JOURNAL OF CELLULAR PHYSIOLOGY 156:294-302 (1993)

Accumulation of Insulin-Like Growth Factor Binding Protein-3 in Conditioned Medium of

Human Fibroblasts Increases With Chronologic Age of Donor and Senescence

In Vitro SAMUEL GOLDSTEIN,* ELENA J. MOERMAN, AND ROBERT C. BAXTER

Departments of Medicine and Biochemistry and Molecular Biology, University of Arkansas for Medical Sciences and Geriatric Research, Education, Clinical Center, 1. L. McClellan

Memorial Veterans' Hospital, Little Rock, Arkansas 72205 (S.C., E.I.M.); and Department of Endocrinology, Royal Prince Alfred Hospital, Camperdown, NSW 2050, Australia (R.C. B.)

We have found that insulin-like growth factor binding protein-3 (IGFBP-3) accu- mulates to higher levels in medium conditioned by a strain of normal fibroblasts at late passage (LP) and a strain derived from subjects with Werner syndrome (WS) of premature aging, compared to medium conditioned by the same normal cells at early passage (EP) (Goldstein et al., Proc. Natl. Acad. Sci. USA, 8819680-9684, 1991). To explore the generality of this phenomenon with respect to chronologi- cal age of donor (in vivo aging) and LP (in vitro senescence) we assayed IGFBP-3 in medium conditioned by 18 normal fibroblast strains at EP and LP and two WS strains at the midpoint of their curtailed replicative lifespans and assessed IGFBP-3 mRNA levels in cells by Northern analysis. The lowest accumulations of IGFBP-3 were found in medium conditioned by fetal cells with progressively increasing amounts postnatally; direct correlations between IGFBP-3 levels and donor age were seen in EP cells 3 days after subculture (during logarithmic growth) r = 0.80, P < 0.001, and 7 days after subculture (at confluence) r = 0.77, P < 0.001. With two exceptions, conditioned medium of cell strains accumulated more ICFBP-3 at LP; IGFBP-3 levels correlated with chronological age after 3 days, r = 0.50, P = 0.05, and after 7 days, r = 0.75, P < 0.001. IGFBP-3 content of WS culture medium fell within the range of LP normal cells. Cumulative IGFBP-3 levels were inversely proportional to the thymidine labeling index, a measure of proliferative vigor. With some exceptions IGFBP-3 mRNA levels were commensurate with the amount of IGFBP-3 accumulated in the medium, suggesting that distal transla- tional and posttranslational mechanisms also regulate IGFBP-3 production in some strains. The trend toward augmented IGFBP-3 output of fibroblasts as a direct function of chronological age and in vitro senescence and as an inverse function of proliferative vigor i s consistent with the known inhibitory effect of excess IGFBP-3 on IGF-mediated DNA synthesis and the reduced regenerative potential that i s evident during biological aging in vivo. o 1993 WiIey-Liss, Inc.

Insulin-l ike growth factors-I and I1 (IGF-I and 11) are found in body f lu ids and culture medium in association w i t h one or more o f a family o f s ix specific binding proteins (Shimasaki et al., 1991). The predominant binding protein in adult serum, IGF binding protein-3 (IGFBPS) exists as part o f a ternary, growth hormone- dependent complex o f -150 kDa, which also includes IGF-I or IGF-I1 and an acid-labile glycoprotein o f -85 kDa (Baxter and Martin, 1989). IGFBP-3 regulates the effects o f I G F - I and I1 o n DNA synthesis and other cellular processes by as yet poorly understood interac- t ions w i t h components o f extracellular ma t r i x and spe- cific IGF receptors o n the plasma membrane (Mc- Cusker et al., 1990, 1991; Conover, 1991). Thus, IGFBP-3 has the capacity to either potentiate or inhibit 0 1993 WILEY-LISS, INC.

IGF-I action on DNA synthesis in human diploid fibro- blasts (HDF) and bovine fibroblasts in v i t ro (DeMellow and Baxter, 1988; Conover e t al., 1990) and in wound heal ing models in vivo (Sommer et al., 1991) depending on the molar rat io of IGFBP-3IIGFs and on precise spa- tiotemporal factors (DeMellow and Baxter, 1988; Conover et al., 1990).

We recently reported that IGFBP-3 mRNA is overex- pressed in normal senescent HDF and prematurely se-

Received September 3,1992; accepted March 22, 1993. *To whom reprint requestdcorrespondence should be addressed.

295 IGFBP-3 OUTPUT BY AGING HUMAN FIBROBLASTS

TABLE 1 Strains of HDF used in this study

Cell Cell Maximum strain Age (code)’ Gender type MPD References

~~

Normal HSC172 MRC5 W138 A30s

csc103 A24

A25? A2 A233

A20

A8

A39

5032 5065 A32

A33

A35b

5088

ws12 WS8a

ws

Fetal (F1) Fetal (F2) Fetal (F3) Newborn (NBI)”

Newborn (NB2) 9 (9a)

9 i9b) 11 23

28

31

51

53 56 61

70

76 (76a)

76 (76b)

46 47

0 P P 9

6 6

9 6 6

6

‘3

6

0 6 d

6

6

Lung Lung Lung Skin

Foreskin Skin

Skin Skin Skin

Skin

Skin

Skin

Skin Skin Skin

Skin

Skin

P Skin

6 Skin 6 Skin

62 60 64 48

70 58

48 54 56

40

56

40

44 44 20

35

33

Buchwald and Ingles (1976) Jacobs et al. (1970) Hayflick (1965) Moerman and Goldstein

(unpublished) Smith et al. (unpublished) Moerman and Goldstein

Wojtyk and Goldstein (1980) Harley and Goldstein 11978) Shmookler Reis and Goldstein

Moerman and Goldstein

Moerman and Goldstein

Moerman and Goldstein

Goldstein et al. (1979) Goldstein et al. ( 1979) Moerman and Goldstein

Moerman and Goldstein

Moerman and Goldstein

(unpublished)

(1983)

(unpublished)

(unpublished)

(unpublished)

(unpublished)

(unpublished)

(unpublished) 44 Goldstein et al. (1979)

19 Moerman et al. (unpublished) 18 Murano et al. (1991)

’9bbreviations of strains used in Northern blots (Figures 4 and 5 ) are given in parentheses or as the age in years. ‘Stillborn at term. 3.425 formerly designnted as N-2 and A23 as TM. Biopsy site in A and J strains: anterior forearm -1 inch below elbow crease; for WS strains: inner aspect of upper arm

nescent HDF derived from subjects with Werner syn- drome (WS), a genetically determined disorder of premature aging, compared to vigorously proliferating early-passage HDF from normal donors (Murano et al., 1991; Goldstein et al., 1991). We also found that IGFBP-3 levels in medium conditioned by all three of- these cell types increased as cells became quiescent both during high-density growth arrest and during se- rum deprivation (Goldstein et al., 1991; Moerman et al., 1993). In the present study we have surveyed several strains of HDF derived from donors of various ages across the lifespan including fetal, postnatal, young a.dult, and elderly donors. We examined the accumula- tion of IGFBP-3 in culture medium conditioned by each strain of HDF at two stages of the limited replicative lifespan, when cells were at early passage (EP) and at late passage (LP). The results indicate a remarkable direct correlation between increasing IGFBP-3 accu- mulation and both chronological donor age and in vitro replicative senescence.

MATERIALS AND METHODS Cultured fibroblasts

Twenty strains of HDF were studied including 17 derived from normal donors, one from a term fetus that

stein, 1986). Other strains were obtained from the Ge- netic Mutant Repository, Camden, NJ, or as gifts from other investigators. All cells were propagated in regu- lar growth medium consisting of Eagle’s minimum es- sential medium supplemented by 15% fetal bovine se- rum without antibiotics, counting the number of mean populations doublings (MPD) as an index of replicative age (Haslam and Goldstein, 1974; Goldstein et al., 1991). Normal HDF at EP were defined as cells capable of vigorous proliferation at a low MPD level, i.e., in the first half of their replicative lifespan, whereas LP cells were studied after 290% of the replicative lifespan had been consumed, when growth had slowed, cells became enlarged, and showed other evidence of replicative se- nescence (Hayflick, 1965; Haslam and Goldstein, 1974; Goldstein, 1989,1990; Cristofalo et al., 1989). In prepa- ration for each experiment HDF were subcultured into replicate 60 or 100 mm Petri dishes in regular growth medium at a split ratio such that they became conflu- ent at day 7 without change of medium. These ratios, which varied from 1:20 for fetal strains to 1:3 for WS cells, did not affect the relative differences in IGFBP-3 output observed in EP and LP cells (V.G. Grigoriev, E.J. Moerman and S. Goldstein, manuscript in prepara- tion).

was stillborn but grossly normal at autopsy, and two

and WS8a had been studied before (Goldstein et al., 1991) and are included for comparative purposes. Cul- tures were established primarily in our own laboratory from skin biopsies as described (Moerman and Gold-

from subjects with classical WS (Table 1). Strains 5065 Determination Of thymidine labeling index (TLI) HDF were seeded into single chamber LAB-TEK

slides (Nunc Inc., Naperville, IL) at the same time and split ratio used for subculture into Petri dishes. Forty- eight hours later 3H-thymidine (Specific activity 6.7

296 GOLDSTEIN ET AL.

0

Fig. 1. Cumulative levels of IGFBP-3 in medium conditioned for 3 days by HDF derived from donors of increasing age. Cells were subcul- tured into regular growth medium. Three days later medium was collected and assayed for IGFBP-3, whereas cell monolayers were sol- ubilized and assayed for total protein. 0 E P HDF; LP HDF; A WS HDF at mid passage. The vertical line between symbols connects EP and LP values from individual cell strains. All cell strains are depicted

Ci/mmole, NEN, Boston, MA) was added to a final con- centration of 0.1 FCi/ml. Following a 30-hour incuba- tion, monolayers were washed three times with phos- phate buffered saline (PBS), fixed for 15 minutes at 4°C in 5% glutaraldehyde (Polysciences, Inc., Warrington, PA) in PBS, rinsed with 70% ethanol, and air dried. Slides were prepared for autoradiography and analyzed microscopically for percent labeled nuclei (Cristofalo and Sharf, 1973).

IGFBP-3 assays Medium was rapidly removed from replicate cell cul-

tures after 3 or 7 days and immediately frozen at - 70°C. Radioimmunoassay for IGFBP-3 was carried out on thawed medium as previously described using R7 antibody (Baxter and Martin, 1986; Martin and Baxter, 1988). Total cellular protein was determined (Smith et al., 1985) on cell monolayers following rising with PBS and dissolving in 0.1 N NaOH-0.4% deoxy- cholate. Results are expressed as the mean nanograms of IGFBP-3 accumulated per microgram total cellular protein in each dish at days 3 and 7.

Northern analysis Total cellular RNA was prepared from fibroblasts

immediately after removal of conditioned medium and 10 pg/lane was resolved on agarose/formaldehyde gels. Gels were blotted onto nylon filters, hybridized with the full-length (2.5 kbp) IGFBP-3 32P-labeled cDNA insert, then washed and prepared for autoradiography (Srivastava et al., 1985; Murano et al., 1991). To ensure equivalent RNA loading and transfer, 32P-labeled DNA specific for 285 ribosomal RNA was probed against the same filters after stripping the IGFBP-3 probe (Murano et al., 1991).

Statistical analysis Linear regression analysis was performed on normal

cell strains a t EP and LP by the method of least squares

in the order listed in Table 1, the three fetal strains to the left of zero and the two newborn strains a t and to the right of zero. The two 76-year-old strains are 76a below and 76b above. Correlations be- tween cumulative levels of IGFBP-3 in conditioned medium and donor age, at EP (n = 18, exclusive of WS strains: r = 0.80,P < 0.001); a t LP r = 0.50, P = 0.05.

and levels of significance were determined by the two- tailed t test (Armitage, 1971).

RESULTS Accumulation of IGFBP-3 in conditioned

medium IGFBP-3 content of medium conditioned by EP and

LP cells 3 days after subculture, i.e., during the period of most rapid growth, is shown in Figure 1. At EP, fetal strains showed the lowest accumulations of IGFBP-3 in conditioned medium. A trend toward increasing IGFBP-3 accumulation was evident in conditioned me- dium of postnatal cells, with the highest levels in me- dium of normal cells at EP recorded in the 76-year-old female subject (76b). The IGFBP-3 level in medium conditioned by HDF of the 47-year-old WS subject was higher than all of the EP normal values; although these WS cells were prematurely senescent when studied at 12 MPD, they were able to grow slowly and achieve a maximum replicative lifespan of 18 MPD. Thus, they were comparable to LP cultures of normal subjects, which with the exception of the 28-year-old donor (A20), all showed higher accumulations of IGFBP-3 in the medium at LP vs. EP. There was a statistically significant direct correlation at EP between IGFBP-3 accumulations and donor age (r = 0.80, P < 0.001) but this correlation just reached the boundary of signifi- cance at LP ( r = 0.50, P = 0.05). Although the lowest levels at both EP and LP appeared in two of the three fetal and one of the two neonatal strains, the largest increments in LP vs. EP cells were seen in one of the fetal strains (F2), one of the newborn strains (NBl), one of the 9-year-old strains (9b), and the 56-year-old strain.

Similar results were obtained for IGFBPS accumula- tion in conditioned medium of EP and LP cells 7 days after subculture (Fig. 2), at which time cells had be- come confluent and had undergone high-density prolif- erative arrest. The change in scale should be noted in

IGFBP-3 OUTPUT BY AGING HUMAN FIBROBLASTS 297 - 160 c Q) 140-

Q D 100- 3 0 80- t

60-

2 40-

0 early 9 late A WS .- I + 9 120-

W

g 20- a - 0-

-10 0 10 20 30 40 50 60 70 80

DONOR AGE (years)

Fig. 2. Cumulative levels of IGFBP-3 in medium conditioned for 7 days by HDF derived from donors of increasing age. Replicate dishes were grown to confluent density at day 7 and treated exactly as in Figure 1. Symbols as in Figure 1. Correlations between cumulative levels of IGFBP-3 in conditioned medium and donor age, at EP: r = 0.77,P < 0.001; at LP: r = 0.75,P < 0.001.

comparing IGFBP-3 levels after 3 and 7 days. With the exception of a few strains that showed mild departures from linearity (both higher and lower), IGFBP-3 accu- mulated in the medium at a relatively constant rate over the 7-day interval. Strain A20 (28 years), as a t day 3 , showed a reversal in the E P to LP trend toward increased IGFBPS accumulation, with A35b (76a) showing reversal a t day 7 only. In contrast to day 3, however, the direct correlation at day 7 between IGFBP-3 accumulation and donor age was significant for both E P cells (r = 0.77, P < 0.001) and LP cells (r = 0.75, P < 0.001).

In order to assess the relation between proliferative rate and IGFBP-3 output, we first sought correlations between the TLI and in vivo age (Fig. 3A). As expected from several previous studies (e.g., Goldstein et al., 1.969, 1978; Schneider and Mitsui, 1976) the TLI of cultures at E P was inversely proportional to donor age. Thus, fetal and newborn strains displayed the fastest proliferative rates while strains derived from older do- nors proliferated more slowly. IGFBPS output in E P cultures correlated inversely with TLI a t day 3 (Fig. 3B) and at day 7 (Fig. 3 0 . For LP cultures no signifi- cant correlations were found between TLI and donor age and between IGFBP-3 output and TLI (data not shown).

Northern analyses of IGFBP-3 mRNA content Total RNA extracted from HDF was examined for

IGFBP-3 mRNA content by Northern analysis at day 3 (Fig. 4) and day 7 (Fig. 5). As with IGFBP-3 levels in culture medium, IGFBP-3 mRNA levels on both days tended to be low in fetal strains with a trend toward increasing mRNA levels in cells derived from postnatal donors, commensurate with the increasing accumula- tions of IGFBP-3 in conditioned medium. However, there was not a strict correlation between these two parameters. For example, NB1 and 9b showed pro- nounced increases in mRNA levels at LP compared to E:P on both days 3 and 7, which correlated with aug- mented IGFBP-3 accumulation (Figs. 1, 2). Other

strains, e.g., 11, 70, and 56 years, showed increased mRNA levels at LP only on day 3, but these levels were nonetheless associated with high LP increments of cu- mulative IGFBP-3 on both days. On the other hand, strain 76b showed low IGFBP-3 mRNA levels a t EP, with essentially no change at LP, whereas IGFBP-3 accumulation in the medium of 76b was high at EP with relatively little change at LP. Strains 61 and 76a, which showed high IGFBP-3 mRNA levels at EP with no increase at LP, generated substantial increments in cumulative IGFBP-3 a t LP. WS cells displayed mRNA levels in the middle range similar to their moderate to high levels of IGFBP-3 accumulation in the medium. Strain 28 had the same IGFBP-3 mRNA level a t EP and LP on day 3 and a lower mRNA level at LP vs. EP on day 7, consistent with the reversed pattern of IGFBP-3 accumulation in the medium.

DISCUSSION The present study demonstrates direct correlations

between the accumulation of IGFBP-3 in medium con- ditioned by HDF and aging in vivo and in vitro, mea- sured by the chronological age of donors and replicative senescence, respectively. Furthermore, cumulative IGFBP-3 levels were inversely proportional to the pro- liferative rate of EP cells. Taken together, these corre- lations indicate that the progressive augmentation of IGFBP-3 output characteristic of HDF undergoing ag- ing in vivo and in vitro depends on a reduction in prolif- eraive vigor.

Steady-state levels of the cognate mRNA were not always commensurate with the cumulative IGFBP-3 levels attained in the medium. It would appear that whereas IGFBP-3 output bears a general relationship to levels of the cognate mRNA, regulation is also ex- erted translationally and/or posttranslationally. Multi- ple translational controls are possible including factors intrinsic to the mRNA sequence. We have found that although the IGFBP-3 mRNA in WS8a cells contains several polymorphic changes compared to wild type, these cannot account for the relative attenuation in

298 GOLDSTEIN ET AL.

>- r= - 0.81 I p c 0.001 $ I- :: -10 0 10 20 30 40 50 a 70

76bO

061

DONOR AGE (years) B

2 10- m (3 5- L

oi I I

60 70 80 90 THYMIDINE LABELING INDEX(%)

h

“ I 076b 053

70 90 1 THYMIDINE LABELING INDEX(%)

I 0

D

Fig. 3. Correlations between proliferative capacity at EP, chronological age of donors, and IGFBP-3 output. Proliferative capacity measured as TLI vs. in vivo age (A). IGFBP-3 accumulation in day 3 medium (B) and day 7 medium (C) vs. TLI. Prematurely senescent WS8a and WS12 cells (TLI, 37 and 19%, respectively) were excluded from these analyses.

IGFBP-3 OUTPUT BY AGING HUMAN FIBROBLASTS 299

Fig. 4. Northern analysis of IGFBP-3 mRNA in EP and LP HDF derived from donors of increasing age 3 days after subculture. Total cellular RNA was isolated from cell strains depicted in Figure 1, frac- tionated on agarose-formaldehyde gels, transferred to nylon filters, and hybridized to 32P-labeled IGFBP-3 cDNA. After autoradiography,

the probe was stripped and blots were reprobed with "P-28s rDNA to demonstrate even transfer and relatively constant expression. Each normal donor is represented by a pair of lanes, at EP and LP, respec- tively, denoted by hash marks. The age of the donor appears above each pair of lanes. WS cells were studied at mid passage.

IGFBP-3 output compared, for example, to strain 5065 (Figs. 1, 2 vs. Figs. 4, 5, and see Thweatt et al., 1993). This disparity between IGFBP-3 mRNA levels and IGFBPS output must relate, therefore, to other compo- nents of the translational apparatus. Regarding post- translation mechanisms such as proteolysis, previous studies indicate that the antibody utilized here mea- sures several proteolytic derivatives of IGFBP-3 ob- tained in vivo (Suikkari and Baxter, 1992) and in vitro (Martin and Baxter, 1988) with M,s ranging from 30-35 kDa that are still capable of binding IGFBP-3. Therefore, in conjunction with the generally linear ac- cumulation of IGFBP-3 in the medium in this and ear- lier reports (Martin and Baxter, 1988; Goldstein et al., 1991) we believe it is unlikely that differential rates of proteolysis account for the disparities between IGFBP-3 mRNA and IGFBP-3 output in some cell strains.

Alternatively, although the quantity of IGFBP-3 pro- duced is generally proportional to mRNA levels sub- stantial amounts of secreted IGFBP-3 may reassociate with the cell. There is now increasing evidence that binding and immobilization of cytokines and growth factors to proteoglycans and glycosoaminoglycans oc- cur a t or near cell surfaces, thereby affecting binding

affinities for receptors and binding proteins (Ruoslahti and Yamaguchi, 1991). Recent studies indicate that fibroblast-associated IGFBP-3 can be released by hep- arin, suggesting the involvement of glycosaminogly- cans in the binding sites (Martin et al., 1992). Such a sequestration mechanism, which would render IGFBP-3 unavailable for assay in the soluble medium, may in part explain the apparent discrepancy between the age-dependent increases in IGFBP-3 accumula- tions reported here and the lower plasma IGFBP-3 lev- els found in older persons (Baxter and Martin, 1986). Thus, higher output of IGFBP-3 by cells such as fibro- blasts in vivo need not necessarily contribute to a mea- surable rise in plasma levels but rather may increase IGFBP-3 concentrations at short range. Indeed, Daughaday and Rotwein (1989), proposing that similar disparities occur for the IGFs, invoke differences be- tween endocrine secretions that are detectable in the plasma and localized autocrine and paracrine secre- tions that are not. Nonetheless, other factors may ac- count for the disparity between in vivo and in vitro findings, including clearance and degradation of IGFBP-3 by organs such as the kidney and liver in vivo.

The molecular genetic basis for the trend toward in- creasing IGFBP-3 accumulation in conditioned me-

300 GOLDSTEIN ET AL.

Fig. 5. Northern analysis of IGFBP-3 mRNA in EP and LP HDF derived from donors of increasing age 7 days after subculture. Total cellular RNA was isolated from cell strains depicted in Figure 2, and processed as before.

dium of HDF as a function of chronological donor age and in vitro senescence is unknown. However, our present observations indicate that different fibroblast strains apparently utilize a specific subset of the sev- eral regulatory steps available, from gene transcription and mRNA stabilization, to translation and posttrans- lational mechanisms (Thweatt and Goldstein, 1993).

The functional role of this age-dependent increase in IGFBP-3 production is also unclear but can be rational- ized by two major mechanisms. Increased IGFBP-3 lev- els may represent an age-dependent adaptation such that progressively older cells, both in vivo and in vitro, are attempting to compensate for the reduced IGF-me- diated stimulation of DNA synthesis observed in senes- cent HDF (Harley et al., 1981; Phillips et al., 1987; Moerman et al., 1993; Goldstein et al., 1991). Alterna- tively, high levels of IGFBP-3 secreted into the medium may reflect an active, genetically regulated cellular response designed to sequester IGFs and thus limit their bioavailability. In this latter mode, IGFBP-3 would play a central role in senescent and quiescent replicative arrest via an autocrinelparacrine mecha- nism (Goldstein et al., 1991; Moerman et al., 1993). Preliminary studies indicate that normal HDF at EP and LP and WS HDF are all susceptible to inhibition of IGF-I-mediated stimulation of DNA synthesis by exog-

enously added recombinant human IGFBP-3 at molar ratios of IGFBP-3/IGF-I 3 1 (Moerman et al., 1993, and work in preparation). Moreover, addition to HDF un- dergoing high-density growth arrest of increasing amounts of IGF-I or IGF-I analogs, which do not bind appreciably to IGFBP-3 but bind normally to the IGF-I plasma membrane receptor, enables resumption of DNA synthesis and cell division in EP cells and to a significant but lesser degree in senescent cells (S. Gold- stein, V.G. Grigoriev and E.J. Moerman, in prepara- tion). Elevated IGFBP-3 levels, therefore, may play a pivotal role in mediating the high-density growth ar- rest in EP cells although the situation is clearly more complex. HDF also secrete IGFBP-4, 5, and 6 (Martin et al., 1992; Camacho-Hubner et al., 1992; J.L. Martin, J.A. Coverley, and R.C. Baxter, in preparation), and IGFBP-3 also appears to inhibit the action of growth factors other than IGFs (Liu et al., 1992). Similarly, we envisage that senescent proliferative arrest a t LP also involves greater complexity because it is not reversed by reducing the IGFBP-3IIGF ratio either via the addi- tion of supraphysiologic levels of IGF-I or depletion of IGFBP-3. In short, although excess IGFBP-3 can be inculpated in HDF senescence, this phenomenon also appears to involve the concerted action of several other overexpressed gene products, many of which are capa-

301 IGFBP-3 OUTPUT BY AGING HUMAN FIBROBLASTS

ble of inhibiting DNA synthesis (Goldstein, 1990; Mu- ran0 et al., 1991; Thweatt and Goldstein, 1993).

Finally, despite the complexity inherent in this sys- tem, the present results have implications for in vivo events. The low levels of IGFBP-3 production in fetal cells at EP, with the trend to increased production in postnatal cells, especially after adolescence, suggest an important inverse regulatory role for IGFBP-3. The so- matic stunting in WS may be rationalized as prema- turely high output of IGFBPS along with the products of other antiproliferative genes leading to growth shut- down before adolescence (Murano et al., 1991; Gold- stein et al., 1991; Thweatt and Goldstein, 1993). IGFBP-3 excess could also predispose, a t any age, to generalized or focal impairment of wound healing as is seen, for example, in WS, in diabetes mellitus, and in many elderly persons (Goldstein, 1989). Further stud- ies in vitro and in vivo are needed to resolve these fimdamental issues of normal and aberrant growth reg- ulation in health and disease.

ACKNOWLEDGMENTS We thank Joseph Daniel and Sarah Holmes for ex-

pert technical assistance and Patricia Spies for prepar- ing the manuscript. These studies were supported by grants to S.G. from the National Institutes of Health (AG08708), the Arkansas Experimental Program to Stimulate Competitive Research (funded by the Na- tnonal Science Foundation, the Arkansas Science and Technology Authority, and the University of Arkansas for Medical Sciences), and the Department of Veterans Affairs, and to R.C.B. from the National Health and Medical Research Council, Australia.

LITERATURE CITED Armitage, P. (1971) Statistical Methods in Medical Research. John

Wiley and Son, New York, pp. 150-166. Baxter, R.C., and Martin, J.L. (1986) Radioimmunoassay of growth

hormone-dependent insulinlike growth factor binding protein in hu- man plasma. J . Clin. Invest., 78.1504-1512.

Baxter, R.C., and Martin, J.L. (1989) Structure of the M , 140,000 growth hormone-dependent insulin-like growth factor binding pro- tein complex: Determination by reconstitution and affinity-label- ing. Proc. Natl. Acad. Sci. USA, 86:689%6902.

Buchwald, M., and Ingles, C.J. (1976) Human diploid fibroblast mu- tants with altered RNA polymerase 11. Somat. Cell Genet., 2r225- 233.

Camacho-Hubner, C., Busby W.H., McCusker, R.H., Wright, G., and Clemmons, D.R. (1992) Identification of the forms of insulin-like growth factor-binding proteins produced by human fibroblasts and the mechanisms that regulate their secretion. J . Biol. Chem., 267:1194%11956.

Conover, C.A. (1991) A unique receptor-indpendent mechanism by which insulinlike growth factor I regulates the availability of in- sulinlike growth factor binding proteins in normal and transformed human fibroblasts. J. Clin. Invest., 88t1354-1361.

Conover, C.A., Ronk, M., Lombana, F., and Powell, D.R.(1990) Struc- tural and biological characterization of bovine insulin-like growth factor binding protein 3. Endocrinology, 227r2795-2803.

Cristofalo, V.J., and Sharf, B.B. (1973) Cellular senescence and DNA synthesis. Thymidine incorporation as a measure of population age in human diploid cells. Exp. Cell Res., 76t419-427.

Cristofalo, V.J., Doggett, D.L., Brooks-Frederich, K.M., and Phillips, P.D. (1989) Growth factors as probes of cell aging. Exp. Gerontol., 24r367374.

Daughaday, W.H., and Rotwein, R. (1989) Insulin-like growth factors I and 11. Peptide, messenger ribonucleic acid and gene structures, serum, and tissue concentrations. Endocrinol Rev., IOt6&&91.

DeMellow, J.S.M., and Baxter, R.C. (1988) Growth hormone-depen- dent insulin-like growth factor (IGF) binding protein both inhibits

and potentiates IGF-I-stimulated DNA synthesis in human skin fibroblasts. Biochem. Biophys. Res. Commun., 156:199-204.

Goldstein, S. (1989) Cellular senescence. In: Endocrinology, 2nd Ed. L.J. DeGroot, G.F. Cahill, Jr., W.D. Odell, L. Martini, J.T. Potts, Jr., D.H. Nelson, E. Steinberger, and A.I. Winegrad, eds. Grune and Stratton, New York, pp. 2525-2549.

Goldstein, S. (1990) Replicative senescence: The human fibroblast comes of age. Science, 249t1129-1133.

Goldstein, S., Littlefield, J.W., and Soeldner, J.S. (1969) Diabetes mel- litus and aging: Diminished plating efficiency of cultured human fibroblasts. Proc. Natl. Acad. Sci. USA, 64:155-160.

Goldstein, S., Moerman, E.J., Soeldner, J.S., Gleason, R.E., and Bar- nett, D.M. (1978) Chronologic and physiologic age affect replicative lifespan of fibroblasts from diabetic, prediabetic and normal donors. Science, 199t781-782.

Goldstein, S., Moerman, E.J., Soeldner, J.S., Gleason, R.E.. and Bar- nett, D.M. (1979) Diabetes mellitus and genetic prediabetes: De- creased replicative capacity of cultured skin fibroblasts. J. Clin. Invest., 63:35%370.

Goldstein, S., Moerman, E.J., Jones, R.A., and Baxter, R.C. (1991) Insulin-like growth factor binding protein-3 accumulates to high levels in culture medium of senescent and quiescent human fibro- blasts. Proc. Natl. Acad. Sci. USA, 88t9680-9684.

Harley, C.B., and Goldstein, S. (1978) Cultured human fibroblasts: Distribution of cell generations and a critical limit. J. Cell. Physiol., 97r509-516.

Harley, C.B., Goldstein, S., Posner, B.I., and Guyda, H. (1981) De- creased sensitivity of old and progeric human fibroblasts to a prepa- ration of factors with insulin-like activity. J. Clin. Invest., 68r988- 994.

Haslam, R.J., and Goldstein, S. (1974) Adenosine 3’5’-cyclic mono- phosphate in young and senescent human fibroblasts during growth and stationary phase in vitro. Effects of prostaglandin E l and of adrenaline. Biochem. J. , 144t253-263.

Hayflick, L. (1965) The limited in vitro lifetime of human diploid cell strains. Exp. Cell Res., 37r61P636.

Jacobs, J.P., Jones, C.M., and Baille, J.P. (1970) Characteristics of a human diploid cell designated MRC-5. Nature, 227:168-170.

Liu, L., Delbe, J., Blat, C., Zapf, J . , and Harel, L. (1992) Insulin-like growth factor binding protein (IGFBP-3), an inhibitor of serum growth factors other than IGF-I and -11. J. Cell. Physiol., 153t15-21.

Martin, J.L., and Baxter, R.C. (1988) Insulin-like growth factor bind- ing proteins (IGFBPs) produced by human skin fibroblasts: Immu- nological relationship to other human IGFBPs. Endocrinology, 123t1907-1915.

Martin, J.L., Ballesteros, M., and Baxter, R.C. (1992) Insulin-like growth factor (1GF)-I and transforming growth factor-p release IGF binding protein-3 from human fibroblasts by different mechanisms. Endocrinology, 132r1703-1710.

McCusker, R.H., Camacho-Hubner, C., Bayne, M.L., Cascieri, M.A., and Clemmons, D.R. (1990) Insulin-like growth factor (IGF) binding to human fibroblast and glioblastoma cells: The modulating effect of cell released IGF binding proteins (IGFBPs). J. Cell. Physiol., 144:24&253.

McCusker, R.H., Busby, W.H., DeHoff, M.H., Camacho-Hubner, C., and Clemmons, D.R. (1991) Insulin-like growth factor (IGF) binding to cell monolayers is directly modulated by the addition of IGF- binding proteins. Endocrinology, 129t939-999.

Moerman, E.J., and Goldstein, S. (1986) Culture of human skin fibro- blasts. In: Methods in Diabetes Research, Vol. 2: Clinical Methods. S.L. Pohl, J . Larner, and W.L. Clarke, eds. John Wiley and Sons, New York, pp. 283-312.

Moerman, E.J., Thweatt, R., Moerman, A.M., Jones, R.A., and Gold- stein, s. (1993) Insulin-like growth factor binding protein-3 is over- expressed in senescent and quiescent human fibroblasts. Exp. Ger- ontol. (in press).

Murano, S., Thweatt, R., Shmookler Reis, R.J., Jones, R.A., Moerman, E.J., and Goldstein, S. (1991) Diverse gene sequences are overex- pressed in Werner syndrome fibroblasts undergoing premature rep- licative senescence. Mol. Cell. Biol., 11 t3905-3914.

Phillips, P.D., Pignolo, R.J., and Cristofalo, V.J. (1987) Insulin-like growth factor-I: Specific binding to high and low affinity sites and mitogenic action throughout the life span of WI-38 cells. J. Cell. Physiol., I33:135-143.

Ruoslahti, E., and Yamaguchi, Y. (1991) Proteoglycans as modulators of growth factor activities. Cell, 642467-869.

Schneider, E.L., and Mitsui, Y. (1976) The relationship between in vitro cellular aging and in vivo human age. Proc. Natl. Acad. Sci. USA, 73t3584-3588.

302 GOLDSTEIN ET AL.

Shimasaki, S., Shimonaka, M., Zhang, H.P., and Ling, N. (1991) Isola- tion and molecular characterization of three novel insulin-like growth factor binding proteins (IGFBP-4,5, and 6). In: Modern Con- cepts of Insulin-like Growth Factors. E.M. Spencer, ed. Elsevier, New York, pp. 343-358.

Smith, P.K., Krohn, R.I., Hermanson, G.T., Mallia, A.K., Gartner, F.H., Provenzano, M.D., Fujimoto, E.K., Goeke, N.M., Olson, B.J., and Klenk, D.C. (1985) Measurement of protein using bicinchoninic acid. Anal. Biochem., 150:7685.

Sommer, A., Maack, C. A., Spratt, S.K., Mascarenhas, D., Tressel, T.J., Rhodes, E.T., Lee, R., Roumas, M., Tatsuno, G.P., Flynn, J.A., Ger- ber, N., Taylor, J., Cudny, H., Nanney, L., Hunt, T.K., and Spencer, E.M. (1991) Molecular genetics and actions of recombinant insulin- like growth factor binding protein-3. In: Modern Concepts of Insu- lin-like Growth Factors. E.M. Spencer, ed. Elsevier, New York, pp. 7 15-728.

Srivastava, A,, Norris, J.S., Shmookler Reis, R.J., and Goldstein, S. (1985) c-Ha-ras-1 proto-oncogene amplification and overexpression during the limited replicative life span of normal human fibro- blasts. J. Biol. Chem., 260:6404-6409.

Suikkari, A.-M., and Baxter, R.C. (1992) Insulin-like growth factor- binding protein-3 is functionally normal in pregnancy serum. J. Clin. Endocrinol. Metab., 74:177-183.

Thweatt, R., and Goldstein, S. (1993) Werner syndrome and biological aging: A molecular genetic hypothesis. Bioessays, 15: (in press).

Thweatt, R., Fleischmann, R., and Goldstein, S. (1993) Normal cDNA sequence of insulin-like growth factor binding protein-3 from Werner syndrome fibroblasts. DNA Sequence (in press).

Wojtyk, R., and Goldstein, S. 11980) Fidelity of protein synthesis does not decline during aging of cultured human fibroblasts. J. Cell. Physiol., I03:299-303.