JOURNAL OF CELLULAR PHYSIOLOGY 1423-38 (1990)

Growth-Inhibiting Effect of Tumor Necrosis Factor on Human Umbilical Vein Endothelial Cells I s Enhanced With Advancing Age In Vitro YOSHlYA SHIMADA,* KAZUHIKO KAJI,HlDEKl ITO, KOUlCHl NODA, AND MlTSUYOSHl MATSUO Tokyo Metropolitan In5titute of Gerontology (Y.S., K.K., K.N., M.M.), Tokyo Metropolitan Geriatric Hospital (ti./.),Sakae-cho, /tabashi-ku, Tokyo 173, japan We have examined the effects of in vitro aging on the growth capacity of human umbilical vein endothelial cells (HUVECs) under the influence of tumor necrosis factor (TNF) with or without interferon-y (IFN-y).The growth and colony-forming abilities of control cells were impaired with advancing age in vitro, especially at later stages (more than 70-80% of life span completed). It was found that treatment with TNF inhibited growth and colony-forming efficiency at any in vitro age. The effects of TNF were shown to increase with increasing in vitro age, as reflected by a more pronounced increase in doubling times, a decrease in saturation density, and a reduction in colony-forming efficiency. However, the characteristics of TNF receptors, including the dissociation constant, and the number of TNF-binding sites per cell-surface area remained rather constant. The effect of TNF was augmented by IFN-y at a dose that alone affected growth and colony formation only slightly. The augmentation by IFN-y was also found to depend on in vitro age; the synergy with TNF in the deterioration of colony-forming ability was observed only in "aged" cells. These results suggest that the intrinsic responsiveness of HUVECs to growth-inhibiting factors, as well as to growth-stimulating factors, changes during aging in vitro.

The endothelium forms the luminal surface of the vascular system and plays a n important role in maintaining homeostasis of the vessel wall and organs. In addition, alterations in the function and repair ability of the endothelium play key roles in atherogenesis. Therefore, knowledge of the factors regulating the proliferation and functions of endothelial cells (ECs) is essential for understanding the normal and abnormal states of blood vessels. Recent progress in this field has revealed that endothelial cell growth factor (ECGF) (Lobb and Fett, 1984) and fibroblast growth factor (FGF) (Gospodarwicz et al., 1978) are important for the proliferation of ECs. Using these factors along with heparin and gelatin- or fibronectin-coated dishes, the proliferative mode of ECs has been examined and serial cultivations have been achieved (Thornton et al., 1983; Nichols et al., 1987). Tumor necrosis factor (TNF), which is released from macrophages and was initially recognized a s a cytotoxic or cytostatic factor for some tumor cells, has been found to influence the functions of ECs profoundly. I t has been reported that TNF increases neutrophil adhesion to ECs, augments the production of coaggulant activity from ECs, stimulates ECs t o produce plateletderived growth factor (PDGF), interleukin 1 (IL-l), or granulocyte-macrophage colony-stimulating factor (GM-CSF), and inhibits the proliferation of ECs :C 1990 WILEY-LISS, INC

(Beutler and Cerami, 1986; H a j a r et al., 1987; Pober 1988; Sato et al., 1986). These effects of TNF on ECs suggest t h a t TNF is closely associated with inflammatory and immune responses, wound healing, and atherogenesi s. Levine and Mueller (1979) demonstrated that endothelial injury occurs more frequently and/or is repaired less efficiently with increasing age. Furthermore, deterioration in the capacity of wound healing and acceleration in the development of vascular diseases have been shown to be related to aging (Kligman et al., 1985; Danon e t al., 1989). It is likely, therefore, that the responsiveness of ECs to factors that regulate EC functions, such as ECGF and TNF, alters during aging, and that this age-related alteration results in dysfunction of blood vessels. We examined changes in the sensitivity of human umbilical vein endothelial cells (HUVECs) to the growth-inhibiting effects of TNF and/or IFN-y during aging in vitro, as a model for aging in vivo. Here, we Received August 31, 1988; accepted August 22, 1989.

"To whom reprint request'correspondenee should be addressed a t present address: Division of Physiology and Pathology, National Institute of Radiological Sciences, Anagawa, Chiha-shi 260, Japan.

32

SHIMADA ET AL

report a n unexpected increase in the responsiveness of HUVECs to a growth-inhibiting factor, TNF, during aging in vitro.

MATERIALS AND METHODS Recombinant human TNF was provided by Asahi Chemical Industry Co. Ltd. (Tokyo, Japan) and had a specific activity of 2.0 x lo6 Uimg when assayed by the L-M cell cytotoxicity assay in the absence of actinomycin D (Yamazaki et al., 1986).It is equivalent to 1-2 x 10' Uimg when assayed for cytotoxic activity on L 929 cells in the presence of actinomycin D (1 pgiml). Recombinant human interferon-gamma IIFN-?) was supplied by Toray Industries Inc. (Tokyo, Japan) and Daiichi Seiyaku Co. Ltd. (Tokyo, Japan) and had a specific activity of 4.0 x lo7 U/mg. Endothelial cell growth factor (ECGFj was prepared as described previously, using heparin affinity chromatography (Lobb and Fett, 1984). Briefly, newborn bovine brains were homogenized 4°C in 0.15 M ammonium sulfate and acidified to pH 4.5. The resulting homogenate was processed with ammonium sulfate, and the fraction precipitating between 20 and 25% ammonium sulfate was isolated by centrifugation. The pellet was redissolved in purified water and extensively dialyzed against 0.1 M sodium phosphate in a 6,000-8,000 molecular weight cutoff dialysis tube. The insoluble substance was removed by centrifugation (11,000 rpm, 30 min), and the supernatant was applied to a CM-Sephadex C 50 column (Pharmacia, Uppsala, Sweden). The fraction eluted by 0.5 M NaCl after 0.15 M NaCl in PBS- was dialyzed overnight against 10 mM Tris-HC1, pH 7.0, containing 0.6 M NaC1, and then applied directly to a heparin-Sepharose column (Pharmaciaj. The fraction eluted with 10 mM 'his-HC1, pH. 7.0, supplemented with 2.0 M NaCl was freeze-dried and stored a t -80°C. A nanogram amount of this protein mixture could stimulate the growth of HUVECs with the effect saturated at 40 ngiml in the presence of both heparin and EGF and at 100 ngiml in their absence. The saturation dose was independent of in vitro age, although the apparent growth-promoting ability was quite small in aged cells (data not shown). The ECGF purified by this method contains about 80% of anionic type (acidic FGF) and 20% of cationic type (basic FGF). This estimation was based on the fractional distribution of growth-stimulating activities of ECGF eluted from a heparin-Sepharose column with different concentrations of NaCl (0.6, 1.0, 1.5, 2.0 M) (data not shown). Fetal bovine serum (FBS) was purchased from Hyclone Sterile Systems Inc. (Logan, UT). Medium MCDB-104 was from GIBCO Laboratories (Grand Island, NY); epidermal growth factor (EGF) was from Collaborative Research Inc. (Lexington, MA); and heparin and gelatin were from Sigma Chemical Company (St. Louis, MO). Tissue culture plasticware (Falcon) was from Becton Dickinson Labware (Oxnard, CAI. Preparation of gelatin-coated dishes and wells A 0.1% gelatin solution was added to dishes or wells a t 37°C and allowed to stand at room temperature for 30 min. The gelatin solution was then removed by aspiration, and the dishes were washed with PBS- and dried. The gelatin-coated dishes and wells were used within 4 h r after preparation.

Tissue culture HUVECs were obtained by the modified method of Jaffe et al. (1973). Immediately after receipt of an umbilical cord in medium MCDB-104 a t 4"C, the vein was vigorously washed with PBS- for removal of coaggulated blood and filled a solution of 0.1% trypsin and 0.02% EDTA. The vein was then incubated at 37°C for 15 min and repeatedly washed out with a small amount of MCDB-104 (10 ml in total). The cell suspension obtained was centrifuged a t 1,000 rpm for 5 min and the supernatant removed. The pelleted cells were resuspended in MCDB-104 supplemented with 10% FBS, 100 ng/ml ECGF, 10 ng/ml EGF, and 100 Kgiml heparin (hereafter, referred to as the MCDB-104 system), and seeded on gelatin-coated 24-well plates. The cells were maintained a t 37°C in a humidified 5% C0,-95% air incubator, and passaged a t a 1:2 split ratio twice per week. The cells were identified as endothelial cells by their positive response to factor VIII and their uptake of acetylated-LDL. When examined 24 h r after seeding, the plating efficiency was 70-90% regardless of the population doubling levels (PDLs). For the experiments, two sublines of HUVECs were used, each of which could be passaged about 60 times a t a 1:2 split ratio. Cell growth assay Cells were seeded at a density of 1.0 x or 1.0 x lo4 cells/cm2 in gelatin-coated wells containing the MCDB-104 system. The cells were allowed to stand for 3-4 h r to attach to the substrata, and then TNF and/or IFN--y was added to the medium. The medium containing TNF and/or IFN-y was exchanged every 4 days. After a given period for growth, the number of viable cells and, if necessary, cell-size distributions were measured with a Coulter counter ZBI ICoulter Electronics Inc., Hialeah, FL). Colony formation Cells were seeded a t a density of 100 cells/cmz in the MCDB-104 system. Three hours later, TNF and/or IFN-y was added, and 3 days later, the medium with the cytokine(s) was replaced with fresh medium with them. The cells were maintained at 37°C for 1week. On day 7, the cells were fixed in 2.5% glutaraldehyde, rinsed with PBS-, and stained with a 3% Giemsa solution. The number of colonies and the number of cells in each colony were counted under a dissecting microscope. Here, a colony is defined as one composed of more than 16 cells. Colony size was designated as the number of doublings necessary for the initial single cell to reach the observed number (Smith et al., 1978). The colony-forming efficiency of HUVECs was rather low; even the efficiency of control cells was 15% and the percentage of nondividing cells was around 70. Thus, the size distribution of colonies composed of more than two cells was examined in order to see the difference more clearly in the colony-size distribution of cytokineuntreated and treated HUVECs. Binding of 12sII-labeled TNF Purified recombinant TNF was radioiodinated by the iodogen method (Aggarwal and Eessalu, 1987). Briefly,

33

EC RESPONSE TO TNF DURING CELLULAR AGING

Fig. 1 Light micrographs of HUVECs at (a) the 12th and (b) the 54th passage (about 20 and 95%of life span completed, respectively). Bar indicates 500 pm.

Serial passages

Serial passages (1:2 split ratio)

0.1 M phosphate buffer (pH 7.4) was added to a glass tube coated with 5 kg/ml of iodogen (Pierce Chemical Co., Rockford, ILI and incubated with 1 mCi of Na'"1 for 10 min a t 4°C. The mixture was then transferred to a tube containing TNF and allowed to react for 5 min a t 4°C. The reaction was stopped by removal of the soluble material by filtration on Sephadex G-25 equilibrated with 0.1 M phosphate buffer containing 0.1% gelatin. I1"1]TNF had a specific activity of 1,180 Ciimmol protein. Binding assays were carried out essentially according to the procedure of Aggarwal and Eessalu (_1987). Briefly, confluent cell monolayers (2.0-4.0 x 10" cells/ well) in 12-well plates were washed twice with 0.5 ml of MCDB-104 medium containing 0.5% BSA (binding buffer) and then incubated with about 200,000 cpm of ['"IITNF in 0.4 ml of the binding buffer containing various amounts of unlabeled TNF. These cultures were used to determine nonspecific binding (nonspecific binding was 20-30%1 of the total binding). The binding reaction was terminated by rinsing the cultures four times with l ml of ice-cold MCDB-104 medium containing 0.2% BSA. The cells were solubilized with 0.4 ml of 1 N NaOH and cell-bound radioactivity was determined in a gamma counter.

Pig. 2. Colony-forming efficiency of HUVECs at various passage levels. Colonies containing more than 16 cells were counted. Inset shows cell yieldlcm' plotted as a function of the number of 1:2 subcultivations on dishes 10 cm in diameter. The data for two lines of HUVECs are shown.

TNF and IFN-y inhibited HUVEC growth in a dosedependent manner. A slight, but significant, inhibitory effect of TNF on the growth of HUVECs was observed a t a concentration of 0.3 Uiml. The inhibitory effect of TNF reached a plateau level a t a concentration of around 30 Uiml. On the other hand, the dose-response curve to IFN-y was quite different from that of TNF. I t has a threshold concentration of around 100 Uiml (2.5 ngiml), and the growth rate of cells decreased rapidly with the increase in IFN-y concentration. Figure 4 shows the change in the response of HUVECs to TNF andior IFN-7 during aging in vitro. The early (13th)-passagecells displayed extensive proliferation with a doubling time (DT) of 20 h r and a saturation density (SD) of approximately 1.0 x lo5 cellsicm'. With TNF at a concentration of 500 Uiml, which is a saturating concentration for growth inhibition of HUVECs (Fig. 3), cell growth was delayed, although the DT and SD remained almost unchanged. With IFN-y at a concentration of 1,000 Uiml, the cell RESULTS growth was slightly delayed and the DT and SD reChanges in the response of HUVECs to TNF mained unchanged. Treatment with both cytokines, and/or IFN-y during aging in vitro however, dramatically changed the growth curve; the When serially cultivated, HUVECs could be subcul- DT was twice a s long as that of control cells and the SD tivated about 60 times at a split ratio of 1:2 and thus was decreased to about one-seventh that of control cells (Fig. 4a). Since the effect of IFN-y alone at the dose were estimated to have finite life spans about 50-55 PDL under our culture conditions. HUVECs a t a late used was quite small, these data suggest TNF and IFNpassage (54th passage) showed typical features of so- y act on HUVECs synergistically. As shown in Figure called senescent cells: an increase in both cellular and 4b, the growth response of control HUVECs (TNF or nuclear sizes and a decrease in colony size (Fig. 1). IFN-y untreated) to ECGF a t the 54th passage (about Growth and colony-forming efficiency declined progres- 95% of life span completed) was lowered significantly sively with advancing age in vitro (Fig. 2). The reduc- compared to early-passage cells; the DT was prolonged tion in growth rate observed in cells at higher PDL three times and the SD was decreased to one-tenth. It could not be overcome by a n excess amount of ECGF is worth noting that a single treatment with TNF significantly affected the DT and SD of late-passage cells (data not shown). Figure 3 shows the effects of various concentrations compared to TNF-untreated cells, whereas there was of TNF and IFN-y on the growth of HUVECs. Both almost no effect on early-passage cells. Compared to

SHIMADA ET AL

34 100

C--rr&

I

3200

-i

v

-

150-

.-E c

0

-n._c

100-

0

500.1

10

1

100

a

10000

1000

Concentration (Ulrnl)

t

Fig. 3. Effect of TNF (0)or IFN-y (e) on the proliferation of IIUVECs. HUVECs at 15-20 passages were seeded at 1.0 x 10‘ cellsi cm2 and received increasing doses of TNF or IFN-7 in the MCDB-104 system. Cells attaching to the substrata were counted on day 4. Cell growth was estimated on the basis of the ratio of harvested cells to seeded cells and calculated as the change in population doubling level IAPDL). Each value represents the mean 2 standard deviation of three independent experiments.

0

bl

10

20

30

40

50

Serial passages

Fig. 5. Age-dependent changes in (a)doubling time and (b) saturation density of HUVECs treated with vehicle (c),TNF at 500 U/ml (01, IFN-y at 1,000 U!ml [A)or TNF plus IFN-y ( 0 ) . The left points indicate the mean i standard deviation of the 13-16th passage cells.

k 0

. 2

.

.

4

6

n 8

. 1012

.

t 0

. 2

. 4

. 6

. 8

.

.

.

101214

T i m e (days)

Fig 4 Time course of the effects of TNF, IFN-y, or TNF plus IFN-y on the growth of HUVECs HUVECs at (a)the 13th and (b)the 54th passages (about 20 and 95%of life span completed, respeclwely) weie seeded at 1 0 x 10’ cellsicm’ No additive ( c ) ,TNF (500 U‘ml) (U), IFN-y (1,000 17iml) (A), or TNF plus IFN-7 (elwere added to each HUVEC culture Each value represents the mean of duplicate cultures

TNF, IFN-y at the dose used had a smaller effect on the DT and SD of late-passage cells. Combined treatment with TNF and IFN-y on late-passage cells caused complete inhibition of cell growth. Some HUVECs treated with both cytokines became detached from the substrata and the number of cells decreased progressively. Figure 5 summarizes the changes in DT and SD of ECs cultured with or without TNF and/or IFN-y a s a function of in vitro age. It can be seen that in control cells both the increase in DT and the decrease in SD occur with increasing in vitro age: especially, SD drastically decreases after the 40th passage (at the begin-

I ning of phase 1111, although it was constant before phase 111. Comparison of TNF-treated ECs with control cells reveals that a single treatment with TNF reduces the SD as early a s the 25th passage (middle phase I1 stage), but the effect on DT was significant only above the 40th passage (beginning of phase 111) and became highly prominent a t the 54th passage (late phase 111). Thus, SD may be more sensitive to the age-associated effects of TNF.

Influence of in vitro age on the colony-forming efficiency of HUVECs treated with TNF and/or IFN--y In order to see the change in the effect of TNF andlor IFN-y on proliferative behavior of “individual” HUVECs during aging in vitro, we examined the effect of TNF and/or IFN-y on the colony-forming efficiency of cells during various passages. As shown in Figure 2, the colony-forming efficiency of control HUVECs was clearly reduced in late-passage cells. The number of colonies containing more than 16 cells at the 52nd and 54th passages reduced to two-fifths and one-fifth, respectively, of that at the 12th passage. Figure 6 shows

z EC RESPONSE TO TNF DURING CELLULAR AGING

100

15

35

1

'1

C T

I T I

C

T

I

T

C T

I

I

I I

C T

1 :

1

Treatments

Fig. 6. The effect of TNF, IFN-7, and TNF plus IFN-y on the colonyforming efficiency of HUVECs. Cells at the 12th, 19th, 52nd, and 54th passages (about 20, 35, 90, and 95% of life span completed, respectively) were treated with TNF (500 Ulml), IFN-y (1,000 Uiml), or TNF plus IFN-y. C, T, I, and T + I stand for control (no additive), TNFtreated, TFN-y-treated, and TNF-plus-IFN-y-treated cultures, respectively. Cells were seeded at a density of 100 cells/cm2 and were fixed 1 week later. The number of colonies with more than 16 cells was counted. Each value represents the mean = standard deviation of four dishes. Significant difference !P < 0.01) from the value for control (") or for both control and TNF (*") by Student's t-test.

that TNF (500 U/ml) reduced colony-forming efficiency. This effect was also more clearly exerted in late-passage cells; by treatment with TNF, the number of colonies in early-passage cells was reduced to onehalf, whereas in late-passage cells, the number was reduced to one-fifth. IFN-y at a concentration of 1,000 Uiml decreased the colony-forming efficiency to a small extent. The effect of TNF and IFN-y together on earlypassage cells was similar t o that of TNF alone with respect to colony-forming efficiency, although the combined treatment retarded both growth rate and saturation density much more effectively than did TNF alone (Fig. 4a). In contrast, the combined treatment decreased the colony-forming efficiency of late-passage cells more than did treatment with TNF alone; the combination decreased the colony-forming efficiency of ECs at the 52nd and 54th passages to 35 and 23% of that of TNF alone, respectively. Therefore, in late-passage cells, the synergistic effect of TNF and IFN-y may be manifested partially via a reduction in the number of cells capable of colony formation. But such a n explanation does not apply to younger cells because their colony-forming efficiency after treatment with both cytokines was almost the same as that after treatment with TNF alone. The effect of TNF on the colony-size distribution of HUVECs at the 19th and 54th passages is shown in Figure 7. The colony-size distribution curves of latepassage cells shifted significantly downward. Thus, the growth ability decreased during aging in vitro. It was also found that the colony-size distribution curves of

Colony size, log cells/colony 2

Fig. 7. The colony-size distributiun curves of HUVECs. Cells t the 19th passage (about 35% of life span completed: squares) and a t the 54th passage (95% of life span completed: circles) were cultivated with (closed symbols) or without TNF (open symbols).

both early- and late-passage cells were shifted markedly downward by treatment with TNF. The TNF-induced decrement ratio of colonies containing more than four cells in late-passage cells is 0.37 (from 43.6 to 27.6%), whereas that of early-passage cells is 0.17 (from 85.2 to 70.4%). The decrement ratios of colonies containing more than eight cells in late- and earlypassage cells are 0.89 (from 23.5 to 2.7%) and 0.28 (from 76.6 to 54.9%), respectively. Therefore, the effect of TNF on the colony-size distribution of late-passage cells seems to be larger than that of early-passage cells. Binding of [lz5I1TNFto ECs at early and late passages Since the response of early- and late-passage it .seemed possible.t h.a t HUVECs t o TNF was. different, - . there is a n age-associated change in the characteristics of TNF receptors. Thus, we examined the time course of ['"IITNF binding t o the early- and late-passage cells (13 and 90% of life span completed, respectively). The results shown in Figure 8 demonstrate that maximal specific binding a t 4°C occurred after 6 h r in both earlyand late-passage cells and remained a t the same level for at least 18 hr. The late-passage cells specifically bound about twice more lignnd than did early-passage cells. To further investigate the ligand-binding characteristics in early- and late-passage cells, we examined specific binding by competition with unlabeled TNF (Fig. 9). Scatchard plot analysis of the binding data from two independent experiments indicated that the mean dissociation constant (Kd) was 0.58 nM for early-passage cells and 0.67 nM for late-passage cells. The mean numbers of receptors per cell were 5,800 and 8,600 for early- and late-passage cells, respectively, and the modal volumes were determined to be 1.8 5 0.1 and 2.8

36

SHIMAUA ET AL

-

8

N

k y

6

0

4

-g C

3

0

m

2

0

0

2

4

6

81012

24

0

2

4

6

81012

24

Time ( h o u r s ) Fig 8 The time course of binding of ['L51]TNFto (a)early- and (b) late-passage HUVECs (13 and 90%. of life span completed, respeclively'l, 0 5 n M L1"I]TNF was incubated with cells at 4°C with or without (13)0 5 phf (1,000-fold excess) unlabeled TNF Specific binding

0

0.125

0.5

2

8

32

T N F (nM)

Fig. 9. Effects of various concentrations of unlabeled TNF on specific binding of ["'IlTNF to HUVECs at early (0)and late (el-passages (13 and 95% of life span completed, respectively). The cells were incubated at 4°C for 8 h r with 0.4 ml binding medium containing 0.5 nM L'"1lTNF with or without the indicated concentrations of unlabeled TNF. Specific binding was determined as described for Figure 8. Inset shows Scatchard analysis of the specific hinding of TNF. 2 0.4 (pm', x lo-')). From these volumes, assuming a spherical shape, a n approximation of the mean surface area in pm2 was made. When the number of TNF receptors was adjusted for surface area, there is little difference between early- and late-passage cells with respect to receptor density.

DISCUSSION The results presented reveal the following: 1) the growth of HUVECs in the presence of ECGF declines during aging in vitro; 2) with TNF, growth is inhibited and colony-forming efficiency is lowered; 3) the sensi-

(El) per l o 5 cells was calculated from the difference between binding in the absence and presence of unlabeled TNF The data shown are mean values of duplicate determinants

tivity to TNF is enhanced with advancing age in vitro; 4)neither the Kd nor density of TNF receptors seems to be altered with aging; 5) the ability of IFN--y to augment the cytostatic effect of TNF also increases a s a function of in vitro age, especially after the stage when the growth of untreated control cells begins to deteriorate. These results suggest that the sensitivity of HUVECs t o growth-inhibiting factors, as well as to growth-stimulating factors, alters during aging in vitro. Relatively little information is available concerning the changes of cells to growth-inhibiting agents during aging in vitro and in vivo. Heparin and ionizing radiation have been reported to inhibit the growth of smooth muscle cells (MaCaffrey et al., 1988) and fibroblasts (Macieira-Coelho et al., 1978), respectively; cells in the late stages of their in vitro life span become more sensitive to these agents than cells in the early stages. IFN-(3 also inhibits the growth of fibroblasts, although with no change in sensitivity during their life span (Tamm et al., 1984; Tan et al., 1975). Together with our finding that the sensitivity of HUVECs to TNF increases with increasing in vitro age, it may be proposed that the responsiveness of cells to growth-inhibiting agents increases during aging in vitro. A mechanism to explain how the growth-inhibiting effect of TNF on HUVECs is enhanced with advancing age is unclear a t present, but some possible explanations may be offered. One is that the repair capacity of cells against TNF-induced damage decreases during aging. It has been proved that TNF induces DNA fragmentation in cells sensitive t o TNF and that IFN-y interacts with 7°F to enhance the extent of DNA fragmentation (Dealtry et al., 1987). TNF and IFN-y have also been reported to cause HUVECs to lose their stainable fibronectin matrix (Stolpen et al., 1986). Thus, the abilities of late-passage cells to repair DNA damage and/or t o reproduce the fibronectin matrix might be deteriorated. Indeed, some abilities of ECs have been demonstrated to decrease with increasing in vitro age. Response to some growth factors (Johnson and Longenecker, 1982) and the release of a n angiotensin I-converting enzyme (Noveral et al., 1987) decrease as the

37

EC RESPONSE TO TNF DURING CELLULAR AGING

passage of cultures progresses. Furthermore, in human diploid fibroblasts a t phase 111, DNA strands broken by X-irradiation are deficiently rejoined (Epstein et al., 1973)and their ability to repair UV-induced damage is lowered (Painter et al., 1973). In addition, the average amount of unscheduled DNA synthesis has been reported to decrease as cells approach phase I11 (Hart and Setlow, 19761, although a finding of no age-related decrease has also been reported (Hall et al., 1982). An alternative explanation is a n alteration in TNF receptor characteristics during aging. We found that the Kd’s were 0.58 nM and 0.68 nM and that the numbers of TNF receptors were 5,800 and 8,600 in early- and late-passage ECs, respectively. However, cell volume, and thus cell surface area, increases with advancing in vitro age (by approximately 1.6-fold in late-passage cells). Thus, there seems to be little difference in receptor number per surface area between early- and latepassage cells. This is in good agreement with a previous report that the number of EGF binding sites per surface area remains constant throughout aging in vitro (Phillips et al., 1983). Thus, it seems unlikely t h a t the differential response to TNF between early- and late-passage ECs is attributable to differential receptor characteristics. It is also possible that during subculturing we were selecting for a cell type that is more sensitive to TNF rather than that each cell became more sensitive to TNF with increasing age. Recent studies have found that TNF alters the phenotype of vascular ECs so as to promote coaggulation, inflammation, and immunity (Pober, 1988). Because we have demonstrated that late-passage ECs are more sensitive to the cytostatic action of TNF, i t would be important to show that cells a t high PDL are more sensitive with regard to other effects of TNF such as stimulation of ECs to produce coaggulant activity, IL-1 or PDGF, and induction of ECs to adhere to neutrophils. Recently, MaCaffrey et al. (1988) reported increased proliferation of arterial smooth muscle cells isolated from old rats. It has also been demonstrated that the adherence of monocytes to HUVECs in vitro is significantly higher in aged than in young cells (Molenaar et al., 1989). Taken together with our results, i t seems possible that the intrinsic character of vascular components (i.e., ECs, smooth muscle cells, and, maybe, macrophages) alters during aging and that this may account for the age-associated increase in atherogenesis or age-dependent impairment of the capacity for wound healing (Danon et al., 1989). We have found that during aging in vitro, the sensitivity of HUVECs to the growth-promoting effect of ECGF decreases while that to the growth-inhibitng effect of TNF increases. Since the function of ECs is regulated by numerous cytokines, including TNF, IL-1, and TGF-P, it is of great importance to understand the details of the age-related changes in the sensitivities of ECs to these cytokines.

ACKNOWLEDGMENTS We are very grateful to Miss Y. Miyata for technical assistance, Asahi Chemical Industry Co. Ltd. for supplying recombinant human TNF, and Toray Industries, Inc. and Daiichi Seiyaku Co. Ltd. for supplying recombinant human IFN-,.

LITERATURE CITED Agganval, B.B.? and Eessalu, T.E. (1987) Induction of receptors for tumor necrosis factor-cu by interferons is not a major mechanism for their synergistic cytotoxic response. J. Biol. Chem., 262: 10000-10007. Beutler, B., and Cerami, A. (1986) Cachectin and tumor necrosis factor as two sides of the same biological coin. Nature, 320r584588. Danon, U., Kowatch, M.A., and Roth, J . (1989) Promotion of wound repair in old mice by local injection of macrophages. Proc. Natl. Acad. Sci. U.S.A., 86t2018-2020. Dealtry, G.B., Naylor, M.S., Fiers, W., and Balkwill, R.F. (19871DNA fragmentation and cytotoxicity caused by tumor necrosis factor is enhanced by interferon-?. Eur. J. Immunol., 17:689-693. Epstein, J., Williams, J.R., and Little, J.B. (19731 Deficient DNA repair in human progeroid cells. Proc. Natl. Acad. Sci. U.S.A., 70: 977-981. Gospodarwicz, D., Brown, K.D., Birdwell, C.R., andzetter, B.R. (1978) Control of prolikeration or human vascular endothelial cclls. J . Cell Biol., 77t774-788. Hajjar, K.A., Hajjar, D.P., Silvestein, R.L., and Nachman, R.L. (19871 Tumor necrosis factor-mediated release of platelet-dcrivcd growth factor from cultured endothelial cells. J . Exp. Med., 166: 235-245. Hall, J.R., Almy, R.E., and Scherer, K.L. (1982) DNA repair in cultured human fibroblasts does not decline with donor age. Exp. Cell Res., 139r351-359. Hart, R.E.. and Setlow, R.B. (1976) DNA repair in late-nassare - ._ human cells. Mech. Agcing Dev., 5r67-77. Jaffe, E.A., Nachman, R.L., Becker, C.G., and Minick, C.R. (1973) Culture 01human endothelial cells derived from umbilical veins. Identification by morphologic and immunologic criteria. J. Clin. Invest., 522t2745-2756. Johnson, I,,, and Longenecker, J.P. (1982) Sensecence of aortic endothelial cells in vitro: Influence of culture conditions and preliminary characterization of the senescent phenotype. Mech. Ageing Dev., 18:l-18. Kligman, A.M., Grove, G.L., and Balin, A.K. (1985) Aging or human skin. In: Handbook of the Biology of Aging. C.E. Finch and E.L. Schneider, eds. Van Nostrand Reinhold Company, New York, pp. 820-84 1. Levine, E.M., and Mucller, S.N.(1979) Cellular senescence and vascular endothelial cells. Int. Rev. Cytol., (Suppl),IOr67-76. Lobb, R.R., and Fett, J.W. (1984) Purification of two distinct growth factors from bovine neural tissue by heparin affinity chromatography. Biochemistry, 23:6295-6299. MaCaffrey, T.A., Nicholson, A.C., Szabo, P.E., Weksler, M.E., and Weksler, B.B. (1988)Aging and arteriosclerosis. The increased proliferation of arterial smooth muscle cells isolated from old rats is associated with increased platelet-derived growth factor-like activity. J. Exp. Med., 167r163-174. Macieira-Coelho, A., Diatloff, C., Billard, M., Fertil, B., Malaise, E., and Fries, D. (1978) Effect of low dose rate irradiation on the division potential of cells in vitro. J . Cell. Physiol., 95t235-238. Molenaar, R., Visser, W.J., Verkerk, A,, Koster, J.F., and Jongkind, J.F. (1989) Peroxidative stress and in vitro ageing of endothelial cells increases the monocyte-endothelial cells adherence in a human in vitro system. Atherosclerosis, 76.193-202. Nichols, W.W., Buynak, E.B., Bradt, C., Hill, R., Aronson, M., Jarrell, B.E., Mueller, S.N., and Levine, E.M. (1987) Cytogenetic evaluation of human endothelial cell cultures. J. Cell. Physiol., 132:453462. Noveral, J.P., Mueller, S.N., and Levine, E.M. (1987) Release of angiotensin I-converting enzyme by endothelial cells in vitro. J. Cell. Physiol., 131:l-5. Painter, R.B., Clarkson, J.M., and Young, H.K. (1973) Ultraviolet induced repair replication in aging diploid human cells (W1-38).Radiat. Res., 56:5(j0-564. Phillips, P.D., Kuhnle, E., and Cristofalo, V.J. (1983) [IZ5IJEGF binding ability is stable throughout the replicative life-span of WI-38 cells. J. Cell. Physiol., 114:311-316. Pober, J.S. 11988) Cytokine-mediated activation of vascular endothehum. Am. J. Pathol., 233t426-433. Sato, N., Goto, T., Haranaka, K., Satomi, N.: Nariuchi, H., ManoHirano, Y., and Sawasaki, Y. (19861 Actions of tumor necrosis factor on cultured vascular endothelial cells: Morphologic modulation, growth inhibition, and cytotoxicity. JNCI, 76:1113-1121. Smith. J.R.. Pereira-Smith. O.M.. and Schneider. ~, E.L. (1978) ~,Colonv size distributions as a mtasure of in vivo and in vitro aging. Proc. Natl. Acad. Sci. U.S.A., 75t1353-1356. ~

~

~.I

38

SHIMADA ET AL

Stolpen, A.H., Guinan, E.C., Fiers, W., and Pober, J.S. (1986) Recombinant tumor necrosis factor and immune interferon act singly and in combination to reorganize human vascular endothelial cell monolayers. Am. J. Pathol., 123t16-24. Tamm, I., Kikuchi, T., Wang, E., and Pfeffer, L.M. (1984) Growth rate of control and p-interferon-treated human fibroblast populations over the course of their in vitro life span. Cancer Res., 4422922296. Tan, Y.H., Chou, E.L., and Lundh, N. (1975) Regulation of chromo-

some 21-directed anti-viral gene(s) as a consequence of age. Nature, 257t310-312. Thornton, S.C., Mueller, S.N., and Levine, E.M. (1983) Human endothelial cells: Use of heparin in cloning and long-term serial cultivation. Science, 222t623-625. Yamazaki, S., Onishi, E., Enami, K., Natori, K., Kohase, M.: Sakamoto, H., Tanouchi, M., and Hayashi, H. (1986) Proposal of standardized methods and reference for assaying recombinant human tumor necrosis factor. Jpn. J. Med. Sci. Biol., 39:106-118.