Ultrastructural localization of proteoglycans in bone in ... - Springer Link

J Bone Miner Metab (2002) 20:288–293

© Springer-Verlag 2002

Ultrastructural localization of proteoglycans in bone in osteogenesis imperfecta as demonstrated by Cuprolinic Blue staining Padmini Sarathchandra1, John P. Cassella2, and S. Yousuf Ali1 1 Institute of Orthopaedics, and Musculo-Skeletal Science, Royal National Orthopaedic Hospital Trust, Brockley Hill, Stanmore, Middlesex, HA7 4LP, UK 2 Division of Biological Sciences, School of Environmental and Applied Sciences, University of Derby, Derby, UK

Abstract The role of proteoglycans in bone in osteogenesis imperfecta (OI) has been examined. Using Cuprolinic Blue staining of whole fetal bone tissue and examining the tissue in the transmission electron microscope, the presence of proteoglycans was observed. Quantitative comparative image-analysis of the proteoglycans from electron micrographs was performed, with measurement of sizes and number of proteoglycan particles. A significant increase in the total number of proteoglycan particles in OI bone osteoid was observed when compared with normal, matched controls. The area of the proteoglycan particles, as measured by pixel-area, using image analysis, was also increased in OI bone osteoid. These findings further suggest a role for proteoglycans in mineral formation by the possible inhibition of mineral growth and alteration of collagen nucleation sites. The increased number and size of proteoglycan particles may be a contributing factor to the previously reported poor mineral formation with subsequent loss of bone strength, making it more prone to fracture, in OI. Key words type I collagen · proteoglycan · osteogenesis imperfecta · image analysis · Cuprolinic Blue staining

Introduction The interactions of proteoglycans (PGs) and collagen have remained speculative for some time, partly because of the unresolved problem of PG translocation and the collapse of PG domains during the processing of tissue (Scott [1]). The development of Cupromeronic and Cuprolinic Blues, two dyes designed for electron histochemical use in a “critical electrolyte concentra-

tion” system, led Scott and coworkers [2,3] to demonstrate the specific interactions of PGs and collagen type I fibrils. It is generally believed that PGs may be involved in matrix formation and in the regulation of mineralization (Boskey [4]; Fisher [5]; Fisher and Termine [6]). It has been suggested that PGs also play a role in controlling collagen fibril diameter (Scott and Parry [7]), and recent evidence supports an important role in fibrillogenesis (Parry et al. [8]; Kadler et al. [9]). Blumenthal et al. [10] emphasised the role of PGs in bone mineralization; it was reported that a decrease in PG content was accompanied by the deposition of mineral in cartilage. Initial calcification probably follows the removal of inhibitory compounds such as PGs (Ali [11]). PG abnormalities have been suspected in osteogenesis imperfecta (OI) for some time; for example, both Bleckmann et al. [12] and Engfeldt and Hjerpe [13] reported increased amounts of glycosaminoglycans (GAGs) in OI type II bone biopsies (old classification, OI congenita; new, type II), when compared with normal bone. Brown et al. [14] also observed an abnormal distribution of GAGs in OI dentine, and Spencer [15] showed histochemical evidence of abnormal GAGs in OI bone. Goldberg [16] found increased excretion of PG in young patients with OI. In the present study, the total amount and the size of PG particles present in the osteoid of fetal bone and normal bone was examined at the ultrastructural level, using Cuprolinic Blue staining.

Materials and methods

Offprint requests to: P. Sarathchandra Received: August 28, 2001 / Accepted: February 8, 2002

Bone specimens from seven OI patients and three normal controls were processed for Cuprolinic Blue staining to demonstrate the PG distribution patterns between the osteoid collagen fibers (Table 1).

P. Sarathchandra et al.: Proteoglycans and osteogenesis imperfecta

289

Table 1. Specimen details Patient

Age OI or Normal OI type Site of bone

1

2

3

4

5

6

7

8

9

10

Fetal 22/40 wks OI II Rib

Fetal 19/40 wks OI II Femur

Fetal 24/40 wks OI II Femur

Infant 40/40 wks OI II/III Femur

Fetal 22/40 wks OI II/III Tibia

Fetal 22/40 wks OI II/III Femur

3 yrs

13 yrs

Fetal

Fetal

OI III Femur

Normal n/a Tibia

Normal n/a Femur

Normal n/a Femur

wks, weeks; OI, osteogenesis imperfecta

All human bone material was obtained through collaborations with clinical colleagues. Local ethics committee approval was obtained before the study took place. Tissue has been collected over many years to build up this unique collection of material. Normal control bone was obtained, with informed consent, from surgical procedures, usually limb amputations. Bone specimens from OI patients were obtained during rodding procedures (patient 7) or at postmortems (patients 1–6). The diagnostic criteria for the determination of the clinical types of OI was made on the basis of the accepted, published literature and clinical experience of the hospital staff. Classification of OI type II and II/III was based on X-ray, clinical signs, and collagen biochemistry from cultured fibroblasts. The three bone specimens categorized as type II/III were actually an infantile form of type III (M. Pope, personal communication). The specimens processed for Cuprolinic Blue staining were fixed in 2.5% glutaraldehyde in 25 mM sodium acetate buffer (pH 5.6), containing 0.1 M magnesium chloride and 0.5% Cuprolinic Blue dye for 18 h, as described by Scott and Orford [17]. The unbound dye was removed by rinsing for 15 min in several changes of buffer solution, followed by further en-bloc staining for 15 min in 0.5% aqueous sodium tungstate. The specimens were dehydrated in a graded ethanol series (first alcohol contained 0.5% sodium tungstate), and embedded in araldite CY212 resin (Agar Scientific, Stansted, Essex, UK). Ultrathin sections were stained with 2% aqueous uranyl acetate and lead citrate for 10 min in each solution, before viewing in a Philips CM12 electron microscope (Philips, FBI UK, Cambridge, UK). Random areas of osteoid were photographed in the transmission electron microscope and prints made to a standardized total magnification of 75 000. The image analysis system could not clearly differentiate the grayscale difference between PG particles and other matrix components, so the PG particles were “traced” onto acetate sheets and the number of PGs and the area of each particle (µm2) could be analyzed. The PG particles were measured automatically on a Centron image analysis system (Centron, Derby, UK).

Fig. 1. Ultrathin araldite section of osteoid collagen from a normal fetal bone en-bloc stained with Cuprolinic Blue (CB). The section shows few CB-positive rods of proteoglycan between collagen fibers. Proteoglycan (arrowheads). ⫻67 500

Results Cuprolinic Blue staining to localize PGs in osteoid The PG-collagen interactions demonstrated by Cuprolinic Blue staining varied between OI and normal bone. There was an orderly pattern of distribution, but a small number of PGs between the collagen fibers in normal fetal bone osteoid (Fig. 1). In OI type II bone, there was a higher content of PG between the osteoid collagen fibrils (Fig. 2) and, furthermore, there was a disturbed pattern of PG distribution in relation to collagen fibers. Normal juvenile bone demonstrated a small number of PG-positive rods (Fig. 3) between the collagen fibers. In OI type II/III, which demonstrated thicker collagen fibrils than OI type II fibrils, there were long PGpositive rods between the collagen fibers, often extending over two D-periods (Fig. 4) and arranged in parallel with the collagen fibers (Fig. 5). Image analysis of PGs in osteoid The numbers of PG particles and the pixel-area (in µm2) of each PG particle in OI and normal fetal bone speci-

290


Fig. 2. Ultrathin araldite section of osteoid collagen from osteogenesis imperfecta (OI) type II fetal bone en-bloc stained with Cuprolinic Blue (CB). The section shows abundant CB-positive rods of proteoglycan between collagen fibers. Proteoglycan (arrowheads). ⫻67 500

Fig. 4. Ultrathin araldite section of osteoid collagen from OI type II/III bone en-bloc stained with Cuprolniic Blue (CB). The section shows very few CB-positive rods of proteoglycan extending over two D-periods of collagen fibers. Proteoglycan (arrowheads). ⫻67 500

Fig. 3. Ultrathin araldite section of osteoid collagen from normal juvenile bone en-bloc stained with Cuprolinic Blue (CB). The section shows very few CB-positive rods of proteoglycan between collagen fibers. Proteoglycan (arrowheads). ⫻67 500

Fig. 5. Ultrathin araldite section of osteoid collagen from OI type III bone en-bloc stained with Cuprolinic Blue (CB). The section shows long CB-positive rods of proteoglycan between collagen fibers. Proteoglycan (arrowheads). ⫻67 500

mens were determined, and graphs of the data were plotted (see Figs. 6 and 7). Statistical analysis was performed on the raw data obtained from the image analysis system. However, due to the small number of specimens, the statistical results were inconclusive. There was not a statistically significant difference between the input groups. This is due to the small number of bone samples; however, the data obtained are still of importance in offering a better understanding of the role and relationship of collagen and proteoglycans in OI. In the present study, the PGs seen between the collagen fibrils in the osteoid of OI bone have not been identified either by immunolocalization or by digestion with chondroitinase ABC or ACII followed by

Cuprolinic Blue staining. (Chondroitinase ABC will digest both chondroitin and dermatan sulfates, but ACII will digest only chondroitin sulfates.)

Discussion In this study, we used a Cuprolinic Blue stain to demonstrate, at an ultrastructural level, PG particles in OI bone. A comparison with normal bone osteoid PG particles showed a quantitative increase in the number of PG particles in bone and an increase in the overall area of these PG particles in bone. To our knowledge, this is the first report to demonstrate the quantitative difference between OI and normal bone.


291

Fig. 6. A graph showing the variation in the number of proteoglycan (PG) particles in the bone specimens examined. Note the increased number of PG particles present in OI bone specimens compared with normal bone

Fig. 7. A graph showing the variation in the size of PG particles in the bone specimens examined. Note the slight apparent increase in the size of PG particles present in OI bone specimens compared with normal bone

Subnormal amounts of glycosaminoglycans (GAGs) were isolated from cortical bone tissue in OI [18]. It was postulated from the study that either a qualitative or quantitative defect of GAGs could hamper normal mineralization. Urinary PG levels in OI patients have been tested, and the levels were found to be especially high in younger OI patients (Goldberg [16]). These findings suggested that there might be alterations of PGs, particularly in fetal bone, which required further investigation and explanation. The abundant PGs in OI bone observed in this current study may affect mineralization, matrix formation, and, indeed the controlling mechanisms involved in col-

lagen fibril diameter, all of which are likely to affect the characteristics of bone, making it more susceptible to fracture, as seen in severe types of OI. Although previous studies have demonstrated small PGs (PG 100, decorin, biglycan) rather than large PGs in human fetal bone (Bianco et al. [19]; Bosse et al. [20]), the OI bone biopsies studied here contained “large” PGs, which have been previously identified by immunogold labelling with specific antibodies (Sarathchandra [21]). The immunogold labelling pattern for decorin and biglycan was rather weak when compared with the labelling intensity of the large PGs— keratan sulfate, chondroitin 4-sulfate, and chondroitin

292


6-sulfate. Takagi et al. [22] have detected PGs containing chondroitin 4-sulfate and dermatan sulfate in the walls of osteocyte lacunae, perilacunar matrix, and canaliculi in rat bone, using biochemical and immunocytochemical methods. Such material lay adjacent to the mineral line, but was markedly less dense than that seen in our specimens. The relationship between PGs and biological mineralization has been well documented, and many different in-vitro studies agree that PGs inhibit hydroxyapatite formation (Bowness and Lee [23]; Cuervo et al. [24]; Chen and Boskey [25]). The ultrastructural study by Sarathchandra et al. [21] clearly demonstrates excessive PGs in the osteocyte lacunae of OI type II/III fetal bone with simultaneous poor mineralization; the latter would undoubtedly predispose to excessive bone fragility and thinning. In particular, the osteoid-like collagen band is devoid of mineral. Vetter et al. [26] reported that, in OI, the small PG, decorin, was not significantly reduced compared with normal bone, while other noncollagenous proteins were decreased and may contribute to the fragility in OI by interfering with mineralization and normal tissue architechture. De Luca et al. [27] found that patients with a more severe clinical picture demonstrated a lower content of PGs with the highest molecular weight. Similar clinical phenotypes generally reflect the nature and localization of the type I collagen mutation [28]; however, in some cases, the clinical phenotypes are different. This variable expression may occur as the result of abnormalities in other connective tissue proteins. If different forms of PG are absent, as appears may be the case in this study, then it may be speculated that this could contribute, in part, to the wide clinical spectrum. Modifications of PG metabolism in the dermis of patients with nonlethal OI have been reported [29]. Crystals were found to form in the dermis, possibly in response to an abnormal collagen-PG interaction. Work by Scott et al. [30] reinforced the importance of PG-collagen interactions in both normal and cultured human fibroblasts. Nogami and Oohira [31] found a 95% decrease in the number of PG granules associated with the cross-banding of collagen fibrils in a patient with severe fragility, while there was no decrease in the number of PG granules in a patient with a “predominantly bowing deformity” of the bones. They discussed that, hypothetically, the separation of PGs from collagen fibrils may be associated with an increased bone fragility and, furthermore, that the mechanism of bowing could be different from the mechanisms resulting in increased fragility. A depletion of bone PG in bovine OI (Australian variant) was reported [32] and it was suggested that the

quantitation of noncollagenous proteins and PGs may, in fact, be a useful technique for differentiating clinically similar syndromes. In a previous report by the same group (Termine et al. [33]), PGs were decreased in the Texas bovine variant, suggesting hard-tissue matrix protein deletions, possibly due to an underlying impairment of cellular development. The variable expression of PGs, either a decrease (De Luca et al. [27]; Termine et al. [33]; Fisher et al. [32]; Nogami and Oohira [31]; Fedarko et al. [34]) or an increase, as is reported here, in fetal OI bone specimens, reflects an underlying imbalance which undoubtedly contributes to the fragility syndrome that is OI. The variable clinical heterogeneity observed may, in part, be explained by the variable expression of all or some of the many components that interact to form bone tissue. The importance of the two connective tissue elements in the extracellular matrix is clear, the inextensible collagenous fibrils and the compression-resistant soluble polymers—the PGs, aid the bone in coping with the “usual stresses” placed upon it [35]. Alterations in the relative amounts of these extracellular matrix components, as shown in this report with respect to PG, would undoubtedly be reflected in stress resistances in OI, making the bone more prone to fracture. As previously reported, the ratio of calcium to phosphorus (Ca/P), as determined by X-ray microanalysis in a transmission electron microscope study in OI bone, was altered, and was lower than that in normal bone (Cassella et al. [36]). When OI bone was categorized into subtypes, the ratio was lowest in OI type II, the lethal form of OI (Sarathchandra et al. [37]). Even the small number of OI type II specimens used in the present study demonstrated a higher number of PG particles than the normal bone; this may indicate the level of effect of PG on mineralization. It has been demonstrated that the mean diameter of type I collagen fibrils in OI bone, examined using transmission electron microscopy, was reduced compared with that in normal controls, and the smallest mean diameter was demonstrated in OI type II (Sarathchandra et al. [38]). All these findings indicate the important role played by the abundant PG in OI bone in mineralization, matrix formation, and the control of collagen fibril diameter, all of which are likely to affect the characteristics of bone and make it more susceptible to fracture, as seen in severe types of OI. This is the first conclusive ultrastructural evidence of the presence of increased amounts of PG in bone in association with type I collagen. The literature on PGs and OI is occasionally conflicting, and further studies are required to better understand how the known mutation in the collagen gene exerts its effects on other moieties required for normal bone strength.


Acknowledgments. The authors would like to thank Action Research for funding this work.

References 1. Scott JE (1985) Proteoglycan histochemistry. The partnership with analytical biochemistry. Coll Relat Res 5:541–575 2. Scott JE, Orford CR, Hughes EW (1981) Proteoglycan-collagen arrangements in developing rat tail tendon. Biochem J 195:573– 581 3. Scott JE, Haigh M, Nusgens B, Lapierre CM (1989) Proteoglycan: collagen interactions in dermatosparactic skin and tendon. An electron histochemical study using cupromeronic blue in a critical electrolyte concentration method. Matrix 9:437–442 4. Boskey AL (1984) Bone structure, composition, and mineralization. Orthop Clin North Am 15:597–618 5. Fisher LW (1985) The nature of proteoglycans of bone. In: Butler TW (ed) The Chemistry and Biology of Mineralised Tissues, EBSCO Media, Birmingham, Alabama, pp 188–196 6. Fisher LW, Termine JD (1985) Purification of the non-collagenous proteins from bone: technical pitfalls and how to avoid them. In: Ornoy A, Harell A, Sela J (eds) Current Advances in Skeletogenesis. Elsevier Science Amsterdam, p 467 7. Scott JE, Parry DAD (1992) Control of collagen fibril diameters in tissues. Int J Biol Macromol 14:292–293 8. Parry DAD, Flint MH, Gillard GC, Craig AS (1982) A role of glycosaminoglycans in the development of collagen fibrils. FEBS Lett 149:1–7 9. Kadler KE, Holmes DF, Trotter JA, Chapman JA (1996) Collagen fibril formation. Biochem J 316:1–11 10. Blumenthal NC, Posner AS, Silverman LD, Rosenberg LC (1979) Effect of proteoglycan on in-vitro hydroxyapatite formation. Calcif Tissue Int 27:75–82 11. Ali SY (1992) Matrix formation and mineralisation in bone. In: Whitehead CC (ed) Bone Biology and Skeletal Disorders. Carfax Publishing, Abingdon, Chapter 2, pp 19–38 12. Bleckmann H, Kresse H, Wollensak J, Buddecke E (1971) Glykosaminoglykan und kollagenanalysen bei Osteogenesis Imperfecta. Z Kinderhelikunde 110:74–84 13. Engfeldt B, Hjerpe A (1976) Glycosaminoglycans of cartilage and bone tissue in two cases of osteogenesis imperfecta congenita. Acta Pathol Microbiol Scand section A 84:488–494 14. Brown DM, Sauk JJ, Corbin KW, Bradford DS, Wiktop CJ (1975) Glycosaminoglycans (GAG) of EDTA soluble and insoluble dentin in autosomal dominant osteogenesis imperfecta. Pediatr Res 9:312 15. Spencer AT (1962) A histochemical study of long bones in osteogenesis imperfecta congenita. J Pathol Bacteriol 83:423–427 16. Goldberg H (1978) A Study of the Urinary Mucopolysaccharides in Osteogenesis Imperfecta and Other Skeletal Disorders. M. Phil thesis, University of London, United Kingdom 17. Scott JE, Orford CR (1981) Dermatan sulphate-rich proteoglycan associates with rat tail-tendon collagen at the d-band in the gap region. Biochem J 197:213–216 18. Solheim K (1969) Osteogenesis imperfecta. Microradiographic and biochemical studies with special reference to the mucopolysaccharides (glycosaminoglycans). J Oslo City Hosp 19:193– 199 19. Bianco P, Fisher LW, Young MF, Termine JD (1993) Expression and localization of the small proteoglycans biglycan and decorin in developing skeletal and non-skeletal tissue. J Histochem Cytochem 38:1549–1559

293

20. Bosse A, Schwarz K, Vollmer E, Kresse H (1993) Divergent and co-localization of the two small proteoglycans decorin and proteoglycan-100 in human skeletal and non-skeletal tissues and tumours. J Histochem Cytochem 41:13–19 21. Sarathchandra P, Pope FM, Ali SY (1996) An ultrastructural and immunogold localization study of fetal bone in osteogenesis imperfecta. Calcif Tissue Int 58:6 435–442 22. Takagi M, Maeno M, Kagami A, Takahashi Y, Otsuka K (1991) Biochemical and immunocytochemical characterization of mineral-binding proteoglycans in rat bone. J Histochem Cytochem 39:41–50 23. Bowness JM, Lee KH (1966) Effects of chondroitin sulphates on mineralisation in vitro. Biochem J 103:382–390 24. Cuervo LA, Pita JC, Howell DS (1973) Inhibition of calcium phosphate mineral growth by proteoglycan aggregate fractions in synthetic lymph. Calcif Tissue Int 13:1–10 25. Chen CC, Boskey AL (1985) Mechanisms of proteoglycan inhibition of hydroxyapatite growth. Calcif Tissue Int 37:395– 406 26. Vetter U, Fisher LW, Mintz KP, Kopp JB, Tuross N, Termine JD, Robey PG (1991) Osteogenesis imperfecta: changes in noncollageneous proteins in bone. J Bone Miner Res 6:501–505 27. De Luca G, Tira ME, Rindi S, Salvini R, Cetta G, Castellani AA (1984) Proteoglycan modifications in cultured osteogenesis imperfecta skin fibroblasts. Ital J Biochem 33:403–412 28. Dyne KM, Valli M, Forlino A, Mottes M, Kresse H, Cetta G (1996) Deficient expression of the small proteoglycan decorin in a case of severe/lethal osteogenesis imperfecta. Am J Med Genet 63:161–166 29. Tenni R, Cetta G, Dyne K, Rossi A, Quacci D, Lenzi L, Castellini AA (1988) Type I procollagen in the severe non-lethal form of osteogenesis imperfecta. Defective pro-alpha1(I) chains in a patient with abnormal proteoglycan metabolism and mineral deposits in the dermis. Hum Genet 79:245–250 30. Scott J, Dyne KM, Thomlinson AM, Ritchie M, Bateman J, Cetta G, Valli, E (1998) Human cells unable to express decorin produced disorganized extracellular matrix lacking “shape modules” (interfibrillar proteoglycan bridges). Exp Cell Res 243:59–66 31. Nogami H, Oohira A (1988) Defective association between collagen fibrils and proteoglycans in fragile bone of osteogenesis imperfecta. Clin Orthop 232:284–291 32. Fisher LW, Denholm LJ, Conn KM, Termine JD (1986) Mineralised tissue protein profiles in the Australian form of bovine osteogenesis imperfecta. Calcif Tissue Int 38:16–20 33. Termine JD, Robey PG, Fisher LW, Shimokawa H, Drum MA, Conn KM, Hawkins GR, Cruz JB, Thompson KG (1984) Osteonectin, bone proteoglycan, and phosphophoryn defects in a form of bovine osteogenesis imperfecta. Proc Natl Acad Sci USA 81:2213–2217 34. Fedarko NS, Robey PG, Vetter UK (1995) Extracellular matrix stoichiometry in osteoblasts from patients with osteogenesis imperfecta. J Bone Miner Res 10:1122–1129 35. Scott JE (1995) Extracellular matrix, supramolecular organization and shape. J Anat 187:259–269 36. Cassella JP, Garrington N, Stamp TC, Ali SY (1995). An electron probe X-ray microanalytical study of bone mineral in osteogenesis imperfecta. Calcif Tissue Int 56:118–122 37. Sarathchandra P, Kayser MV, Ali SY (1999). Abnormal mineral composition of osteogenesis imperfecta by electron probe X-ray microanalysis on conventional and cryosections. Calcif Tissue Int 65:11–15 38. Sarathchandra P, Pope FM, Ali SY (1999). Morphometric analysis of type I collagen fibrils in the osteoid of osteogenesis imperfecta. Calcif Tissue Int 65:390–395