T3-L1 fibroblasts (A) Confluent 3T3-L1cells were incu-bated for 12 h without or with 0.2% EtOH (vehicle) or 10–20 l M DIF-1, the incubation media were discarded, and all the cells were further incubated with fresh media in the absence of EtOH or DIF-1 for 18 h The glucose concentration of each well was measured at the indicated time points, and the mean ± SD (bars) values of the triplicate (n ¼ 3) are presented Note that the rate of glucose consumption by the con-trol cells appear to increase along with the incubation time, which would be the allowable margin of error by the simple assay system (B) Confluent 3T3-L1cells were incubated for 18 h without or with 0.2% EtOH or 10–20 l M DIF-1 The glucose concentration of each well was measured, and the approximate rate of glucose consumption was calculated The cell number of each well was then assessed using a cell-number indicator The mean ± SD (bars) values of the triplicate (n ¼ 3) are presented **P < 0.001 versus others n.s., not significant (C) Confluent 3T3-L1cells were incubated for the indicated period with 20 l M DIF-1, and the 2-deoxy-glucose uptake was assessed as des-cribed in Experimental procedures All samples were prepared in quartettes, and the mean ± SD (bars) values of each quartette (n ¼ 4) are presented Note that it takes approximately 4 h for glucose uptake activity to reach a maximal level **P < 0.001 versus *0-time control n.s., not significant (D) Confluent 3T3-L1cells were incubated for 4 h without or with 0.2% EtOH (vehicle), 10–20 l M DIF-1, or 20 l M

2-MIDIF-1 (2-MID-1), and the 2-deoxy-glucose uptake was assessed The mean ± SD (bars) values of the triplicate (n ¼ 3) are presented The results agree well with those of the simple assay in (B).

Effect of DIF-1 on GLUT1 translocation

in 3T3-L1 fibroblasts

Insulin and some agents are known to promote glucose

uptake by inducing translocation of glucose

transport-ers from cytosolic store vesicles to the plasma

mem-brane [21,22] In 3T3-L1 fibroblasts, the major activity

of glucose uptake can be attributed to the glucose

transporter isoform GLUT1 [23,24] We therefore

examined the effect of DIF-1 on GLUT1 translocation

in 3T3-L1 cells and found that DIF-1 at 20 lm

stimu-lated GLUT1 translocation from low-density

micro-somes (LDMs) to the plasma membrane (PM) by

approximately two-fold (Fig 7) These results indicate

that DIF-1 promotes glucose uptake, at least in part,

by stimulating GLUT1 translocation

Effects of DIF-1 and PI3K inhibitors on glucose

uptake and GLUT1 expression

DIF-1 has been shown to activate the PI3K⁄ Akt

path-way in human leukemia K562 [13,18] and human gastric

cancer AGS cells [15] On the other hand, the activation

of PI3K and Akt is involved in insulin actions to promote glucose uptake and glycolysis [25,26]

In order to assess whether PI3K and Akt are involved

in the action of DIF-1 in 3T3-L1 cells, we examined the effect of DIF-1 on glucose consumption (Fig 8A) and glucose uptake (Fig 8B) in the presence of wortmannin

or LY294002 [2-(4-morpholinyl)-8-phenyl-1(4H)-benzo-pyran-4-one], inhibitors for PI3K However, DIF-1 pro-moted glucose consumption even in the presence of the inhibitors Western analysis showed that DIF-1 might slightly activate (phosphorylate) Akt, which was per-fectly inhibited with wortmannin (Fig 8C) Import-antly, DIF-1 did not affect the total amount of GLUT1 present in the cells (Fig 8C), but coaddition of DIF-1 and wortmannin decreased the expression of GLUT1 (Fig 8C), resulting in a due decrease in glucose con-sumption (uptake) (Fig 8A,B) Nevertheless, DIF-1 promoted glucose consumption (uptake) even in the presence of the inhibitor (Fig 8A,B) Similar results were obtained with confluent RGM-1 cells (data not shown) These results suggest that DIF-1 should promote the glucose uptake in some way, at least in part, via a PI3K⁄ Akt-independent pathway without

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Time of incubation (h)

withdrawal of DIF-1 at 14 h addition of DIF-1

none

15 µ M DIF-1

30 µ M DIF-1

0 10 20 30 40 50

**

B

0 10 20 30 40

DIF-1 (µM )

0 20 40 60 80 100

120

DIF-1 (µM )

**

*

*

Fig 4 Effects of DIF-1 and its analogs on glucose consumption in RGM-1 cells (A) Confluent RGM-1 cells were incubated for 14 h with 15–30 l M DIF-1, the incubation media were discarded, and all the cells were incubated with fresh media in the absence of EtOH or DIF-1 for

12 h The glucose concentration of each well was measured at the indicated time points, and the mean ± SD (bars) values of the triplicate (n ¼ 3) are presented (B) Confluent RGM-1cells were incubated for 12 h without or with 8–30 l M DIF-1 The glucose concentration of each well was measured, and the approximate rate of glucose consumption was calculated The cell number of each well was then assessed using a cell-number indicator The mean ± SD (bars) values of the triplicate (n ¼ 3) are presented *P < 0.02 versus others **P < 0.001 ver-sus others (C) Confluent RGM-1 cells were incubated for 12 h without or with 0.2% EtOH (vehicle) or 20 l M DIF analogs The glucose con-centration of each well was measured, and the approximate rate of glucose consumption was calculated **P < 0.01 versus others.

affecting GLUT1 expression in mammalian cells It

should be noted that other glucose transporter

iso-forms, such as GLUT2, GLUT3, and GLUT4, were

not detectable in 3T3-L1 cells regardless of the

pres-ence of DIF-1 (data not shown)

Effect of DIF-1 on the enzymes involved in glucose metabolism in 3T3-L1 fibroblasts

To further investigate the mechanism of the action of DIF-1, we examined the effects of DIF-1 on the enzy-matic activities of glucose 6-phosphate dehydrogenase and three glycolytic enzymes, such as hexokinase, fluc-tose 6-phosphate kinase, and pyruvate kinase How-ever, incubation of the cells with 20 lm DIF-1 did not affect the total activities of the enzymes (data not shown) We then examined the direct effects of DIF-1

on the enzyme activities in vitro, but the activities were scarcely affected by DIF-1 (data not shown)

Effects of DIF-1 and insulin on the glucose uptake and GLUT translocation in 3T3-L1 adipocytes 3T3-L1 adipocytes are well-established model cells suitable for the analysis of the actions of insulin and other insulinomimetic substances in adipocytes Thus,

we examined and compared the effects of DIF-1 and insulin on glucose uptake in 3T3-L1 adipocytes As shown in Fig 9A, DIF-1, as well as insulin, promo-ted the glucose uptake by two to three-fold However, while the insulin-promoted glucose uptake and the phosphorylation (activation) of Akt were almost com-pletely suppressed by the addition of wortmannin, the DIF-1-promoted glucose uptake was only slightly suppressed by the inhibitor These results indicate that DIF-1 promotes the glucose uptake of 3T3-L1 adipocytes, at least in part, via a PI3K⁄ Akt-independ-ent pathway

We finally examined the effects of DIF-1 and insulin

on the translocation of GLUT1 and GLUT4 in 3T3-L1 adipocytes (Fig 9B) As expected, insulin induced both GLUT1 and GLUT4 translocation from intracel-lular stores to the plasma membrane but, in contrast, DIF-1 induced only GLUT1 translocation without affecting the total amount of GLUT1 The total amounts of GLUT4 were reduced after stimulation with DIF-1 or insulin (Fig 9B) for unknown reasons, while the other species of glucose transporters, GLUT2 and GLUT3, were not detectable (data not shown) Our results suggest that DIF-1 may promote glucose uptake, at least in main part, via GLUT1 transloca-tion

Discussion

A novel function of DIF-1 in mammalian cells DIF-1 was originally identified as a stalk-cell differen-tiation-inducing factor in the cellular slime mold

A ** P < 0.01 versus others

** P < 0.01 versus others

C

0

1

2

3

**

B

0

1

2

3

**

*

*

*

** P < 0.01 versus *

0

1

2

3

g 7

**

Fig 5 Effects of DIF analogs and some reagents on glucose

conc-umption in 3T3-L1 fibroblasts Confluent 3T3-L1cells were

incuba-ted for 10–14 h with 0.2% EtOH (vehicle), 20 l M of the indicated

DIF derivatives, 20 n M Tg, 0.2 l M A23187, 0.4% dimethylsulfoxide

(vehicle), or 0.4 m M 8-MIBMX The glucose concentration of each

well was measured, and the approximate rate of glucose

con-sumption was calculated as a ratio of control (EtOH or

dimethyl-sulfoxide) The mean ± SD (bars) values of three independent

experiments (n ¼ 3) are presented DMPH,

1-(2,6-dihydroxy-4-methoxyphenyl)hexan-1-one; DMSO, dimethylsulfoxide.

D discoideum [1], and DIF-1 and its analogs were

found to possess antiproliferative activity, occasionally

inducing cell differentiation in mammalian cells

[8–19] In the present study, we have found for the

first time that DIF-1 induces the translocation of

GLUT1 from intracellular stores to the plasma

mem-brane, at least in part, via a PI3K⁄ Akt-independent

pathway and thereby promotes glucose uptake in

3T3-L1 fibroblasts (Fig 8) and 3T3-L1 adipocytes

(Fig 9) Furthermore, we have shown here that DIF-1

possesses the most potent activity among its analogs

tested (Fig 5) Because DIF-3 and its derivatives,

rather than DIF-1, exhibit stronger antitumor activities

[12,16,17,19], the present results suggest that the

mech-anism of the action of DIF-1 for stimulating glucose

uptake is different from that for suppressing tumor cell

growth If this is the case, we may be able to

develop some DIF-1 derivatives that possess strong

glucose uptake-promoting activity but little

anti-tumor activity DIF-1 and such derivatives may have

therapeutic potential in the treatment of obesity and

diabetes

Mechanism of the action of DIF-1

As described already, since DIF-1 activates PI3K and

Akt in K562 cells [13] and human gastric cancer AGS

cells [15], we first expected that DIF-1 may have the

insulinomimetic function of promoting glucose

meta-bolism However, DIF-1 activated Akt slightly, if at

all, in 3T3-L1 fibroblasts (Fig 8C), adipocytes

(Fig 9A), and RGM-1 cells (data not shown)

More-over, because DIF-1 promoted the glucose uptake in

the cells even in the presence of wortmannin or

LY294002 (PI3K inhibitors) (Figs 8 and 9A), we

concluded that DIF-1 promotes glucose uptake, at least in part, via a PI3K⁄ Akt-independent pathway It should also be noted that DIF-1 did not affect the

0.9 1 1.1 1.2 1.3 1.4 1.5 DIF-1

0 200 400 600 800

Time of monitoring (s)

0.8 1 1.2 1.4 1.6 1.8

Tg

*

0.9 1 1.1 1.2 1.3 THPH

0.9 1 1.1 1.2 1.3 1.4 1.5

Tg

0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 DIF-3

0 200 400 600 800

n.s.

n.s.

Fig 6 Effects of DIF analogs on [Ca 2+ ]cin 3T3-L1 fibroblasts (A) Fura-2-loaded 3T3-L1 cells were stimulated with 20 l M of DIF-1, DIF-3, or THPH, or 20 n M Tg, and the changes in fluorescence ratio (R340 ⁄ 380) were monitored The representative traces of three independent cells are presented in each graph (B) The intensities of the maximal fluorescence ratio (R340 ⁄ 380) induced by the reagents are presented as the ratio of the basal fluorescence ratio The mean ± SD (bars) values of five independent cells (n ¼ 5) are presented *P < 0.001 n.s.; not significant.

A

B

0 1 2

**

Syntaxin4

-+

DIF-1:

0 1 2

Glut1-C

Fig 7 Effect of DIF-1 on GLUT1 translocation in 3T3-L1 fibroblasts (A) Crude membrane fractions of confluent 3T3-L1 cells were applied to sucrose density-gradient centrifugation, and the PM and LDM thus obtained were analyzed by western blotting for syntax-in4 (a marker for PM) (B) Confluent 3T3-L1 cells were incubated for 4 h with 0.2% EtOH (–) or 20 l M DIF-1 (+), and crude mem-brane fractions of the cells were applied to sucrose density-gradi-ent cdensity-gradi-entrifugation PM and LDM were then analyzed by western blotting for GLUT1, and the relative amounts of GLUT1 in the blot are shown (C) The relative amounts of GLUT1 present in PM were assessed after stimulation with 0.2% EtOH or 20 l M DIF-1, as des-cribed in (B) The mean ± SD (bars) values of three independent (n ¼ 3) experiments are presented **P < 0.01.

activities of four key enzymes involved in the glucose

metabolism (data not shown)

Most importantly, in the present study, we have

shown that DIF-1 induces the translocation of GLUT1

in 3T3-L1 fibroblasts (Fig 7) and adipocytes (but not

that of GLUT4) (Fig 9B), without affecting the

amounts of GLUT1 (Figs 8C and 9B) GLUTs2–4 in

3T3-L1 fibroblasts and GLUTs2–3 in adipocytes were

undetectable in western bolts (data not shown) In

addition, since the rates of DIF-1-promoted glucose

uptake (Figs 3 and 9A) correlate well with the amounts of GLUT1 in the plasma membranes (Figs 7 and 9B), the major glucose-uptaking activity of DIF-1 should be attributed to GLUT1 translocated to the plasma membrane in the cells, although we cannot exclude the possibility of the involvement of some GLUT isoform(s) other than GLUTs1–4

Glucose uptake and GLUT1 expression are enhanced in cancer cells and especially in malignant cells [27–30], indicating that GLUT activity should be associated with cell proliferation However, in regard

to some colorectal adenocarcinomas, the highest level

of GLUT1 was found in the most slowly proliferating cell line and the induction of cell cycle arrest increased both GLUT1 expression and glucose consumption [31] Although a simple comparison should be avoided and further research in this area is awaited, the action

of DIF-1 is different from that in the above cases in that DIF-1 never increased the expression levels of GLUTs as far as we are aware

It has been shown that DIF-1 increases [Ca2+]c in mammalian cells [9–11,20] and that DIF-1 is a direct inhibitor of PDE1 [19] In addition, increases in [Ca2+]c have been shown to stimulate glucose uptake via GLUT1 in avian erythrocytes [32], Swiss 3T3 fibro-blasts [33], and rat epithelial cells [34] However, neither Tg, nor A23187 ([Ca2+]c-increasing agents), nor 8-MIBMX (PDE1 inhibitor) promoted glucose consumption in 3T3-L1 fibroblasts (Fig 5A) Further-more, we have shown that the [Ca2+]c-increasing activ-ities of DIF analogs (Fig 6) hardly correlate with the glucose consumption-promoting effects of the analogs (Fig 5A) but do correlate well with the antileukemic effects of the analogs [12] These results strongly sug-gest that DIF-1 induces GLUT1 translocation and thereby promotes glucose consumption (uptake) in a way other than by increasing [Ca2+]c or inhibiting PDE1

The natural coenzyme a-lipoic acid, a small lipophi-lic substance that may be similar to DIF-1, can also promote glucose uptake in 3T3-L1 adipocytes [22,35] However, a-lipoic acid stimulates glucose uptake, at least in part, via PI3K⁄ Akt-dependent GLUT4 trans-location [22], which is different from the actions of DIF-1 in several aspects (Figs 7–9)

Finally, it is noteworthy that DIF-1 did in fact promote glucose consumption in all the mammalian cell lines tested to date Therefore, although the pre-cise mechanism of DIF-1-stimulated GLUT1 trans-location and glucose uptake is unclear at present, DIF-1 and its derivatives appear to be good tools for the study of glucose metabolism and cell biology in general

A

C B

0

1

2

3

4

wort

-+

-+

+ +

*

**

wort

-+

-+

+ +

0

1

2

3

4

**

**

-+

-+

+ +

DIF-1 LY

0 1 2 3

4

**

**

Akt pAkt

-wort

-+

-+

+ +

GLUT1

-Fig 8 Effects of DIF-1 and PI3K inhibitors on glucose consumption

in 3T3-L1 fibroblasts (A) Confluent 3T3-L1 cells were incubated for

8 h without or with 20 l M DIF-1 and ⁄ or 0.1 l M wortmannin (wort)

(left panel) or 30 l M LY294002 (LY) (right panel) All the incubation

media contained 0.2% EtOH (left panel) or both 0.3%

dimethylsulf-oxide and 0.1% EtOH as vehicles (right panel) The glucose

con-centration of each well was measured, and the approximate rate of

glucose consumption was calculated as a ratio for the control All

samples were prepared in triplicate, and the mean ± SD (bars)

val-ues of the triplicate (n ¼ 3) are presented **P < 0.001 (B)

Conflu-ent 3T3-L1cells were incubated for 4 h without or with 20 l M DIF-1

and ⁄ or 0.1 l M wortmannin (wort) (all the incubation media

con-tained 0.2% EtOH as vehicle), and the 2-deoxy-glucose uptake was

assessed as described in Experimental procedures Values are the

means ± SD (bars) of three independent experiments (n ¼ 3) in

which four-well determination was performed for each sample.

*P < 0.01 **P < 0.05 (C) Confluent 3T3-L1cells were incubated

for 6 h without or with 20 l M DIF-1 and ⁄ or 0.1 l M wortmannin

(wort) (all the incubation media contained 0.2% EtOH as vehicle),

and the cell proteins were analyzed by western blotting for Akt,

phospho-Akt (p-Akt), and GLUT1 Representative results of several

similar experiments are presented Note that wortmannin can

inhi-bit the phosphorylation of Akt at least for 9 h of incubation (data

not shown).

Experimental procedures

Reagents

DIF-1, DIF-3, and their derivatives were synthesized as

previously described [17] and stored at)20 C as 5–10 mm

solutions in EtOH Tg and A23187 were obtained from

Wako Pure Chemical Industries, Ltd (Osaka, Japan) and

stored at)20 C as 10 lm and 0.1 mm solutions in EtOH,

respectively 8-MIBMX was obtained from Calbiochem

(Darmstadt, Germany) and was stored at 4C as

25–100 mm solutions in dimethylsulfoxide LY294002 was

purchased from the Sigma-Aldrich Corp (St Louis, MO,

USA) and stored at 4C as a 10 mm solution in

dimethyl-sulfoxide Sheep anti-syntaxin4 serum, a generous gift from

J E Pessin (Stony Brook University, NY, USA), was used

for western blotting (1 : 1000 dilution) Rabbit

anti-(GLUT1 serum) (FabGennix, Inc., Shreveport, LA, USA)

and rabbit anti-(GLUT4 serum) were used as described

previously [36] Rabbit anti-(Akt serum) and

anti-(phospho-Akt serum) (Ser473) were obtained from New England

BioLabs (Beverly, MA, USA), and horseradish peroxidase-conjugated anti-(rabbit serum) and anti-(mouse serum) were obtained from Amersham (Little Chalfont, UK) The anti-bodies were used for western blotting according to the manufacturer’s instructions

Cells

Mouse 3T3-L1 fibroblast cells, kindly provided by O Ezaki (National Institute of Health and Nutrition, Tokyo, Japan), differentiated 3T3-L1 adipocytes, and rat gastric mucosal RGM-1 cells [20] were used in this study

3T3-L1 fibroblast cells were maintained in vitro at 37C (5% CO2) in DMEM containing a high concentration (4500 mgÆL)1) of glucose (DMEM-HG) (Sigma, D5796) supplemented with 75 lgÆmL)1penicillin, 50 lgÆmL)1 strep-tomycin and 10% (v⁄ v) fetal bovine serum RGM-1 cells were maintained in vitro at 37C (5% CO2) in DMEM containing a low concentration (1000 mgÆL)1) of glucose (DMEM-LG) (Sigma, D6046) supplemented with the anti-biotics, 10% fetal bovine serum and 10 mm Hepes-NaOH

B

**

**

EtOH DIF wort wort

+DIF Ins wort +Ins

0 1 2 3 4

DMEM (LG, -Serum) BA

DIF ± wort or vehicle (EtOH)

Ins ± wort

3T3-L1

0 1 2 3 4 4.5 (h)

EtOH

adipocytes

Assay for 2-deoxy-Glu uptake &

Western blot

A

GLUT4 GLUT1

0 1 2 3

**

n.s.

**

C: Control D: DIF-1

I :insulin

Cell lysates

GLUT1 GLUT4

Akt p-Akt

Fig 9 Effects of DIF-1 and insulin on glucose uptake (A) and the translocation of GLUT1 and GLUT4 (B) in 3T3-L1 adipocytes (A) As des-cribed schematically, adipocytes were stimulated for 4 h with 0.2% EtOH (vehicle), 10 l M DIF-1 and ⁄ or 0.1 l M wortmannin (wort) in DMEM-LG (serum-free) and then for 30 min in Buffer A (BA) containing the same reagents Alternatively, 3T3-L1 adipocytes were stimulated for 30 min with 100 n M insulin (Ins) and ⁄ or 0.1 l M wortmannin (wort) in BA All samples were then assayed for 2-deoxy-glucose uptake Note that the maximal effect was attained with 10 l M of DIF-1 under the conditions probably because the active concentration of DIF-1 was high due to the absence of calf serum in the incubation media (calf serum reduces the effects of DIF-1; data not shown) The values in the graph are the mean ± SD (bars) values of three independent experiments (n ¼ 3) **P < 0.0001 The adipocyte cell proteins stimulated with the reagents were analyzed by western blotting for Akt and phospho-Akt (p-Akt) (B) 3T3-L1 adipocytes were incubated for 4.5 h with 0.2% EtOH (C) or 10 l M DIF-1 (D) or for 30 min with 100 n M insulin (I), and the crude membrane fractions of the cells were applied to sucrose density-gradient centrifugation The PM and LDM thus obtained were analyzed by western blotting for GLUT1 and GLUT4 (representative blots are shown; upper panel) The relative amounts of GLUT1 and GLUT4 present in PM were assessed, and the mean ± SD (bars) values

of three independent experiments (n ¼ 3) are shown **P < 0.01 The total cell lysates of the three samples were also analyzed by western blotting for GLUT1 and GLUT4.

(pH 7.4) Unless otherwise mentioned, cultures were fed

every 2–3 days during growth and every 2 days after

con-fluence

3T3-L1 fibroblasts (preadipocytes) were differentiated

into adipocytes essentially as described by Student et al

[37] Briefly, 2 days after confluence, the medium was

removed and replaced with fresh DMEM-HG containing

0.5 mm 3-isobutyl-1-methylxanhine, 1 lm dexamethasone,

and 1.7 mm insulin Forty-eight hours later, the medium

was withdrawn and replaced with fresh DMEM-HG

con-taining 1.7 mm insulin After 48 h, insulin was withdrawn

from the culture medium and cells were maintained in

DMEM-HG until use

Measurement of the glucose concentration in the

incubation media of 3T3-L1 fibroblasts, a simple

assay system for glucose consumption

For the assay for glucose consumption, 3T3-L1 cells were

transferred into 12-well plates, each containing 1 mL of

DMEM-HG, and incubated for 5–6 days until they became

a confluent state with exchanging DMEM-HG every

2–3 days The cells were then preincubated for 2–3 days

with 1 mL of DMEM-LG with exchanging the medium

every day The confluent cells thus obtained were treated

for the appropriate number of hours with 1 mL of

DMEM-LG in the presence or absence of EtOH (usually

0.2%; vehicle), DIF-1 (5–20 lm), DIF analogs (20 lm), or

some reagents Aliquots of the incubation media were then

assayed for glucose concentration using a blood glucose test

meter and sensor chips for it (Sanwa Chemical Institute,

Nagoya, Japan), and the approximate rate of glucose

con-sumption was calculated It should be noted that the

glu-cose consumption by 3T3-L1 fibroblasts was assessed in

DMEM-LG because DMEM-LG rather than DMEM-HG

was suitable for the precise measurement of the glucose

concentration by the use of a blood glucose test meter

Therefore, to standardize the experimental conditions, we

used the DMEM-LG culture system in other experiments

(i.e confluent 3T3-L1 cells were preincubated for 2–3 days

in DMEM-LG and used to assess the actions of DIF-1 and

other reagents in DMEM-LG unless otherwise mentioned)

RGM-1 cells were also transferred into 12-well plates,

each containing 1 mL of DMEM-LG, and incubated for

several days until they reached a confluent state with

exchanging DMEM-LG every 2–3 days The confluent cells

were used for the assay for glucose consumption as

des-cribed above

It should be noted here that the simple assessment for

the medium glucose concentration (Fig 3A,B) can provide

only approximate rates of glucose consumption because the

measurement of the glucose concentration by the test meter

and the sensor chips is bound to have some degree of error

(data not shown) and there is a lag for the glucose uptake

promoted by DIF-1 to reach a maximum rate (Fig 3C)

Thus, to obtain reliable and comparable data for the rates

of glucose consumption, we usually incubated the cells with the drugs for more than 8 h

Assay for cell number

Confluent 3T3-L1 or RGM-1 cells in 12-well plates were treated for the appropriate number of hours with 1 mL of DMEM-LG in the presence or absence of EtOH (vehicle)

or DIF-1 After the glucose concentration was measured, the incubation media were discarded, and the cells were incubated with 1 mL of a fresh DMEM-LG containing 5% (v⁄ v) of Alamar blue (a cell number indicator; Wako Pure Chemical Industries, Osaka, Japan) until the color changed The relative cell number was assessed as described previ-ously [12,19]

Assay for glucose uptake

Confluent 3T3-L1 fibroblasts in a 12-well plate were incu-bated for 4 h in DMEM-LG with or without reagents Cells were then washed twice with Buffer A [Krebs-Ringer Hepes (25 mm), pH 7.4 containing 0.4% BSA and 3 mm sodium pyruvate] and incubated for 30 min in the same buffer with the reagents Then, 2-[1,2-3H]deoxy-d-glucose (PerkinElmer, Fremont, CA, USA) (0.8 lCiÆwell)1) was added to the cells to the final concentration of 0.1 mm and incubated for 20 min at 37C The nonspecific uptake was measured in the presence of 1 lm cytochalasin B At the end of incubation, cells were washed twice with ice-cold Buffer A and lysed in 0.4 mL of 0.4% SDS The radioactiv-ity in the lysate was counted with a scintillation counter The net values for glucose uptake stimulated with the rea-gents were obtained by subtracting the nonspecific uptake

in the presence of cytochalasin B

As shown in Fig 9A, 3T3-L1 adipocytes in a 12-well plate were incubated for 4 h in serum-free DMEM-LG with

or without the reagents (cells were incubated under serum-free conditions to establish a basal condition for the stimu-lation with insulin) Cells were then washed twice with Buffer A and incubated for 30 min in the same buffer con-taining the reagents They were assayed for glucose uptake

as described above, except for the incubation time (10 min) with 2-[1,2-3H]deoxy-d-glucose

Subcellular membrane fractionation

Confluent 3T3-L1 fibroblasts in 10-cm dishes were stimula-ted for 4 h in 10 mL of DMEM-LG with 0.2% EtOH (vehicle) or 20 lm DIF-1 Differentiated 3T3-L1 adipocytes

in 10-cm dishes were stimulated for 4 h in serum-free DMEM-LG with 0.2% EtOH or 10 lm DIF-1 and then for

30 min in Buffer A with the same reagents, or 3T3-L1 adi-pocytes were incubated for 4 h in DMEM-LG (serum-free)