Bone and Mineral, 17 (1992) 269-272
Elsevier This paper was presented at the Fifth International Conference on Cell-Mediated Calcification and Matrix Vesicles, held November 16-20,1991, Hilton Head, South Carolina.
Bone Research Branch Research Associate program, National Institute of Dental Research, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, USA
Synthetic lipid vesicle (liposome) suspensions have been used to experimentallymodel many of the calcium phosphate precipitation steps observed in matrix vesicle (MV) calcification. In particular,precipitate development in liposomes can be made to preferentiallyfollow the progression seen in MV, i.e. to occur initially in intraliposomal spaces and then to expand into the surrounding suspending medium. This paper reviews results from studies by us which show that certain phospholipid (PL) constituents of the liposomal membranecan modulate this progression. Of greatest relevance to MV calcificationis the observation that phosphatidylserine and sphingomyelin, ~wirLipidsselectively enriched in MV, slow the expansion of the precipitation from inside to outside the liposome.
The first mineral to form in growth-plate cartilage occurs in association with submicron size, membrane-bound, extracellular structures known as matrix vesicles (MV) [1,2]. A unique feature of MV calcification is the sequential involvement of two distinctly different, membrane-separated fluid compartments. Mineral crystals first appear in the membrane-confined microspaces of the vesicle interior. Once established, mineralization then rapidly expands beyond the membrane barrier into the surrounding extracellular fluid space. This sequence of membrane-separated precipitation events has been experimentally modeled in artificial lipid vesicle (i.e., liposome) suspensions by encapsulating inorganic phosphate (PO,) solutions in the aqueous cores of the liposomes and then making the enclosing lipid membranes permeable to external solution Ca2’ ions with lipophilic ion carriers such as Basp’I., cid acid (X-53716) [3,43. Apatite mineral readily forms inside the liposomes, and if the external solution is supersaturated with respect to apatite, mineral eventually forms outside the liposomes as well. This latter location is seeded by intraliposomally-formed crystals which penetrate the membrane .
270 The earliest work done with this precipitation model utilized liposomes made from a 7:2:1 molar mixture of phosphatidylcholine (PC), dicetylphosphate (DCP), and cholesterol (Chol) [3,4]. This choice of lipids was fortunate in that subsequent studies [5-71 with other lipid combinations showed that precipitation (both inside and outside the liposome) occurred most rapidly with this particular lipid combination. Adding and/or substituting other lipids often reduced and/or delayed precipitation to varying degrees. By observing the particular manner in which these other lipids, especially phospholipids (PL) substituted for DCP, affected the precipitation reactions, and correlating these effects with the stereoelectrical properties of the lipids, it has been possible to obtain from substituted 7PC:2DCP:lChol liposomes experiments some insights into the manner in which membrane lipid/mineral interactions may affect MV calcification in vivo. WC:2PL;lChol liposome mineralization Table 1 summarizes the effect various acidic (APL) and neutral (NPL) phospholipid substitutions were found to have on precipitation in our model liposome suspensions. In general, PLs appear to affect precipitation in three different ways: (1) by reducing the amount of intraliposomal precipitate formed, (2) by increasing the time gap between intraliposomal precipitate formation and the onset of extraliposomal precipitation, and (3) by slowing down or inhibiting subsequent growth of the extraliposomal precipitate. Of the APLs examined, only phosphatidic acid (PA) significantly affected intraliposomal precipitate formation. The reduced yields (Table 1) most probably re-
Table 1 Summary of the general effects various phospholipids have on apatite precipitation in the interior and exterior fluid spaces of 7PC:2PL: 1Chol” liposome suspensionsb (based on data presented in ). Phospholipid”
DCP PA PE FG PI PS
Intraliposomal precipitationc Yield (5 h)
Yield (48 h)
No effectd No effectC No effect’ Reduces(-aO%) Suppresses Suppresses Reduces(-40%) Delays (-5 h) No effect No effect No effect No effect No effect Delays (-2 h) No effect No effect Delays (-24 h) Reduces(-80%) Reduces(-60%) Delays (-5 h) Sph No effect ~-s-~x_p-_l__l__ -1__111 ’ Lipid symbolsdefined in text. ’ All liposomeeinitially contained 50 mM encapsulatedPO.,, primed with X-537A, and suspendedin a buffered NaCl solution (pH 7.4.240 mOsmol) containing2.2 mM Ca2+ and 1.5 mM PO,. ’ Onset of intraliposomalprecipitation is immediate in all cases. d Maximum yield occursby 2-5 h. ’ Onset of precipitation is immediate. ’ MaxLmumyield ogurs by 24-48 h after onset.
271 suited from @aPA-coupled 8, leakage across the membrane decreasing substantially (UP to 75%) the amount of initially encapsulate PO, available for reaction
[6,7]. On the other hand, the lower intraliposomal precipitation in NPL liposomes  was directly related to a reduction in the actual volume of PO, reaction solution initially encapsulated. In either case, there was no evidence that the substituted PLs directly interfered with the precipitation reaction itself. However, this was not the case with extraliposomal precipitation. PL-mineral interactions appear to strongly influence both the initial appearance and the subsequent growth and proliferation of mineral crystals outside the liposomes. Transmission electron microscopy (TEM) observations [6,8] show that the most rapid extraliposomal precipitate development occurred in those iiposomes, e.g. 7PC:2DCP:lChol, in which the endogenous seed crystals made only incidental contact with the lipid membrane. In contrast, extraliposomal precipitation was totally suppressed in PA liposomes where intraliposomal crystals were observed to be in extremely close physical association with the inner membrane surface. Although the mechanism by which crystals trate the membrane to initiate extralip precipitation is unknown, these observations suggest that the rapidit transit is governed in part by the extent of the crystal/membrane attachmen too extensive, as in the case with PA liposomes, the crystals adhere to the inner membrane surface rather than penetrate the membrane. The degree of crystal/membrane contact and the subsequent delay effected in extraliposomal precipitation is, in turn, probably a reflection of the comparative strength of the electrostatic attractions bringing the two surfaces together. Because of steric constraints imposed by the lipid bilayer structure, these interactions would be limited largely to the PL terminal groups lining the membrane surface. The strength of the terminal charges on the APLs, therefore, could explain the relative differences in their effect on the onset of extraliposomal precipitation. The negative PO, terminus of PA forges the strongest bond with the crystal surface, probably via anhydrous Ca *+ bridges to surface-exposed lattice PO, , followed in strength b the zwitterionic serine group of phosphatidylserine (PS), with the neutral 0 groups terminating phosphatidylinositol (PI) and phosphatidylglycerol (PG) forming the weakest Ca*+-bridges. In the case of the NPL liposomes, the relatively brief delays in the onset of extraliposomal precipitation could possibly br attributed to direct, i.e., nor&a*+ mediating, interactions between the N+-containing terminal groups of PC and phosphatidylethanolamine (PE) or sphingomyeiin (Sph) and the apposing surface-situated lattice PO, slowing seed crystal penetration. On the other hand, the lack of appreciable membrane restraints on extraliposomal precipitation in 7PC:2DCP: 1Chol liposomes probably resulted from a combination of offsetting factors. In particular, the terminal PO, groups of DC are too recessed in the body of the membrane  to be able to establish Ca2+-links with crystal Pop, yet the net negative charge they impart to the membrane sufficiently repels the anionic crystal surface that effective terminal (N+)PC-crystal PO, interactions do not readily occur. The same electrostatic attractions that appreciably delayed crystal transit through the membranes of PS-substituted liposomes could also account for the ob-
272 served slowdown in subsequent extraliposomal mineral growth in metastable suspensions of these liposomes (Table 1). Once the seed crystals made contact with the external solution phase, the attachment of PS-containing exterior membrane surfaces of neighboring liposomes to these crystals could delay their further growth and/or proliferation by blocking access of external solution Ca*’ and PO, ions to active accretion sites on their surfaces [S]. However, since in native MV PS is largely restricted to the inner leaflet of the bilayer, this effect may be of less significance in vivo. M V-like lipasomc minerdiza~iun Of the three ways described above in which PLs can affect precipitation in liposoma1 suspensions, only the second is probably of biological interest. There is no evidence that calcifying MVs aggregate in vivo in the manner postulated for PS-containing liposomes around exposed seed crystals. Recent studies (Skrtic and Eanes, unpublished) likewise showed that intraliposomal precipitation rapidly reach maximum levels in liposomes with MV-like membrane lipid compositions [9,10]. In contrast, extraliposomal precipitation was largely suppressed in these suspensions. Of the most common lipids in MVs (PA is not among them f9,10]), only PS and Sph ap peared to be involved in this slowdown. Inhibition was reversed when both of these PLs were removed from the MV-simulated lipid mixture. Based on the model liposome experiments discussed in this review, one could postulate that membrane PS and Sph may influence MV calcification by affecting the release of intravesicularlyformed crystals into the extracellular fluid space.
References 1 Bonucci E. Fine structure ofearly cartilagecalcification. J Ultrastruct Res 1967;20:33-50. 2 Anderson NC. Vesicles associatedwith calcificationin the matrix of epiphysealcartilage. J Cell Biol 1969;41:59-72. 3 Eanes ED, Hailer AW, Costa JL. Calcium phosphateformation in aqueoussuspensionsof multilamellar liposomes.Calcif Tissue Int 1984;36:421-430. 4 Eanes ED, Hailer AW. Ltposome-mediatedcaicium phosphateformation in mctastab!: so!u?ions. Calcif Tissue Int 1985;37:390-394. 5 Eanes ED, Hailer AW. Calcium phosphateprecipitation in aqueoussuspensionsof phosphatidylserine-containing anionic liposomes.Calcif Tissue Int 1987;40:43-48. 6 Eanes ED, Hailer AW, Heywood BR. Modulation of calcium phosphate formation by phosphatidate-containinganionic liposomes.Calcif Tissue Int 1988;43:226-234. 7 Skrtic D. Eanes ED. Effect of different phospholipid-cholesterolmembrane compositionson liposome-mediatedformation of calciumphosphates.Calcif Tissue fnt (in press). 8 Heywood BR, Eanes ED. An ultrastructural study of the effectsof acidic phospholipid substitutions on calcium phosphateprecipitation in aqionic liposomes.Calcif Tissue Int (in press). 9 Peress NS, Anderson HC, Sajdera SW. The lipids of matrix vesicles from bovine fetal epiphyseal cartilage. CalcifTissue Res 1974;14:275-281. JO Wuthier RE. Lipid composition of isolated epiphyseal cartilage cells, membranes and matrix vesicles. Biochim Biophys Acta 1975;409:128-143.