High glucose and hyperosmolarity increase platelet-derived growth factor mRNA levels in cultured human vascular endothelial cells

Vol.

187,

September

No.

2, 1992

16,

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS Pages

1992

HIGH

GLUCOSE

AND HYPEROSMOLARITY

PLATELET-DERIVED CULTURED

HUMAN

GROWTH VASCULAR

FACTOR

664-669

INCREASE mRNA

ENDOTHELIAL

LEVELS

IN

CELLS

Masakazu Mizutani, Yukichi Okuda, Takashi Yamaoka, Kenichiro Tsukahara, Masaaki Isaka, Chieko Bannai and Kamejiro Yamashita Division of Endocrinology and Metabolism, Institute of Clinical Medicine, University of Tsukuba, l-l-l Tennoudai, Tsukuba, Ibaraki, 305 Japan Received

July

15,

1992

SUMMARY: We investigatedthe effects of high glucoseandhyperosmolality on platelet-derivedgrowth factor (PDGF) production andPDGF-B chain mRNA levels in cultured humanumbilical vein endothelial cells. Under an excessof ambient glucose (13.8 and 27.5mM) and a hyperosmolar condition (22.0mOsm/L with mannitol), PDGF concentrations in the culture medium were both significantly increased(10.3f 6.0%, 36.2 f7.2%, 48.5* 9.0% increaserespectively compared with 5.5mM glucose). Parallel to protein secretionlevels, PDGF-B chain mRNA levels showeda significant increase(57.7%, 103.7%, 210.8% increase),while no change of /3 -actin mRNA levels was observed. Thus, elevated PDGF released from endothelium by high glucosemay play an important role in the pathogenesis of diabetic angiopathy. o 1~2 Academic press,I~~.

Hyperglycemiais believedto be the majorcauseof diabeticvascularcomplications (1, 2). However, it is unclear by which mechanismhyperglycemia alters the metabolism in vascular cells and their functions. Endothelial cells regulate the exchange of materials and information between the blood and arterial wall. For example, it hasbeenreportedthat exposureof endothelialcellsto high glucosecan reducecell proliferation ability andprostacyclin production, aswell asincreasethe synthesis of an extracellular matrix, such as fibronectin, laminin, and type IV collagen (3). Cultured endothelial cells are also known to produce somegrowth factors which migrateandproliferate vascularsmoothmusclecells. The platelet-derived growth factor (PDGF) is an ubiquitous growth factor regulating many biological processesin cells of mesenchymalorigin, including proliferation and chemotaxis. PDGF is composedof two peptides,A and B, that are products of two distinct genesand are regulated independently (4). PDGF is active only in the dimeric form, and it is synthesizedand secretedby various other cell types in addition to endothelialcells(5) includingmacrophages(6) and smooth

Vol.

187,

No.

2, 1992

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

muscle cells (7). Although smooth muscle cells secrete the AA-dimer exclusively, endothelial cells secrete a mixture of PDGF-BB, -AA, and -AB (8). It is suggested that altered endothelium may cause an increase in permeability for plasma components and produce growth factors such as PDGF.

In order to

investigate the pathogenesis of diabetic vascular complications, we have examined the effect of high glucose and hyperosmolarity levels in human umbilical vein endothelial cells.

MATERIALS

on PDGF and PDGF-B

mRNA

AND METHODS

Cell cultHuman umbilical vein endothelial cells for the primary culture were obtained from fresh human umbilical cords using established methods (9). Cell cultures were maintained in a MCDB107 medium with 15% fetal calf serum (FCS), HEPES (7.15mg/ml), endothelial mitogen (50 fi g/ml), heparin (5 Y g/ml), penicillin (50 unit/ml) and streptomycin (50~ g/ml) in an atmosphere of 95%air and 5%CO2 at 37°C. Confluence was obtained at 25cm’ flask. At confluence, cells were passed using 0.25% trypsin in 0.02% EDTA. Media were changed twice weekly. All studies were performed at the 5-10th passage. v of PDGF oroduced bv endotheltal cells. Confluent 4 35mm mates of endothelial cells were washed twice in a DhosDhate buffer saline and then 2ml each of the above mentioned medium was added, and subsequently the glucose (G) concentration (G5.5mM, 13.8mM and 27.5mM) was adjusted. The hyperosmotic condition was adjusted by mannitol (Man) (G5.5mM+Man8.3mM, G5.5mM+Man22.0mM). After 24 hours incubation, the conditioned media were collected and stored at -70°C until the PDGF assay. The stored media were concentrated by 10 times by vacuum centrifugation before the assay. This assay of PDGF was performed by radioimmunoassay ( [‘251]Platelet-derived growth factor assay system, Amersham, England ). . . ti 1OOmm plates of RNA lsolatmn and Northern blot analysis. Confluent endothelial cells were exposed to the above mentioned conditions for 48 hours. Cells were lysed in the guanidium solution. The total RNA was isolated from the cell lysate by centrifuging it through cesium chloride. After quantification by measuring absorbance at 260 nm, 10 p g of the RNA sample (the same amount per lane for each gel ) was electrophoresed in 1% agarose gel containing formaldehyde. RNA was transferred to the nylon membrane by PosiBlo? (Stratagene, USA), and fixed by baking. The filters were hybridized with [ 32P]dCTP-labeled probes specific for the c-sis gene [ 1300 base pair Pst I fragment of Simian Sarcoma Virus Clone (lo), a gift of the Japan Cancer Research Resources Bank, Tokyo, Japan]. Hybridization was carried out in 10% dextran, 5x saline sodium citrate (5x SSC), (0.75M NaCl, 0.075M sodium citrate and 50% formamide) at 42°C for 24 hours. The final washing of the blot was in 0.1x SSC, 0.1% SDS at 60°C. The membrane was exposed to Kodak XR autoradiography film at -70°C with an intensifying screen for 2-7 days. The relative concentration of PDGF-B mRNA compared with that of the fi-actin was calculated by densitometry of the radioautograph and expressed as the percentage of control. m . . Values were expressed as mean? SEM. Statistical analysis was performed using the Duncan’s multiple range test.

RESULTS PDGF levels in the culture medium secreted by endothelial cells were measured by RIA (Fig.lA, B). PDGF levels secreted by endothelial cells cultured with high glucose (G13.8mM, 27.5mM) were significantly higher (~~0.05, pcO.Ol), 665

Vol.

187,

No.

BIOCHEMICAL

2, 1992

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

respectively than that of the control (G5SmM). In a hyperosmolar condition made by the mannitol addition, PDGF levels were significantly increased. Compared with the same osmotic pressure, marmitol increased PDGF levels higher than those of glucose (Fig.lA).

When lO”M phorbol 1Zmyristate 13-acetate (PMA) or 10-Q

phorbol 1Zmyristate 13-acetate 4-o-methyl ether (PMAM) (4, 11) was added to the control medium, PDGF levels increased significantly (peO.05, p
To determine the cross-reaction

of bovine PDGF in FCS, a medium

without culture was assayed at the same time. However, cross-reactivity was negligible in this assay (data not shown). Parallel to the PDGF secretion, the mRNA glucose- dependent increase (Fig.2A).

levels of PDGF-B

Hyperosmolarity

(22.0mOsm/L

also increased PDGF-B mRNA levels. When phorbol ester (PMAM

with FCS showed

a

with Man) 10-h)

was

added to the control medium, mRNA levels of PDGF-B were also significantly increased (Fig.2B). densitometric intensity

The relative levels of PDGF-B mRNA were determined by a

scanning of each hybridization

of individual

glucose-dependent

fl-actin

band. After normalization

bands, mRNA

levels of PDGF-B

increase (57.7%: G13.8mM;

103.7%: G27.5mM

by the

showed

a

increase

compared with G5.5mM) (Fig.3).

DISCUSSION As PDGF is a potent smooth muscle cell mitogen, it has been hypothesized that PDGF is responsible for intimal proliferation in atherosclerosis (12). The source of PDGF was presumed to be external to the vessel wall, namely, in the form of blood borne platelets.

P 2

F 25 i! 5

40-

However,

in hypertension

or hyperlipidemia,

60

A

-

B

smooth muscle

*

60

20.

40

2 2

20 0.

Glucose Mannitol

(mM) (ItIM)

5.5 0

13.6 0

27.5 0

0 5.5 6.3

5.5 22.0

Control

PMA

PMAM

F i p: 1. Effect of high glucose, mannitol and phorbol ester on PDGF production in human umbrhcai vein endothelial cells. Cells were exposed to A: glucose(G) SSmM, G 13.8mM, G 27SmM, G 5SmM + Mannitol (Man) 8.3mM, or G 5SmM t Man 22.0mM. B: 10-w phorbol 12-myristate 13-acetate (PMA) or 10% phorbol 12-myristate 13-acetate 4-o-metyl ether (PMAM) for 24 hours. Values are mean*SEM, n=8 (A),and n=5 (B) per group. * p
Vol.

187,

No.

2, 1992

BIOCHEMICAL

4

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

PDGF-B

PMAM

0

-

PDGF-B

-

f3 -Actin

l@

(M)

Fig .2. Northern blot analysis of PDGF-B mRNA levels. A 1300 base Pst I fragment of Simian Sarcoma Virus Clone was used as a probe. Total RNA was isolated from the endothelial cells exposed to A: Glucose(G) 5SmM, 13.8mM, 27SmM, or G 5.5mM t Mannitol22.0mM, B: 5.5mM glucose with or without PMAM (10-Q) for 48 hours. The amount of RNA analyzed was 10 Y g per lane.

and platelet aggregation(13). Subsequentstudiesrevealed that various cell types, including endothelialcells (5),

proliferation

occurs without

endothelial

denudation

macrophages(6), and even arterial smoothmusclecells (7), can producePDGF, at leastin vitro. Sincethe abovementionedcell typesarepresentin humanatherosclerotic

5.5

13.8

0

0

Mannitol(mM) PMAM

(hl)

-

-

27.5 0 -

5.5

5.5

22.0

0

-

1o-6

Fig.3. Relative PDGF-B mRNA level normalized by p -actin mRNA level. The density of each band of radioautograph was measured by densitometry, and the ratio of PDGF-B/ /3 -actin was calculated and expressed as a percentage of the control. 667

Vol.

187,

No.

2, 1992

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

plaques, the PDGF hypothesis of atherogenesis has been modified to include the possible production of PDGF from the developing human intima (14). Indeed, the presence of local PDGF in the atheroma is supported by recent studies showing that PDGF-B chain mRNA has been detected by Northern blots of human carotid plaques removed at surgery (7). Our study demonstrated that in cultured human endothelial cells high glucose increased the mRNA levels of PDGF-B and PDGF secretions. Another study showed that in cultured bovine aortic endothelial cells PDGF levels were reduced by high glucose (15). However, in that study PDGF was measured not in the medium, which we did, but in the cytosol. Furthermore, that study used anti-human PDGF monoclonal antibodies for the bovine PDGF assay. Recently, it has been reported that PDGF expression

can be regulated by

physical factors, such as shear stress (16) or hypoxia (17), as well as thrombin (18, 19), endotoxin (20), phorbol myristate acetate (PMA) (4, 1 l), and the transforming growth factor-fi(TGF-fl)

(19). Our study demonstrated that not only glucose but

also mamritol increased PDGF mRNA levels. In the same osmotic pressure, marmitol increases PDGF mRNA levels significantly higher than those of glucose. Thus, hyperosmolarity itself can induce PDGF mRNA transcription. It has been reported that shear stress leads to enhance the inositol phospholipid turnover in endothelial cells which, in turn, activates protein kinase C (21).

Our study demonstrated that

phorbol ester stimulated both the expression of PDGF mRNA and release of PDGF. Since phorbol ester directly activates protein kinase C, it is proposed that PDGF expression by hyperosmolarity

might be regulated through the activation of protein

kinase C. On the other hand, another study has reported that high glucose induces activation of protein kinase C (22). Therefore, high glucose might induce PDGF production through a signal transduction mechanism involving protein kinase C activation.

In diabetic ketoacidosis

or non-ketotic hyperosmolar

coma, plasma

osmolarity is extremely high (23). Therefore, the pathophysiological significance of our results might be considered to contribute somewhat to the cause of diabetic angiopathy. Recently, it has been demonstrated that cultured microvascular endothelial cells can bind and respond to PDGF (24). Thus, PDGF molecules synthesized by vascular endothelial cells seem to function in an autocrine manner. Therefore, our study suggests that hyperglycemia and hyperosmolarity may cause diabetic angiopathy by the autocrine mechanism of PDGF secretion. Whether post-transcriptional effects are also contributing to the enhanced PDGF release or whether the PDGF A-chain is also expressed and / or regulated will require further studies.

ACKNOWLEDGMENTS The authors wish to thank Ms. Yoko Yokokura

for her technical assistance.

Support was provided by the University of Tsukuba Research Project Fund and the

Vol.

187,

No.

2, 1992

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

Scientific Research Fund from the Ministry of Education, Science and Culture of Japan.

REFERENCES 1. Yorek, M.A. and Winegrad, AI. (1989) Metabolism, 38, 16-22. Brownlee, M. and Cerami, A. (1984) Ann. Intern. Med. 101,527-537. i: Cagliero, E. and Lorenzi, M. (1991) Diabetes 40,102-110. Starksen, N.F. and Williams, L.T. (1987) J. Biol. Chem. 262, 14381-14384. :: DiCorleto, P.E. and Bowen-Pope, D.F. (1983) Prc. Natl. Acad. Sci. USA 80, 1919-1923. 6. Shimokado, K. and Ross R. (1985) Cell 43,277-286. 7. Barrett, T.B. and Benditt, E.P. (1987) Proc. Natl. Acad. Sci. USA 84, 1099-1103. Bowen-Pope, D.F. and Seifert, R.A. (1989) J. Biol. Chem. 264,2502-2508. t* Jaffe, E.A. and Minick, C.R. (1973) J. Clin. Invest. 52,2745-2756. 10: Robbins, K.C. and Aaronson, S.A. (1981) Prc. Natl. Acad. Sci. USA 78, 2918-2922. 11. Fox, P.L. and DiCorleto, P.E. (1984) J. Cell. Physiol. 121, 298-308. 12. Ross, R. (1981) Arteriosclerosis 1, 293-311. 13. Owens, G.K. and Schwartz, S.M. (1981) Proc. Natl. Acad. Sci. USA 78, 7759-7763. 14. Ross, R. (1986) N.Engl.J.Med. 314, 488-500. 15. Umeda, F. and Tatematsu, A. (1991) Horm. Metab. Res. 23, 274-277. 16. Hsieh, H. and Frangos, J.A. (1991) Am. J. Physiol. 260 (Heart Circ. Physiol. 29) H642-H646. Kourembanas, S. and Faller, D.V. (1990) J.Clin.Invest. 86, 670-674. ii- Daniel, T.O. and Williams, L.T. (1986) J. Biol. Chem. 261(21), 9579-9582. 19: Daniel, T.O. and Williams, L.T. (1987) J. Biol. Chem. 262 (25) 1189311896. 20. Albelda, S.M. and Kern, J.A. (1989) Am. J. Physiol. 257 (Lung Cell.Mol. Physiol. l), L65-L70. 21. Nollert, M.U. and McInire, L.V. (1990) Biochem. Biophys. Res. Commun. 170, 281-287. 33 King, G.L. et af. (1990) Problems in Methodology and clinical Aspects, Karger, Basel, 86. 23. Louis Vegnati, A. etal. (1987) Joslin’sDiabetesMellitus 12th edition, 526-552. 24. Bietz, J.G. and Frackelton, A.R.Jr. (1991), Proc. Natl. Acad. Sci. USA 88, 202112025. ai.

669