Nonenzymatic glycation of type I collagen modifies interaction with UMR 201-10B preosteoblastic cells

BoneVol. 21, No. 3 September 1997:237-242

ELSEVIER

Nonenzymatic Glycation of Type I Collagen Modifies Interaction With UMR 201-10B Preosteoblastic Cells Y. K A T A Y A M A , 1 S. CELIC, 1 N. N A G A T A , 2 T. J. MARTIN, 1 and D. M. FINDLAY 1. 1 St. Vincent's Institute of Medical Research, Fitzroy, Victoria, Australia 2 Third Department oflnternal Medicine, National Defense Medical College, Saitama, Japan

237-242; 1997) reserved.

Advanced glycation endproduct (AGE), whose formation is accelerated on long lived extracellular matrix proteins in diabetes, is implicated in diabetic complications in various tissues. Type I collagen is the predominant matrix protein of bone and plays an important role in bone cell-matrix interactions. We have previously reported the accelerated accumulation of AGE collagen in bone tissue in diabetes mellitus (DM), in which reduced bone mineral density was observed. In addition, when cultures of mature primary rat osteoblasts were plated onto an in vitro AGE-modified collagen substrate, they showed altered cell functions, in terms of alkaline phosphatase (ALP) activity, osteocalcin secretion, and nodule formation (J Bone Miner Res 11:931-937; 1996). To determine whether AGE collagen might also affect differentiation of preosteoblasts, we compared the effects of plating the preosteoblastic UMR 201-10B cell line onto AGE-modified collagen with plating onto unmodified collagen. The latter had been shown previously to promote differentiation of UMR 201 cells. We have also explored whether these effects might be partly mediated by the transforming growth factor beta (TGF-13) receptor. Growth of UMR 201-10B cells on a type I collagen substrate significantly inhibited cell growth and retinoic acid (RA)-indueed upregulation of ALP activity, compared to cells on plastic. These inhibitory effects were reduced by prior glycatlon of collagen, in a dose-dependent manner with respect to AGE content. Unmodified collagen stimulated production of osteopontin mRNA, which was reduced by AGE modification to levels attained in cells on plastic. Growth on control collagen inhibited TGF-I~ type II receptor binding in 10B cells, while this inhibition was reduced by AGE modification. These data suggest that glycatlon of collagen interferes with specific interactlon(s) between UMR 201-10B cells and collagen. Based on our previous results in UMR 201 cells, these results would be compatible with the notion that glycated collagen has reduced ability to promote differentiation of prensteoblasts to mature osteoblasts. These data further suggest that collagen-mediated events in these cells may be at least in part mediated by regulation of the TGF-13 receptor expression. (Bone 21:

Key Words: Collagen; Glycation; Osteopenia; Diabetes mellitus; Preosteoblast,

Introduction Glycation of long lived extracellular proteins such as collagen increases as a function of age. 19 The process of glycation occurs nonenzymatically, and involves reaction of a sugar with an amino group on a protein molecule to form an Amadori product. 19 The reactive Amadori product can then bind with amino groups on other protein molecules to form advanced glycation end products (AGEs) through intermolecular cross-links. This process is greatly accelerated in diabetic states, and AGE-modified proteins are thought to play a central role in chronic diabetic complications. 5 Type I collagen is the most abundant protein of bone, comprising about 85% of the bone organic matrix, s In addition to its structural role, accumulating evidence suggests that it plays an important role in cell-matrix interactions in bone, which may be important in the regulation of osteoblastic cell growth and differentiation. 1,9,20,22,23We have recently reported that accumulation of AGE collagen is accelerated in the bone tissue of diabetes mellitus (DM) rats in vivo. Moreover, altered cell functions were observed in primary cultures of fetal rat calvariaderived osteoblasts grown in vitro on AGE-modified type I collagen, including an altered expression of alkaline phosphatase (ALP) activity and osteocalcin, and bone nodule formation, l° Taken together, these results support the hypothesis that AGE modification of type I collagen might be important in the reduced bone formation that is characteristic of diabetic osteopenia. 2'3'15 To determine whether AGE collagen might also affect the maturation of preosteoblasts, we have used the phenotypically preosteoblastic UMR 201-10B cell line. We have previously obtained evidence that differentiation of UMR 201 cells is promoted by their growth on type I collagen. 22'23 In this study, we have compared the growth of UMR 201-10B cells on control type I collagen or AGE-modified collagen and investigated expression of pro(a 1) type I collagen (1PC), ALP, and osteopontin (OP), which represent parameters of osteoblastic differentiation. Takeuchi et al.2° have recently reported that differentiation of osteoblasts is dependent on production of type I collagen, and that this is mediated by a reduced expression of the transforming growth factor beta (TGF-13) receptor. 2° Therefore, we have investigated the involvement of the TGF-[3 receptor, in the action

Address for correspondence and reprints: Dr. David M. Findlay,Depart-

ment of Orthopaedics and Trauma, Royal Adelaide Hospital, North Terrace, Adelaide 5005, South Australia, Australia. E-mail: dfindlay@ medicine.adelaide.edu.au *Present address: Department of Orthopaedics and Trauma, Royal Adelaide Hospital, Adelaide, South Australia,Australia. © 1997by ElsevierScienceInc. All rightsreserved.

© 1997 by Elsevier Science Inc. All rights

237

8756-3282/97/$17.00 Pn $8756-3282(97)00128-2

238

Y. Katayamaet al. Glycated collagen and preosteoblasts

of type I collagen on UMR 201-10B cells, using cross-linking analysis to assess changes in the receptor binding. Materials and Methods

Materials [J25I]TGF-[31 was purchased from Dupont/NEN, human recombinant TGF-[31 from Life Technologies (Victoria, Australia), and glucose-6-phosphate (G6P) from Sigma (St. Louis, MO). Rat OP cDNA was a gift from Dr. L. Fisher (National Institutes of Dental Research, NIH, Bethesda, MD) (K26); rat ALP cDNA was from Dr. G. Rodan (Merck, Sharp and Dome, West Point, PA) (k27); and rat 1PC cDNA was from Dr. J. Bateman (Royal Children's Hospital, Victoria, Australia). All chemicals were purchased from BDH (Kilsyth, Victoria, Australia) unless otherwise specified.

Cell and Cell Culture The rat preosteoblast cell line, UMR 201-10B, was used in all experiments and was grown in culture as previously described, z7 UMR 201-10B cells are an immortalized, phenotypically similar variant of UMR 201 cells, z7 Cells were maintained in ~x-modified minimum essential medium (MEM) (a-MEM) (Gibco Laboratories, Grand Island, NY), containing 10% fetal bovine serum (FBS), with the addition of 200 I~g/mL G418 to maintain selection of the neomycin-resistantclonal cell line. G418 was not included in cells prepared for experiments. All media contained gentamicin (80 rag/L) (DBL, Mulgrave, Victoria, Australia) and minocycline (1 mg/L) (Sigma Chemical Company, St. Louis, MO). Cells were routinely used between passages 35 and 45.

Preparation of AGE-Modified Type I Collagen Preparation of AGE-modified type I collagen (Collaborative Research, Waltham, MA) was performed as previously described, l° Briefly, 10 cm or six well multiwell culture dishes, respectively, were precoated with rat tail tendon type I collagen at 50 ixg/cm2 and incubated with 0.5 mol/L G6P in 0.2 mol/L phosphate buffered saline (PBS) at 37°C in a sterile condition for 1, 3, or 5 weeks. After each incubation time, the buffer was removed, the dishes were extensively washed with PBS, and then left with PBS for the remainder of the incubation period (a total of 5 weeks). Collagen-coated dishes incubated in PBS alone for 5 weeks were used as controls. AGE modification of collagen was quantified by measuring relative fluorescence, as previously described, lOBriefly, AGE specific fluorescence-emission spectra at 370 nm excitation wavelength and 440 nm emission wavelength were determined for collagenase digested samples using a fluorimeter (F-850, Hitachi Industrial Co. Ltd., Tokyo, Japan). The quantity of AGE was expressed as relative fluorescence per mg collagen. 7

Bone Vol. 21, No. 3 September 1997:237-242

Assay c~[ALP Activity Osteoblasts were plated at l × l05 cells/mL onto six well collagen-coated dishes and grown in culture media for 48 h. The cell layer was then extracted by sonication for 30 s in a 10 mmol/L Tris/HCl buffered solution (pH 7.4). Total DNA content was measured fluorometrically using Hoechst dye 33238 (American Hoechst Corp., Somerville, NJ). ALP activity was determined by the method of Lowry 13 and expressed as nmol/min/p~g DNA.

Northern Blot Analysis Ceils were subcultured onto 10 cm tissue culture dishes that were either uncoated or coated with the different substrata. Cells were plated at the appropriate density to reach confluence at the time of RNA harvesting. Total RNA was extracted as previously described, z2 20 Ixg total RNA was loaded per lane of 1.2% agarose-formaldehyde gels and transferred to nylon filters (Amersham, International, Buckinghamshire, UK). 16 Complementary DNA probes were random prime labeled with [3Zp]c~-dCTPto a specific activity of 1 × 109 dprn/~g DNA according to the manufacturer's specifications (Boehringer Mannheim, GmbH, Mannheim, Germany). A specifically bound probe was visualized by autoradiography and signals representative of mRNA levels were quantified by Phosphorimager analysis (Molecular Dynamics, Sunnyvale, CA).

Affinity Cross-linking Analyses of Cell Surface TGF-[3 Receptors Affinity cross-linking analysis of cell surface TGF-[3 receptors were determined as previously described, z° Briefly, cells in 10 cm culture dishes, coated or uncoated with different substrata, were washed twice with ice-cold PBS, and 100 pM [IZSI]-TGF[31 (total radioactivity of approximately 300,000 cpm) was added in the binding buffer (128 mmol/L NaC1, 5 mmol/L KC1, 5 mmol/L MgSO4, 1.3 mmol/L CaCIz, and 25 mmol/L HEPES, pH 7.4) containing 0.3% BSA. Nonspecific binding was estimated in the presence of 10 nmol/L unlabeled TGF-[31. After 4 h of incubation at 4°C, cells were washed three times with ice-cold PBS and bound [J25I]-TGF-[31 was chemically cross-linked with its receptors by addition of 0.3 mmol/L disuccinimidyl suberate at 4°C for 20 min. After several washes, cells were harvested in PBS containing 1 mmol/L EDTA, 10 mmol/L N-ethylmaleimide, 1 mmol/L phenylmethylsulfonylfluoride, and 10 ixg/mL pepstatin A. Cross-linked TGF-[3 receptors with [125I]-TGF-[31 were solubilized with 0.1 mol/L Tris-HC1, pH 7.4, 0.5% Nonidet P-40 containing protease inhibitors and were resolved on SDS-PAGE. Sample loadings were adjusted to contain equal protein. Signals were visualized by autoradiography and quantified by Phosphoimager analysis (Molecular Dynamics, Sunnyvale, CA).

Statistics Cell Growth Cells were plated at 5 x 10 4 cells per well onto plastic or collagen-coated six well dishes and grown in culture media (c~-MEM containing 10% FCS), in the presence or absence of 1 txmol/L retinoic acid (RA), for 48 h. Cell numbers were determined using a Coulter counter (Coulter Electronics Ltd., Hertfordshire, England), as previously described, z2

Cell counting and ALP activity studies were performed in six wells per group and all experiments were repeated at least three times. The results are expressed as mean -~ SEM of the data from representative experiments. Intergroup comparisons were made with an analysis of variance and Dunnett t-test using a statistical package (Statview IV, Abacus Concepts Inc.). A p value of less than 0.05 was considered significant.

Bone Vol. 21, No. 3 September 1997:237-242

Y. Katayama et al. Glycated collagen and preosteoblasts

239

0.8]

12-

1-

N,S,

10-

_T_

0.5

"-r-

,<

z r~

8

0.4

ol -1-

o.3 o

~- 0.2'

4

0,1 '

4 0.0 C

A1

A3

A5

I

P

A1

A3

A5

RA10 M

RA 10"SM

Figure 1. Effect of AGE collagen on cell growth of UMR 201-10B cells, with and without additio~ of 1 p~mol/LRA. Cells were plated at 5 × 1 0 4 cells/well onto plastic or collagen-coated six well dishes and grown for 48 h. P: plastic. C: control type 1 collagen. A1, A3, A5: glycated type I collagen. Results are mean -4- SEM of six determinations. **p < 0.01 vs. plastic. This experiment has been repeated three times with similar results. Results

Treatment of type I collagen with G6P for up to 5 weeks increased the amount of AGE in a time-dependent manner, as indicated by the increase in relative fluorescence, and as previously described. 1° Thus, collagen-coated dishes after 1, 3, or 5 weeks of incubation contained 16.2 - 10.6, 31.2 ~ 11.2, 68.2 -411.1 relative fluorescence units/mg protein, respectively. These increasingly modified collagen preparations were then used to examine dose-dependent effects, with respect to AGE, on UMR 201-10B cells. The control collagen significantly inhibited the growth of UMR 10B cells compared to cells on plastic (p < 0.01) (Figure 1). These inhibitory effects were abolished by glycation of collagen, whether glycation was for 1, 3, or 5 weeks, so that growth on none of these modified substrates was different from that of cells on plastic. In RA-treated cells, the control collagen had no significant inhibitory effect compared to plastic (Figure 1). In control cells, ALP activity was very low and no significant changes were obserw'.d by growth on the control or glycated collagen. In contrast, RA treatment of cells greatly stimulated ALP activity, as previously repoited. 27 Control collagen significantly inhibited the degree of RA-induced upregulation of ALP activity in UMR 201--10B cells, compared to cells on plastic. These inhibitory effects were progressively reduced by glycation of collagen as a funcl:ion of the extent of glycation. The ALP activity in cells plated on 3 and 5 weeks glycated collagen was significantly greater than in cells on the control collagen (p < 0.01); 5 weeks glycation completely abolished the inhibitory effect of collagen compared to ALP activity in cells on plastic alone (Figure 2). Figure 3 demonstrates the effects of AGE collagen on the expression of three genes associated with the mature osteoblastic phenotype, which are induced by RA treatment of UMR 201-10B cells. As we have previously described, 22'23"27 RA treatment increased the expression of mRNA encoding 1PC, ALP, and OP, although to different extents, in these cells. When we examined the effect of plating cells onto type I collagen, no consistent substrate-induced changes, compared to cells on plastic, were observed in the 1PC or ALP mRNA expression (Figure 3a). In

Figure 2. Effect of AGE collagen on ALP activity in UMR 201-10B cells, with and without addition of 1 p,mol/L RA. Cells were plated at 1 × 105 cells/well onto plastic or six well dishes coated with control or AGE collagen and grown for 48 h. Values were adjusted for DNA concentration, as described in Materials and Methods. *p < 0.05, **p < 0.01 vs. plastic, tp < 0.05, ~p < 0.01 vs. control collagen. Abbreviations used are the same as those in Figure 1. This experiment is representative of three experiments in which similar results were obtained. contrast, in control cells, unmodified collagen increased the OP mRNA expression, compared to cells on plastic (Figure 4). This collagen-mediated effect was inhibited by glycation of collagen. This pattern was obscured in RA-treated cells, in which no significant effects were observed by collagen or glycated collagen on the OP mRNA expression, compared with cells on plastic. In a different osteoblastic cell line, MC3T3-E1 cells, Takeuchi et al. have obtained evidence to suggest that collagenmediated effects are secondary to the changed TGF-[3 receptor expression. 2° We therefore investigated whether the changes in the properties of UMR 201-10B cells grown on the control collagen might also be associated with the altered cell surface expression of TGF-[3 receptors. The receptor expression was determined by chemical cross-linking of the cell surface bound [12sI]-TGF-[31 to its receptors, as described by Takeuchi et al. 2° As shown in Figure 5a, affinity cross-linking studies have revealed that type I, type II, and type III TGF-[3 receptors are abundantly present in UMR 201-10B cells, as observed in other osteoblastic cells. 21 Type I and type III receptor binding was similar on each of the cell substrates. In contrast, the control collagen dramatically inhibited TGF-[3 type II receptor binding relative to plastic, as shown in Figure 5. Glycation of the collagen strongly inhibited its activity, in a dose-responsive manner, with respect to AGE content. In RA-treated cells, no significant differences were observed between any of the cell groups. Discussion

This study extends our earlier work, which indicated that the expression of a number of osteoblast phenotypic markers is suppressed in primary osteoblastic cells derived from fetal rat calvaria, grown on AGE collagen, when this is compared with growth on unmodified collagen, lO Furthermore, the expression of a number of osteoblastic genes was modulated in UMR 201 cells by growth on type I collagen as well as the responsiveness of these cells. 22'23 Therefore, we hypothesized that AGE modification of type I collagen might alter the ability of this substrate to promote the expression of osteoblastic differentiation, as assessed by the expression of osteoblast markers. Here we have

240

Y. Katayama et al. Glycated collagen and preosteoblasts

P

A RA

C

A1

A3

A5

-

P +

C

Bone Vol. 21, No. 3 September 1997:237-242

A1

A3

A5

+

+

+

+

~1(,)

O

ALP

OP R A 10"SM

GAPDH

B

Figure 4. Effect of AGE collagen on mRNA expression of OP in UMR 201-10B cells, with and without addition of 1 p~mol/L RA. Cells were cultured for 24 h on plastic, control collagen, and AGE collagen. Signals are expressed as the ratio of OP mRNA to GAPDH mRNA, where the result for untreated cells on plastic was given the value of 1. The values represent means of pooled data from three independent experiments and are expressed as mean +_ SEM. **p < 0.01 vs. plastic. Abbreviations used are the same as those in Figure 1.

2

o~I(I)/GAPDH 1 P

C A1 A3 A5

RA C

P

C A1 A3 A5

+

+

+

+

+

P

C A1 A3 A5

+

+

2

ALP/GAPDH 10t P RA

C A1 A3 A5

+

+

+

Figure 3. Effect of AGE collagen on gene expression in UMR 201-10B cells, with and without addition of 1 p~mol/LRA. Cells were cultured for 24 h on plastic, control collagen, and AGE collagen. Shown is a Northern blot with signals corresponding to mRNA for pro(a) type I collagen (cd(1)), OP, ALP, and the housekeeping gene used for normalization, glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Signals are expressed as the ratio of cd(I) or ALP to GAPDH mRNA, where the result for untreated cells on plastic ~1(I) or treated cells on plastic (ALP) was given the value of 1. This experiment has been repeated three times and consistent results have been obtained. Abbreviations used are the same as those in Figure 1. shown that in an immortalized variant of U M R 201 cells, phenotypically preosteoblastic U M R 201-10B cells, growth on collagen-reduced cell growth and TGF-[3 type II receptor binding, and increased osteopontin mRNA expression is relative to growth on plastic. AGE modification reduced or abolished the ability of the collagen to alter the above parameters. These data are consistent with the notion that collagen induces differentiation of preosteoblasts, as we and others have previously indicated, 2°'22'23 and that AGE modification of collagen interferes with its ability to do so. When U M R 201-10B cells were treated with RA, which promotes a more mature osteoblastic phenotype in this cell line, differences between the cells grown on different matrices were less apparent. However, in the case of ALP activity, which is only expressed in U M R 201-10B cells following RA treatment, there was a reduction by collagen but not by AGE collagen, except at low AGE concentration. These changes in ALP activity were not observed at the level of mRNA. This dissociation between ALP activity and mRNA levels has been observed elsewhere. 12 The data described here show that in every circumstance where collagen was effective in altering the properties of these cells, AGE modification reduced or abolished its ability to do so.

TGF-[3 plays an important role in bone formation and is a potent stimulator of bone matrix protein synthesis, including that of type I collagen. 6 Although TGF-[3 induces bone formation in vivo, its effects on the growth and the expression of differentiation parameters in osteoblastic cells depend upon the cell system studied. 14'25 In U M R 201-10B cells, addition of 1 ng/mL of TGF-[31 strongly inhibited OP mRNA expression, both in cells grown on plastic and on control collagen, and suppressed the RA-induced upregulation of ALP activity (data not shown). Takeuchi et al. 2° have recently reported that type I collagen enhanced the differentiation of another osteoblastic cell line, MC3T3-E 1, and concomittently suppressed the actions of TGF-[3 by reducing the receptors competent to bind TGF-[3. These authors proposed that the change in the responsiveness to TGF-[3 allowed these cells to "escape" from the inhibitory effect of TGF-[3 on osteoblastic differentiation and to further differentiate into more mature osteoblastic cells. In the present study, cells grown on control collagen displayed greatly decreased TGF-[3 type I! receptor binding. However, RA treatment, which also stimulated osteoblastic differentiation, did not change TGF-[3 receptor binding compared with untreated cells and abolished the substrate-induced changes seen in the absence of RA. These differences in the presence and absence of RA are not understood at present but are likely to reflect the different intracellular pathways by which collagen and RA modulate the maturational state of U M R 201-10B cells. There are at least two mechanisms to account for the observations presented here. Cellular interaction with extracellular matrix molecules, including collagen, occurs via binding to cellular receptors, the best understood being the integrin receptor family. 17'26 It is possible that AGE modification of collagen interferes with these specific interactions, leading to the reduced or abolished collagen-mediated effects seen here. An alternative mechanism may be interaction between AGE and its specific receptor(s), the best characterized of which is the receptor for AGEs (RAGE), a member of the immunoglobulin superfamily.11,18 These receptors have been identified on endothelial cells, smooth muscle cells, and mononuclear phagocytes. 4 However, using immunohistochemical analysis, we have found only very weak or absent expression of RAGE on U M R 201-10B cells in culture (not shown). One of the principal consequences of AGERAGE interaction is the induction of cellular oxidant stress, which can be blocked by antioxidant agents, such as vitamin E and probucol. 24 In preliminary experiments, neither vitamin E

Bone Vol. 21, No. 3 September 1997:237-242

Y. Katayama et al. Glycated collagen and preosteoblasts

A P R~

.

C .

.

A1 A3 A5 .

.

.

P

C

+

+

A1 A3 A5 +

+

÷

MW (kDa)

TYPE 3

200 TYPE 21

97.4

TYPE 1

66.2

241

differentiative effects o f type I collagen on preosteoblasts, suggesting in turn that glycation o f collagen m a y interfere with specific interaction(s) between U M R 201-10B cells and type I collagen. If these events also occur in diabetic bone, A G E modified collagen m a y have reduced ability to promote the maturation of preoseoblasts and the appropriate activity o f mature osteoblasts.

Acknowledgments: This work was supported by grants from the National Health and Research Council of Australia. D.M.F. was a Senior Research Fellow of the National Health and Research Council of Australia. Y.K. is supported by a travel fellowship from National Defense Medical College. The authors gratefully acknowledge the immunohistochemical analysis of RAGE performed by Dr. Tina Soulis.

B

References

Ratio/plastic 1.2

1. Andilanarivo, A. G., Robinson,J. A., Mann, K. G., and Tracy, R. P. Growth on type I collagen promotes expression of the osteoblastic phenotype in human osteosarcoma MG-63 cells. J Cell Physiol 153:256-265; 1992. 2. Bouillon, R. Diabetic bone disease. Calcif Tissue Int 49:t55-160; 1990. 3. Bouillon,R., Bex, M., Van Herck, E., Laureys, J., Dooms, L., Lesaffre, E., and Ravussin, E. Influence of age, sex, and insulin on osteoblast function: Osteoblast dysfunctionin diabetes mellitus. J Clin EndocrinolMetab 80:1194-1202; 1995. 4. Brett, J., Schmidt, A. M., Yan, S. D., Zhou, Y. S., Weidman, E., Pinsky, D., Nowygrod, R., Neeper, M., Przysiecki, C., and Shaw, A. Survey of the distribution of a newly characterized receptor for AGEs in tissue. Am J Pathol 143:1699-1712; 1993. 5. Brownlee, M. Glycation and diabetic complications. Diabetes 43:836-841; 1994. 6. Centrella, M., Horowitz, M. C., Wozney, J. M., and McCarthy, T. L. Transforming growth factor-13 gene family members and bone. Endocr Rev 15:2739; 1994. 7. Crowley, S. T., Brownlee, M., Edelstein, D., Satriano, J. A., Moil, T., Singhal, P. C., and Schlondorff, D. O. Effects of nonenzymatic glycosylation of mesangial matrix on proliferation of mesangial cells. Diabetes 40:540-547; 1991. 8. Eyer, D. R. Collagen: Molecular diversity in the body's protein scaffold. Science 207:1315-1322; 1980. 9. Ikeda, K., Michelangeli, V. P., Martin, T. J., and Findlay, D. M. Type I collagen substrate increases calcitonin and parathyroid hormone receptor-mediated signal transduction in UMR 106-06 osteoblast-like cells. J Cell Physiol 156:130137; 1993. 10. Katayama, Y., Akatsu, T., Yamamoto, M., Kugai, N., and Nagata, N. Role of nonenzymatic glycosylation of type I collagen in diabetic osteopenia. J Bone Miner Res 1996; 11:931-937; 1996. 11. Khoury, J., Thomas, C., Loike, J., Hickman, S., Cat, L., and Silverstein, S. Macrophages adhere to glucose-modified basement membrane via their scavenger receptors. J Biol Chem 269:10197-10200; 1994. 12. Kiledjian, M. and Kadesch, T. Post-transcriptional regulation of the human liver/bone/kidney alkaline phosphatase gene. J Biol Chem 266:4207-4213; 1991. 13. Lowry, O. Micromethods for the assay of enzyme. II. Specific procedures. Alkaline phosphatase. Meth Enzymol 4:371-372; 1955. 14. Noda, M. Transcriptionalregulation of osteocalcin production by transforming growth factor-J3in rat osteoblast-likecells. Endocrinology 124:612-617; 1989. 15. Rico, H., Hernandez, E. R., Cabranes, J. A., and Comez-Castresana, F. Suggestion of a deficient osteoblastic function in diabetes mellitus. Calcif Tissue Int 45:71-73; 1989. 16. Sambrook,J., Filtsch, E. F., and Maniatis, T. Molecular Cloning: A Laboratory Manual, 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory; 1989. 17. Schiro, J. A., Chan, B. M. C., Roswit, W. T., Kassner, P. D., Pentland, A. P., Hemler, M. E., Eisen, A. Z., and Kupper,T. S. Integiln a213~ (VLA-s)mediates reorganization and contraction of collagen matrices by human cells. Cell 67:403-410; 1991. 18. Scbmidt, A. M., Hori, O., Brett, J., Yan, S. D., Wantier, J. L., and Stern, D. Cellular receptor for AGEs. Arterioscler Thromb 14:1521-1528; 1994. 19. Schnider, S. L. and Kohn, R. R. Effect of age and diabetes mellitus on the

1

m

0.8 0.6 0.4 0.2 0 P RA 10"6M

C

A1

A3 A5

]

P

C

A1

A3

A5

-I-

Figure 5. Effect of AGE collagen on cell surface expression of TGF-13 receptors. Ceils in 10-cra culture dishes, coated or uncoated with control or AGE collagen, were cultured for 24 h prior to incubation with [lZSI]-TGF-131 and chemical cross-linking, as described in Materials and Methods. Cells were then solubilized and proteins resolved on SDSPAGE. Signals were visualized by autoradiography and quantified as described in Materials and Methods. A: Shown is an affinity crosslinking analysis with signals corresponding binding to the type 1, 2, or 3 TGF-13 receptor. B: Quantification of type 2 receptor binding. Values were expressed as the ratio relative to cells on plastic in the absence of RA. This experiment has been repeated three times and consistent results were obtained.

nor probucol altere6 the effects of A G E collagen on U M R 201-10B cells observed in this study. A l t h o u g h these results do not eliminate the possible involvement of A G E - R A G E interactions in those effects, oxidative stress does not appear to be important in the prese,nt context. Taken together, these results are m o r e consistent with A G E modification o f collagen interfering with collagen-receptor interactions on these cells than with R A G E - m e d i a t e d effects. In s u m m a r y , we have demonstrated that growth on control collagen has a n u m b e r of effects on immortalized U M R 201-10B cells, which w e have previously described for the nontransformed U M R 201 cells. These effects are upregulation o f OP m R N A expression, inhibition o f cell growth, and inhibition of RA-induced alkaline phosphatase, effects that we have previously argued are consistent with p r o m o t i o n by collagen o f a m o r e differentiated phenotype. 22 The above collagen-mediated effects are closely correlated with reduction of TGF-[3 type II receptor binding. A G E modification was found to interfere with these

242

20.

21.

22.

23.

24.

Y. K a t a y a m a et al. Glycated collagen and preosteoblasts

solubility and nonenzymatic glycosylation of human skin collagen. J Clin Invest 67:1630-1635; 1981. Takeuchi, Y., Nakayama, K., and Matsumoto, T. Differentiation and cell surface expression of transforming growth factor-[3 receptors are regulated by interaction with matrix collagen in murine osteoblastic cells. J Biol Chem 271:3938-3944; 1996. Takeuchi, Y., Fukumoto, S., and Matsumoto, T. Relationship between actions of transforming growth factor-13 and cell surface expression of its receptors in clonal osteoblastic cells. J Cell Physiol 162:315-321; 1995. Traianedes, K., Ng, K. W., Martin, T. J., and Findlay, D. M. Cell substratum modulates responses of preosteoblasts to retinoic acid. J Cell Physiol 157:243252; 1993. Traianedes, K., Martin, T. J., and Findlay, D. M. Regulation of osteopontin expression by type I collagen in preosteoblastic UMR 201 cells. Connect Tissue Res 34:63-74; 1996. Wautier, J. L., Soukourian, C., Chappey, O., Wautier, M. P., Guillausseaa, P. J., Cat, R., Hori, O., Stem, D., and Schmidt, A. M. Receptor-mediated

B o n e Vol. 21, No. 3 September 1 9 9 7 : 2 3 7 - 2 4 2

endothelial cell dysfunction in diabetic vasculopathy. J Clin Invest 97:238243; 1996. 25. Wrana, J. L., Maeno, M., Hawrylyshyn, B., Yao, K. L., Domenicucci, C., and Sodek, J. Differential effects of transforming growth factor-beta on the synthesis of extracellular matrix proteins by normal fetal rat calvarial bone ceil populations. J Cell Biol 106:915-924; 1988. 26. Yamada, K. M. and Miyamoto, S. Integrin transmembrane signaling and cytoskeletal control. Current Opinion Cell Biol 7:681-689; 1995. 27. Zhou, H., Hammonds, R. G., Findlay, D. M., Fuller, P. J., Martin, T. J., and Ng, K. W. Retinoic acid modulation of mRNA levels in malignant, nontransformed, and immortalized osteoblasts. J Bone Miner Res 6:767-777; 1991.

Date Received: January 24, 1997 Date Revised: April 30, 1997 Date Accepted: May 9, 1997