A search for apatite crystals in the gap zone of collagen fibrils in bone using dark-field illumination

Highlights

• Dark field (DF) TEM was used to search for alignment of apatite mineral crystals along gap zones in collagen fibrils.

• DF-TEM confirms that most mineral is in polycrystalline plates aligned with fibrils.

• Only traces of mineral aligned with gap zone.

• Most of calcium phosphate in gap zone cannot be crystalline, could be amorphous (ACP).

Abstract

Bright-field transmission electron microscope (TEM) images of ion milled or focused ion beam (FIB) sections of cortical bone sectioned parallel to the long axis of collagen fibrils display an electron-dense phase in the gap zones of the fibrils, as well as elongated plates (termed mineral lamellae) comprised of apatite crystals, which surround and lie between the fibrils. Energy dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS) studies by others have shown that the material in the gap zones is calcium phosphate. Dark-field (DF) images are capable of revealing the projected position of crystals of apatite in a section of bone. We obtained bright field (BF) images of ion milled sections of bovine femoral cortical bone cut parallel to fibril axes (longitudinal view), and compared them with DF images obtained using the (002) apatite reflection to test a widely held theory that most of the mineral in bone resides in the gap zones. Most apatite crystals which were illuminated in DF images and which projected onto gap zones were extensions of crystals that also project onto adjacent overlap zones. However, in BF images, overlap zones do not appear to contain significant amounts of mineral, implying that the crystals imaged in DF are actually in the interfibrillar matrix but projected onto images of fibrils. However a small number of “free” illuminated crystals did not extend into the overlap zones; these could be physically located inside the gap zones. We note that projections of gap zones cover 60% of the area of any longitudinal field of view; thus these “free” crystals have a high random probability of appearing to lie on a gap zone, wherever they physically lie in the section. The evidence of this study does not support the notion that most of the mineral of bone consists of crystals in the gap zone. This study leaves uncertain what is the Ca-P containing material present in gap zones; a possible candidate material is amorphous calcium phosphate.

Introduction

Bone is widely recognized to be a composite material made up principally of two components: collagen and mineral. Triple helices of Type I collagen are assembled into fibrils which are closely associated with plate-like crystals of apatite. The c-axes of the apatite are aligned approximately parallel to the long axes of the collagen fibrils [1]. Using sections prepared by either ion milling or focused ion beam (FIB) methods, we have shown in previous work that the majority of the mineral in bone occurs outside the collagen fibrils in the form of elongated plates, 100 s of nm long, approximately 5 nm thick, and 60 to 80 nm wide [1,2]. Using dark field (DF) imaging, it was possible to show that the plates are mosaics of single crystals, 5 nm thick and a few tens of nm across [3], approximately the same dimensions as single crystals that had been previously obtained by removing collagen from samples of powdered bone by chemical oxidation [4]. These plates range from being curved and wrapped around the fibrils, to flat plates which lie between adjacent fibrils. They have been referred to as mineral lamellae (MLs) [5]. This term refers specifically to the polycrystalline, 5–6 nm thick plates which we have observed in all bone samples studied by us. We use this term in preference to less specific terms such as “platelet” and “mineral particle” which have been used elsewhere in the literature. McNally et al. [1] showed that most of the mineral in bone (~80%) is in the form of these extrafibrillar plates.

However, it has long been apparent that some mineral also occurs within the collagen fibrils. In bright-field (BF) images of sections of bone prepared either by ion milling or by microtome sectioning, periodic electron-dense bands 40 nm wide are seen oriented perpendicular to the long axes of the fibrils, and spaced about 67 nm apart (Fig. 1). These have been interpreted as showing the presence of mineral in the so-called gap zones along which occur 40 nm-long gaps between the ends of collinear collagen molecules [[6], [7], [8]]. Adjacent to each gap zone is an “overlap zone” in which no excess electron density is observed. Using either electron energy loss spectroscopy (EELS) or energy-dispersive X-ray spectroscopy (EDS) it has been found that the gap zones contain substantial concentrations of Ca and P [1,9,10], supporting the notion that they contain crystals of a Ca phosphate mineral.

For each BF TEM image of bone, we can obtain a selected-area electron diffraction pattern (SAED) showing rings which correspond to the positions of Bragg-reflections from crystal lattice planes (Fig. 2). The rings can be indexed to the lattice planes of apatite (all apatites including hydroxyapatite give essentially the same diffraction pattern). If we place an aperture on part of one of these rings and create an image using only the electrons scattered through that aperture, then we obtain an image of only those crystals which were appropriately oriented so that they could diffract electrons [11]. These are called dark-field (DF) images. They allow us to test two kinds of questions: a) what are the shapes and sizes of single crystals of apatite in bone?; and b) what is the spatial relationship of the apatite crystals to each other and to features previously visualized in BF images of the same field of view?

Since gap zones contain significant amounts of Ca and P, we could use the DF method to test whether crystals of apatite are located in them. In essence, this was the goal of the present paper.

BF images of sections of bone such as Fig. 1 represent projections of the three-dimensional (3D) form of an approximately 100 nm-thick section of bone onto the projection plane. As a result there is some uncertainty in interpreting the relation between mineral and collagen from the 2D images. Fortunately, as a result of our previous studies of images of orthogonal sections bone, it is possible to reconstruct some general features of the 3D form of the section and recognize the effects of superposition of objects within this section. Fig. 3 shows a schematic view of a 100 nm-thick block of bone cut so that it can be viewed in the TEM through a plane oriented parallel to the collagen fibrils (the X-Y plane in Fig. 1). Sections cutting transversely through the fibrils are seen in the X-Z planes; mainly these appear to be empty holes or partially filled with low-contrast material. We have shown that in bone these holes are filled with collagen but that in TEM sections the collagen is partially or wholly eroded away by the ion beam [12]. From Fig. 3 we see that any electron path directed normal to and passing through the X-Y plane will pass through one or two 50 nm-diameter fibrils within the sample. Gap zones extend continuously across the X-Y plane, defined by zones of higher electron density. However, from Fig. 3 we see that this continuity of gap zones is a result of two facts: a) the fibrils are vertically superimposed in a staggered array; and b) the gap zones in adjacent fibrils are always in register.

The gap zones provide a convenient means for identifying the presence of a collagen fibril at some depth (Z coordinate) in the section beneath a given point on the X-Y plane. However, note that the gap zones extend continuously across the width of the fields of view of all longitudinal sections of bone. This tells us that collagen fibrils are universally present at some depth under any point in the X-Y plane.

As a result of the superposition of fibrils, it is in general difficult to know whether a particular feature (e.g. an apatite crystal) seen in a BF or DF image is inside a fibril or in the surrounding mineralized matrix. However, we have just shown that every object seen on the X-Y plane must be projected onto a collagen fibril whether or not the object physically resides inside the fibril. Recollecting that gap zones occupy 60% of the volume of fibrils [13], there is a 0.6 probability that any object seen projected onto a fibril will overlap the image of a gap zone, whether or not the object is physically located inside the gap zone.

The goal of the present study is to test the widely held hypothesis that most or at least many of the apatite crystals found in bone are physically located inside the fibrils, and more specifically within the gap zones. The approach to this is as follows. First we demonstrate typical BF images of longitudinal sections of bone, to show the apparent location and form of some of the crystalline plates (mineral lamellae) that surround collagen fibrils, as well as the partial electron opacity associated with the 40 nm-wide gap zones. We then present DF images of the same field of view, using the (002) reflections. We have also superimposed the positions of the gap zones (taken from the BF image) on the DF image. Following this, we report on the occurrence of DF images of apatite crystals that appear to be only confined to gap zones, rather than DF images of crystals which are clearly parts of MLs that extend beyond the boundaries of the gap zone in question. Presence of images of the former type would partially confirm the existence and allow us to estimate the abundance of crystals physically within the gap zones. We also show one example of the corresponding BF/DF images obtained using a number of other Bragg reflections to test whether they add information to this inquiry.