JOURNAL

OF SURGICAL

52,625-630

RESEARCH

(19%)

Intracellular Cyclic AMP Levels in Endothelial Subjected to Cyclic Strain in Vitro TOSHIAKI

IBA, M.D., IRA MILLS,

Department Presented

at the Annual

Meeting

of Surgery,

Yale University

of the Association

PH.D., AND BAUER E. SUMPIO, School

for Academic

of Medicine, Surgery,

New Colorado

Haven,

Cells

M.D., PH.D.

Connecticut

Springs,

Colorado,

06510 November

20-23,199I

that human saphenous vein EC exposed to a similar cyclic strain regimen expressed and secreted significant amounts of tissue plasminogen activator (tPA) but not plasminogen activator inhibitor-l. Since other investigators have reported an inverse relation between intracellular levels of cyclic AMP (CAMP) and tPA production [15], the aim of this study was to determine intracellular CAMP levels in EC subjected to a cyclic strain regimen which has been shown to enhance EC production of tPA as well as activate adenylate cyclase activity. The results presented herein demonstrate the general lack of increase in CAMP levels despite the previously reported increase in adenylate cyclase activity even in the presence of a phosphodiesterase inhibitor. Furthermore, pretreatment of the cells with forskolin, or galanin (an inhibitor of adenylate cyclase) could not abrogate the rise in tPA induced by cyclic strain.

Human saphenous vein endothelial cells (EC) were grown to confluence in fibronectin-coated culture plates with flexible membrane bottoms and maintained in M-199 supplemented with substrates. One hour prior to experimentation 5 mM IBMX, a phosphodiesterase inhibitor, was added. Vacuum was used to deform the membrane bottoms to 24% strain at 60 cycles/ min (0.5 set elongation alternating with 0.5 set relaxation). After lo-60 min of cyclic strain, or upon exposure of EC to 100 p&f forskolin or 0.1 PM galanin, intracellular cyclic AMP (CAMP) was measured by radioimmunoassay. In parallel experiments, tissue plasminogen activator (tPA) secretion was determined after 24 hr of cyclic strain in the absence or presence of forskolin or galanin. The results demonstrate that exposure of EC to cyclic strain led to no change in CAMP levels and confirmed our previous observation that tPA secretion was enhanced with cyclic strain. Addition of forskolin, which led to an almost lo-fold increase in CAMP levels, or galanin, which led to a 34% decrease in CAMP levels, did not significantly alter the rise in tPA induced by cyclic strain. o 1992 Academic PWS, I~C.

METHODS Endothelial Cell Culture

INTRODUCTION Recent evidence from a number of laboratories, including ours, suggeststhat external forces, such as shear stress or cyclic strain, can modulate the phenotype of endothelial cells (EC) in culture, manifested by alterations in morphology, proliferation rate, and secretion of molecules [l-11]. However, the molecular basis and intracellular signals governing these changes in cell function are complex and not well elucidated [12, 131. Letsou et al. [ 141, reported that subjecting bovine aortic EC to 60 cycles/min, 24% maximum strain resulted in a progressive increase in the basal membrane adenylate cyclase activity which peaked between 3 and 5 min after the initiation of the strain regimen then declined to control values in 7-10 min. Stimulation of the stretched membrane preparation with forskolin resulted in a parallel increase in adenylate cyclase activity which was maximal at 5 min. Recently, Iba et al. [5-81 demonstrated 625

EC from saphenous vein were isolated [ 161and grown at 37°C on fibronectin (Biochemical Products, Bedford, MA)-coated flasks in a 5% CO, incubator. The culture medium consisted of M-199 (Biomedical Technologies, Inc., Stoughton, MA) supplemented with 17% fetal bovine serum (Hyclone Laboratories Inc., Logan, UT), 68 mM L-glutamine (Sigma Chemical Co., St. Louis, MO), 18 UPS units/ml grade 1 heparin (Sigma), 37 FANhypoxanthine (Sigma), and 0.25 mg/ml endothelial mitogen (Biomedical Technologies, Inc.). EC displayed typical “cobblestone” morphology, density-dependent inhibition of growth, and positive uptake of acetylated Di-ILDL (Biomedical Technologies, Inc.). The cells were subcultured with 1.0% trypsin and used from passages2 to 6. Apparatus The stress unit (Flexercel, McKeesport, PA) has been previously described and characterized in detail [l, 17201. It consists of a vacuum unit regulated by a solenoid valve and a computer program. Cells are cultured on 0022-4804/92

$4.00

Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

626

JOURNAL

OF

SURGICAL

RESEARCH:

flexible-bottomed culture plates with a hydrophilic surface. When a precise vacuum level is applied to the system, the culture plate bottoms are deformed to a known percentage elongation, which is translated to the cultured cells. When the vacuum is released, the plate bottoms return to their original conformation. Thus, the magnitude, duration, and frequency of the applied force can be varied in this system. Force analysis of the strain on the flex plate during stretch at various vacuum levels has been calculated mathematically by finite element analysis [19] and empirically by measuring with a micrometer the distance between concentric circles (radial strain) or diameter axes (axial strain) marked on the membrane. Very little change is observed in the latter, hence, the force on the attached cells is predominantly uniaxial.

Experimental

Protocol

Saphenous vein EC were grown to confluence (x 75,000 cells/cm2) in fibronectin-coated culture plates with flexible membrane bottoms. One hour prior to experimentation, 5 mM 3-isobutyl-1-methyl-xanthine (IBMX), a phosphodiesterase inhibitor, was added. Vacuum was used to deform the membrane bottoms to 24% strain at 60 cycles/min (0.5 set elongation alternating with 0.5 set relaxation). In some experiments, 100 PM forskolin (Sigma) or 0.1 &f galanin (Sigma) was added to the culture media for 30 min. The cell number of an aliquot of scraped cells was determined by Coulter counter (Model ZM, Coulter Electronics, Inc., Hialeah, FL). Since there is a heterogeneity of strain across the membrane, we constructed a polycarbonate “fence” (Fig. 1) to selectively seed cells in either the area of high strain (peripheral zone) or an area of low strain (central zone) based on previous data on the strain gradient across the membrane bottom [19]. The surface area of the peripheral zone is approximately twice that of the inner central zone. When a vacuum of -20 kPa is applied, cells grown in the peripheral zone experience strain of 7-24%. Cells grown in the inner circumference experience ~7% strain when -20 kPa of vacuum is applied, with the majority receiving ~1% strain. This enables us to subject cells to a narrow range of strain and to minimize contribution of fluid mixing, convective forces, shear stress, or agitation in producing a specific response, since both low and high strain zones are exposed to these forces. In additional experiments, EC were selectively seeded to confluence in either the peripheral or central zone. When the cells were confluent, the fence was removed and the fibronectin-coated plates were placed on the strain unit. Membrane bottoms were subjected to -20 kPa vacuum at 60 cycles/min.

VOL.

52, NO.

Cyclic AMP

6, JUNE

1992

Radioimmunoassay

After EC underwent the cyclic strain regimens described above, the regimen was automatically terminated at the predetermined cycle by the computer and the reactions stopped immediately by washing the cells twice in ice-cold PBS then scraped from the wells in the presence of 500 ~1 of 5 mM EDTA. Ice-cold ethanol was added to the cell suspension (65% final volume), and the solution was vortexed, then centrifuged at 2000g for 15 min at 4°C. The supernatant was concentrated by a Speed-Vat for 3 hr and redissolved in 500 ~1 of assay buffer from the CAMP assay kit (Amersham, Arlington Heights, IL). The radioimmunoassay for CAMP was conducted following the standard protocol in the Amersham kit and the results expressed as fmoles of cAMP/lO’ cells. Determination

of tPA levels in the medium

After 24 hr of cyclic strain, media were collected and tPA levels determined as previously outlined [5,6]. The media was first acidified by adding one tenth of the total volume of 1 M sodium acetate buffer. The pH of the media was adjusted to 4.0 then saved at -80°C. All assays were performed within 2 weeks after the collection of samples. Total tPA (free and bound) in culture media from each well was measured in triplicate by double antibody enzyme-linked immunosorbent assay (ELISA) (Imubind tPA, American Diagnostica Inc., New York, NY) utilizing a highly specific goat anti-human tPA. Statistical

Analysis

The data are reported as means * standard deviation. Quantitative analysis was performed by comparing group mean values. Analysis of variance with NewmanKeuls or Dunnet’s multiple comparison test was employed to assess the significance of differences between groups. A P value co.05 was considered significant. RESULTS

Figure 2 illustrates intracellular CAMP levels in human saphenous vein EC attached to flexible membranes, subjected to 60 cycles/min, 24% maximum strain. The average value for six experiments is shown (n = 3-6 wells/experiment). Although there was a decrease in CAMP levels at 10 min after the initiation of cyclic strain, this was not significant. There was no statistical change in the mean CAMP levels after 10, 20, 40, or 60 min of cyclic strain compared to the stationary control (0 min). At earlier time points (1,3, or 5 min) and up to 24 hr of cyclic strain, there was no detectable change in CAMP levels (110 +_ 32 fmole/105 cells, n = two experiments, 4 wells/experiment). EC utilized in these studies

IBA,

MILLS,

AND

SUMPIO:

CAMP

LEVELS

IN

STRETCHED

EC

627

FIG. 1. A polycarbonate “fence” was constructed in order to selectively seed cells in an area of high strain (the outer periphery of the membrane) or an area of low strain (in the center of the membrane). The fence was designed based on previous data on the strain gradient across the membrane bottom [19]. When a vacuum of -20 kPa is applied cells in the periphery experienced a strain of 7-24%, while cells grown in the inner zone experience ~7% strain when -20 kPa of vacuum is applied, with the majority receiving no strain. The fence is shown in profile at the top. The bottom portion shows a 6-well culture plate with the flexible membrane bottom. EC were grown to confluence in either the inner, low strain region (top three wells) or the outer, high strain area (bottom three wells); the fence was removed, and the EC were stained with crystal

were capable of producing CAMP, since Fig. 3 shows that they responded to the addition of 100 pM forskolin (n = 9 wells) with a significant increase in CAMP levels. Furthermore, addition of 0.1 pM galanin (n = 9 wells), an adenylate cyclase inhibitor, reduced intracellular CAMP levels by 30%. The cyclic strain experiments were also performed utilizing bovine aortic EC (unpublished ob300 1

servations) and these studies showed exactly the same pattern as above, without any change in CAMP levels. Since the flexible membranes utilized in these studies produce a heterogenous strain [ 191, a polycarbonate “fence” (Fig. 1) was designed to selectively seed EC in the periphery (r( = 6 wells) or center of the membranes (n = 6 wells). Figure 3 demonstrates that after 15 or 30 min of cyclic strain, EC from the high strain zone (periphery) and the low strain zone (center) had similar CAMP levels and were not different from the stationary control (0 min). The relationship of tPA production in response to cyclic strain with intracellular CAMP levels was also explored (n = four experiments, 3-6 wells/experiment). Figure 4 shows that the rise in tPA levels seen in EC subjected to 24 hr of cyclic strain was unaffected by incubating the cells with either 100 PM forskolin or 0.1 JLM galanin. DISCUSSION

01

0

10

20 Time

30 of stretch

40

50

so

(mid

FIG. 2. Time course of intracellular CAMP levels in EC subjected to cyclic strain. Results are means + standard deviation of six experiments (n = 3-6 wells/experiment).

While the importance of hemodynamic factors in modulating the proliferation and production of secretory products by cells of the vessel wall has been intuitively recognized, it is only recently that experimental models which apply physical forces to cells in culture have been

628

JOURNAL

OF

SURGICAL

RESEARCH:

center periphery

-

Time of stretch (mid

-

FIG. 3. Cyclic AMP production by EC in the presence of forskolin, galanin, or cyclic strain. 100 pM forskolin, a direct activator of adenylate cyclase, or 0.1 pM galanin, an inhibitor of adenylate cyclase [24], was added to confluent monolayers of EC for 30 min to determine their effects on CAMP production. EC were also seeded in the periphery (dark bars) and center (clear bars) of the plates (see Fig. 1) and after 15 or 30 min of cyclic strain at 60 cycles/min, CAMP levels were determined. The results are means + SD. *P -c 0.01, tP < 0.05 compared to control (0 min).

utilized. Much of our previous knowledge of EC biology has come from studies of cultured cells grown and maintained in a strictly defined stationary environment in vitro, which may be suboptimal or inappropriate for the study of their behavior in viva Our laboratory has characterized a device which can apply different regimens of cyclic tensional deformation on attached monolayers of cultured EC. This repetitive force in vitro may be analogous to the pulsatile distension experienced by cells of the vessel wall in vivo. Using this apparatus, we have demonstrated that chronic cyclic strain can modulate vascular cell phenotype in culture, manifested by alterations in growth [l, 81, morphology [2,7], and secretory capacity of prostacyclin, tPA, and endothelin [3-6, 81. However, the pathways by which EC membranes sense changes in pressure or flow are a seminal problem awaiting to be resolved [ll, 12, 211. Adenylate cyclase is a well-characterized membranebound enzyme, with ATP-binding sites on the cytoplasmic face of the plasma membrane and associated with guanine nucleotide regulatory proteins (G proteins), [22, 231. The latter are involved in coupling the diverse cellular receptors to the adenylate cyclase catalyst or to PIspecific phospholipase C. Several polypeptide hormones exert their effects through activation or inactivation of membrane-bound adenylate cyclase, which in turn regulates intracellular protein phosphorylation. Forskolin bypasses the receptors and G-proteins and is thought to act by directly activating adenylate cyclase or facilitating the activation of adenylate cyclase by the G-proteins. Galanin is a ubiquitous neuropeptide which has been shown to inhibit adenylate cyclase through a path-

VOL.

52, NO.

6, JUNE

1992

way involving a pertussis toxin-sensitive inhibitory Gprotein [ 241. Somjen and co-workers [25], using an orthodontic device glued to the bottom of the culture dish to apply constant tension on osteoblasts, observed stimulation of cell division. The cells responded to the continuous deformation by increasing prostaglandin Ez synthesis to a maximum in 20 min followed by an increase in cyclic AMP release. Both products decreased to baseline with time. These investigators also noted a rise in adenylate cyclase activity within 5 min of tension. Reich and colleagues [26] have reported that cyclic AMP levels were elevated fourfold in human umbilical vein EC subjected to flow with a shear rate of 430 set-’ for 15 min. In our laboratory we have noted that subjecting EC to cyclic strain resulted in a progressive increase in the basal- and forskolin-stimulated activity of adenylate cyclase which peaked between 3 and 5 min [14]. The results of the present study, however, indicate that CAMP production by EC subjected to the same cyclic strain regimen is unchanged. Why this occurs in the presence of adenylate cyclase activation is unclear. While the possibility exists that a significant portion of intracellular CAMP has been compartmentalized in a cellular fraction that is not available to the antibody, we were unable to document any changes in CAMP levels in the media or a crude membrane fraction (unpublishedobservations). It is possible that the lack of rise in CAMP despite activation of adenylate cyclase may reflect either a rapid entry of Ca2+ with stretch which is capable of activating phosphodiesterases [27], or to a temporal loss of activity of the exogenously added phosphodiesterase inhibitor, IBMX. With regard to the latter we have noted that in the absence of IBMX, there was a decline in CAMP levels to near undetectable levels under basal conditions and with cyclic strain at all time points tested. Furthermore, our find-

0

control

B

forskolin galanin

*

1:

*

: center

periphery

FIG. 4. Tissue plasminogen activator (tPA) production by EC in response to cyclic strain, in the presence or absence of forskolin or galanin. EC were seeded in the center or periphery of the membranes and subjected to 24 hr of cyclic strain at 60 cycles/min. Forskolin (100 GM) or 0.1 pM galanin was added to the culture media in some experiments prior to cyclic strain. The results are means f SD of four experiments, n = 3-6 wells/experiment. *P < 0.05 compared to center.

IBA,

MILLS,

AND

SUMPIO:

CAMP

ings are consistent with the findings by Wirtz and Dobbs [28] who found no alteration in CAMP levels in lung epithelial cells subjected to 17% stretch despite an early rise in intracellular calcium. In contrast to the studies with osteoblasts or EC, Mills et al. [ 291, have shown that smooth muscle cells subjected to a similar regimen of cyclic strain had a decrease in adenylate cyclase activity with stretch. They also noted that varying the degree of strain inversely affected adenylate cyclase activity. The mechanism of increased synthesis and secretion of tPA by EC in response to cyclic strain is undetermined. Activation of the phosphoinositide pathway may play a major role since elevated intracellular calcium levels are required for the acute release of tPA [30] and the production of diacylglycerol and protein kinase C regulates the synthesis of tPA [ 15,301. Recent data from our laboratory have demonstrated that the phosphoinositide pathway is also activated by cyclic strain. EC subjected to 60 cycles/min, 24% strain had an enhanced phosphatidylinositol turnover with the production of inositol triphosphate and diacylglycerol [31] and activation of protein kinase C [31] within seconds of the initiation of cyclic strain. The adenylate cyclase/cAMP system is also an important regulator of tPA release by EC. However, in the present study, incubation of EC with forskolin at a dose which results in a significant increase in CAMP levels (Fig. 3), failed to abolish the increase in tPA production induced by cyclic strain. Likewise, addition of an inhibitor of adenylate cyclase activity, galanin, at a dose which reduced CAMP levels but did not affect tPA secretion. Based on these findings, we hypothesize that CAMP does not play an important modulatory role in the production of tPA by EC. Further studies will help elucidate the transduction pathway by which an external force, such as cyclic strain, results in the increased tPA production demonstrated in this study.

LEVELS

2.

3.

4.

5.

6.

7.

Sumpio, B. E., Banes, A. J., Levin, L. G., and Johnson, G. Mechanical stress stimulates aortic endothelial cells to proliferate. J. Vase. Surg. 6: 252, 1987. Sumpio, B. E., Banes, A. J., Buckley, M., and Johnson, G. Alterations in aortic endothelial cell morphology and cytoskeletal protein synthesis during cyclic tensional deformation. J. Vast. Surg. 7: 130, 1988. Sumpio, B. E., Banes, A. J. Prostacyclin synthetic activity in cultured aortic endothelial cells undergoing cyclic mechanical deformation. Surgery 104: 383, 1988. Sumpio, B. E., Widmann, M. D. Enhanced production of an endothelium-derived contracting factor by endothelial cells subjected to pulsatile stretch. Surgery 108: 277, 1990. Iba, T., Shin, T., Sonoda, T., Rosales, 0. R., and Sumpio, B. E. Stimulation of endothelial secretion of tissue-type plasminogen activator by repetitive stretch. J. Surg. Znuest. 50: 457, 1991. Iba, T., and Sumpio, B. E. Tissue plasminogen activator expression in endothelial cells exposed to cyclic strain in vitro. Cell Transplant. 1: 43, 1992. Iba, T., and Sumpio, B. E. Morphologic evaluation of human

STRETCHED

629

EC

endothelial cells subjected uusc. Res. 42: 245, 1991.

to repetitive

stretch

in vitro.

Micro-

8.

Iba, T., Maitz, S., Furbert, T., Rosales, O., Widmann, M., Spillane, R., Shin, T., Sonoda, T., and Sumpio, B. E. Effect of cyclic stretch on endothelial cells from different vascular beds. Circ. Shock 35: 193,1991.

9.

Davies, P. F., Dewey, C. F., and hemodynamics forces on vascular Znnuest. 73: 1121, 1984.

10.

11.

Bussolari, endothelial

S. R. Influence of function. J. Clin.

Dewey, C. F., Bussolari, S. R., Gimbrone, M. A., and Davies, P. F. The dynamic response of vascular endothelial cells to fluid shear stress. J. Biochem. Eng. 103: 177, 1981. Diamond, S. L., Eskin, S. G., and McIntire, L. V. Fluid flow stimulates tissue plasminogen activator secretion by cultured human endothelial cells. Science 243: 1483, 1989.

12.

Davies, P. F. How do vascular NIPS 4: 22, 1989.

13.

Sumpio, B. E. Hemodynamic forces and the biology of the endothelium: Signal transduction pathways in endothelial cells subjected to physical forces in uz’tro. J. Vast. Surg. 13: 744, 1991. Letsou, G. E., Rosales, 0. R., Maitz, S., Vogt, A., and Sumpio, B. E. Stimulation of adenylate cyclase activity in cultured endothelial cells subjected to cyclic stretch. J. Curdiouasc. Surg. 31: 634, 1990. Francis, R. B., and Neely, S. Inhibition of endothelial cell secretion of tissue-type plasminogen activator and its rapid inhibitor by agents which increase intracellular cyclic AMP. Biochim. Biophys. Acta 1012: 207,1989. Watkins, M. T., Sharefkin, J. B., Zajtchuk, R., Maciag, T. M., D’Amore, P. A., Ryan, U. S., Van Wart, H., and Rich, N. M. Adult human saphenous vein endothelial cells: Assessment of their reproductive capacity for use in endothelial seeding of vascular prostheses. J. Surg. Res. 36: 58%596,1984. Buckley, M. J., Banes, A. J., Levin, L. G., and Sumpio, B. E. Osteoblasts increase their rate of division and align in response to cyclic mechanical tension in vitro. J. Bone Miner. Res. 4: 225, 1988.

14.

15.

16.

17.

18.

19.

REFERENCES 1.

IN

20.

21. 22. 23. 24.

25.

endothelial

cells respond

to flow?

Banes, A. J., Link, G. W., Gilbert, J. W., Son Tay, R. T., and Monbureau, 0. Culturing cells in a mechanically active environment. Am. Biotech. Lab. May, pp. 13-22, 1990. Gilbert, J. A., Banes, A. J., Link, G. W., and Jones, G. J. Surface strain of living cells in a mechanically active, in vitro environment. In D. E. Dietrich (Ed.), In Proceedings, ANSZS Conference Houston, PA: Swanson Analysis Systems. pp. 13.2-13.7, 1989. Gilbert, J. A., Banes, A. J., Link, G. W., and Jones, G. L. Video analysis of membrane strain: An application in cell stretching. Exp. Tech. 5: 43, 1990. Lansman, J. B. Going with the flow. Nature 331: 481, 1988. Gilman, A. G. G Proteins and regulation of adenylyl JAMA 262: 1819,198S. Levitski, A. From epinephrine to cyclic AMP. Science 1988.

cyclase.

241: 800,

Amiranoff, B., Lorinet, A., Lagny-Pourmir, I., and Laburthe, M. Mechanism of galanin-inhibited insulin release. Eur. J. Biothem. 177:147,1988. Somjen, D., Binderman, I., Berger, E., and Harell, A. Bone remodeling induced by physical stress is prostaglandin E, mediated. Biochim. Biophys. Acta 637: 91, 1980.

26.

Reich, K. M., Gay, C. V., and Frangos, J. A. Fluid shear a mediator of osteoblast cyclic adenosine monophosphate tion. J. Cell. Physiol. 143: 100, 1990.

27.

Rasmussen, 314: 1094,

28.

Wirtz,

H. The 1986.

calcium

H. R. W., and

Dobbs,

messenger

system.

L. G. Calcium

stress as produc-

N. Engl. J. Med. mobilization

and

630

JOURNAL exocytosis after one mechanical Science 250: 1266,199O.

29.

30.

stretch

OF

SURGICAL

of lung

epithelial

RESEARCH:

VOL.

cells.

Mills, I., Letsou, G., Rabban, J., Sumpio, B., and Gewirtz, H. Mechanosensitive adenylate cyclase activity in coronary vascular smooth muscle cells. Biochem. Biophys. Res. Commun. 171(l): 143, 1990. Santell, L., and Levin, E. G. Cyclic AMP potentiates phorbol ester stimulation of tissue plasminogen activator release and in-

52, NO.

hibits secretion man endothelial

6, JUNE

1992

of plasminogen cells. J. Biol.

activator Chem. 263:

inhibitor-l from 16802, 1963.

hu-

31.

Rosales, O., and Sumpio, B. E. Pulsatile stretch enhances phosphatidylinositol turnover and 1,2 diacylglycerol formation in bovine aortic endothelial cells. Am. J. Physiol. 262: C956, 1992.

32.

Rosales, 0. R., and Sumpio, B. E. Protein kinase C is a major mediator of the adaptation of vascular cells to hemodynamic forces. Surgery, 112: 459-466, 1992.

Intracellular cyclic AMP levels in endothelial cells subjected to cyclic strain in vitro.

Human saphenous vein endothelial cells (EC) were grown to confluence in fibronectin-coated culture plates with flexible membrane bottoms and maintaine...
1MB Sizes 0 Downloads 0 Views