Calcified Tissue Research
Calcif. Tiss. Res. 25, 217-222 (1978)
9 by Springer-Verlag 1978
The Morphology of Bone Mineral Crystals S.A. Jackson, A.G. Cartwright, and D. Lewis Department of Mechanical Engineering,Universityof Surrey,Guildford,Surrey,England
Summary. Electron microscopical observations of the size and shape of bone mineral crystallites have not been in complete agreement with X-ray diffraction findings. The two prevalent viewpoints consider bone mineral crystals to be either rod, or plate like in habit. There appears to be agreement that the smallest dimension of the crystals is about 5 nm, but there is discrepancy in the reported c-axial lengths. The method of dark field imaging is used to obtain a quantitative measurement of the c-axial length distribution in rabbit, ox and human bone: mean c-axial lengths 32.6 nm, 36.2 nm and 32.4 nm, respectively, show no significant difference at the 5% level to the mean c-axial length measured by X-ray line broadening. Both bright and dark field images strongly suggest that bone mineral has a plate like form. Reasons for past discrepancies are discussed.
Key words: Bone mineral - - Electron microscopy - X-ray diffraction - - Dark field.
Introduction The determinations of bone mineral morphology in the past have employed two major techniques, namely Xray diffraction, and electron microscopy. With respect to the former, using the method of line broadening, Carlstr6m and Glas (1959) succeeded in measuring the (002), (004), (008) and (130) diffraction profiles of fish bone. They calculated a c-axial length of about 60 nm, a width of 5 rim, and concluded that the line broadening was due to both lattice size and strain effects. Myers and Engstr6m (1965), using calcified fowl tendon, Send requests for offprints to S.A. Jackson, Departmentof Medical
Physics, UniversityHospital, NottinghamNG7 2UH, England
calculated a c-axial dimension of 35 nm, and a width of 5 nm. Sttihler (1938), using various types of bone, estimated the c-axial length to lie between 3 nm and 29 nm, and Carlstr6m (1955), also using various types of bone, calculated a mean c-axial length of 23 _+ 2 nm. Using the low angle scattering of X-rays, Engstr6m and Finean (1953), found that the crystals in various bones were of uniform size, elongated along the c-axis and measuring 21 nm x 7.5 rim. Carlstr6m and Finean (1954) confirmed these findings using fish bone, which is a highly oriented structure. They concluded that the crystals were probably rod shaped, having dimensions of 22 n m • 6.5 rim. Engstr6m (1972) using the same technique and type of bone calculated a c-axial dimension of 33-35 nm x 35 nm. Using electron microscopy, Robinson (1952) measured an average crystallite size in human bone of 50 nm x 25 nm x 10 nm with the log direction parallel to the c-axis. In a later investigation, Robinson and Watson (1952), concluded that the mature crystal length was 35 nm-40 nm, with a width of about the same value. The thickness of these plates was estimated to lie between 2.5 nm and 5 nm. It has been suggested by Fernandez-Moran and Engstr6m (1957), that the observed plates are actually lateral aggregations of rods. However, this does not account for the difference in contrast observed with the microscope between the two types of structures; the rod like structure appearing more dense than the plate like objects. Steve-Bocciarelli (1969) made a study of ox bone with the electron microscope, and by using a tilting specimen stage, he was able to compare images of the same crystallite taken at +_20~ to the normal position. He arrived at the following conclusions: (a) Rod like images of the crystallites were actually edge views of plate like structures.
0008-0594/78/0025-0217501.20
218 (b) T h e plates m e a s u r e d 2 - 4 n m in thickness and up to 70 n m in length and width. (c) T h e plates s h o w e d no well defined profile, but s o m e t i m e s an angle o f 90 ~ c o u l d be o b s e r v e d . T h e s e o b s e r v a t i o n s are c o r r o b o r a t i o n o f w o r k by J o h a n s e n a n d P a r k s (1960), who, basically, arrived at the s a m e conclusions. In s u m m a r y , m i c r o s c o p i c a l o b s e r v a t i o n s on the size and shape o f b o n e mineral crystallites h a v e not been in a g r e e m e n t with X - r a y diffraction findings. T h e t w o p r e v a l e n t v i e w p o i n t s c o n s i d e r b o n e mineral to be either rod like or plate like in habit. T h e r e a p p e a r s to be a g r e e m e n t t h a t the smallest d i m e n s i o n o f the crystals is a b o u t 5 nm, but there is d i s c r e p a n c y in the r e p o r t e d caxial lengths; X - r a y diffraction p r o d u c i n g values o f a b o u t 20 n m - 3 5 nm, whilst m i c r o s c o p i c a l estimations tend to be 2 or 3 times larger. Whilst v a r i o u s h y p o theses h a v e been suggested to explain this d i s c r e p a n c y , e x p e r i m e n t a l e v i d e n c e is still i n c o n c l u s i v e in delineating both the size and shape o f b o n e m i n e r a l crystallites. In an a t t e m p t to explain p a s t discrepancies, and h e n c e obtain a m o r e a c c u r a t e k n o w l e d g e o f b o n e mineral m o r p h o l o g y , b o t h X - r a y diffraction and bright field electron m i c r o s c o p y h a v e been used t o g e t h e r with the d a r k field i m a g i n g technique, w h i c h to date, has n o t been applied to this particular problem.
Materials and Methods Initially, all techniques were developed using cortical bone, cut from the midfemoral region of 6-month-old female Dutch rabbits. Later, cortical bone from the midfemoral regions of a 2-year-old ox and a 65-year-old human male were used. The human sample was obtained at postmortem, and death was not caused by a primary bone disorder. In experiments to determine the effect of specimen preparation on bone mineral, Johansen and Parks (1960), Ascenzi et al. (1968), and Jackson (1976), have concluded that there are no significant differences (at the 5% level), in the mean size and composition of bone mineral crystals between fixed and unfixedunembedded samples. Consequently, unfixed unembedded bone has been used for this work. The bone samples, approximately 1 mm square, were cemented into an epoxy resin block for support in the LKB4800A ultramicrotome jig. The samples were oriented such that sections were cut in a plane, parallel to the periosteal surface. Using glass knives, the samples were trimmed to approximately 0.1 mm square, then sectioned using a diamond knife with a cutting speed of 5 cm/s and a feed of 60 nm-80 nm between cutting strokes. Distilled water was used as a flotation medium, the sections being picked up immediately on 400 mesh copper grids, and allowed to dry on filter paper before being examined. Microscopical examination was performed using a JEOL.100b electron microscope operated at 80 kV. The conventional observation of a biological specimen is by bright field imaging, where the final image is formed by a beam of electrons transmitted through tbe specimen. The dark field method uses diffracted beams to form the image, and this is achieved by tilting the main beam so that the selected diffracted beam passes through the objective aperture, as
S.A. Jackson eta[.: Bone Mineral Crystal Morphology
Brightfield b,earn
Darkfield
A&0
--
"
beam
Fig. 1. Dark field image formation. The main beam is tilted to allow the desired diffracted beam to pass through the objective aperture
shown in Figure 1. By using the (002) diffracted beam, (the (002) planes being perpendicular to the c-axis in hydroxyapatite), the dark field imaged crystals have their c-axes in the plane of the section, and in a direction within that plane determined by the position of the objective aperture on the diffraction pattern, see Figure 2. A 10 /lm diameter aperture was reqaired to isolate the (002) beam from its neighbours in order to reduce interference effects. After electronoptical rotation had been allowed for, approximately 300 measurements of c-axial length per sample were taken from many micrographs, of crystals which appeared well defined and isolated. The measurements were analysed statistically. Beef catalase was used as a magnification standard. In the X-ray diffraction studies the peak profiles of the (002) and (004) diffracted beams were recorded in triplicate using a standard X-ray diffraetometer and nickel filtered copper radiation. Instrumental broadening of the peak profiles were measured using a sample of rabbit bone powder which had been heated to 800 ~ C for two h, and containing crystals with c-axial lengths well in excess of 100 nm, which produced negligible broadening. To calculate the
Fig. 2. Diffraction pattern and objective aperture images superimposed. The (002) diffracted beams pass through the aperture to form the dark field image
S.A. Jackson et al.: Bone Mineral Crystal Morphology
219
mean c-axial length (t) and the lattice strain factor (e) from the measured peak profiles the method described below was adopted. The correction of Warren (1938), was used to allow for instrumental broadening, i.e., (B~)Z = (/~)2 + (~)z
(1)
where B-~= measured broadening = broadening by crystal = instrumental broadening (~), the peak broadening due to the crystal, was assumed to have components due to both crystal size and crystal strain, where fl89size =
0.92
Results
In the bright field m i c r o g r a p h , F i g u r e 3, b o n e mineral crystals c a n be seen as b o t h l o n g thin d a r k objects, a n d less dense irregular plate like objects. Since an electron m i c r o g r a p h is a t w o d i m e n s i o n a l p r o j e c t i o n o f a three d i m e n s i o n a l object, this c o u l d m e a n t h a t the rod like images are plates viewed on edge, or t h a t the two types of structures are actually present. C o l l a g e n c a n n o t be
(2)
t Cos 0B
strain = 2e tan 0 B
(3)
= X-ray wavelength = c-axiallength 0 B= Bragg angle e = Lattice strain, including both tensile and compressive components. Using the Warren formula for the addition of Gaussian peak s: t
(f189 = ~ size)2 + ~ strain)Z,
(4)
and combining (2) (3) and (4) yields (fl89 Cos
zO~=
+ 4e2 Sin z0~
(5)
The (004) reflection is a second order reflection from the (002) planes (i.e., n = 2 in the Bragg Law 2d Sin 0 B= n2), hence by measuring both the ~ values of the (002) and (004) peak profiles two simultaneous equations can be produced by substitution in Equation (5), there by enabling t and ~ to be calculated.
Fig. 4. High magnification bright field micrograph. Original magnification 350,000x. (100) lattice plane resolution is indicated
Fig. 3. Bright field electron micrograph of rabbit femur. Showing both rod and plate like objects Original magnification 35,000x
Fig. 5. Dark field image formed by using (002) diffracted beams. Original magnification 75,000x. The c-axial direction is indicated, and interference fringes are arrowed
220
seen in this micrograph since unstained collagen is electron transparent. Because of the dense nature of fully calcified bone matrix, single crystals are difficult to isolate, as their outlines tend to merge and overlap. The high magnification micrograph, Figure 4, shows many fringe patterns due mainly to (100) lattice plane resolution. (The (100) planes having a spacing of 0.8169 nm.) The resolved lines can be seen to extend over distances of a few tens of nanometres, indicating that crystals of this size are in fact single crystals, and not composed of aggregations of smaller units. The crystals themselves are almost impossible to isolate, making any form of size measurement difficult. Probably the most important observation to be drawn here, is the fact that even if crystals could be isolated for measurement, the orientation of crystal axes cannot be determined in any bright field image, and so measurements have little relation to the crystal lattice. However, both the problem of isolation and orientation have been overcome by using the method of dark field imaging as described earlier. Figure 5 shows a dark field image of a section of rabbit femur, formed using the (002) diffracted beam, the c-axial direction is indicated. It is known from previous X-ray work that the (002) planes of bone mineral are perpendicular to the longest dimension of the crystal, which is the c-axial direction, and therefore dark field images produced using the (002) beams are images of crystals whose c-axes are in the plane of the section. It is fortunate that the (002) diffracted beam is strong and well isolated. This has meant, by using a 10 #m diameter objective aperture, pure (002) dark field images could be produced, and thereby, reducing Moir6 fringe effects from interfering diffracted beams in overlapping crystals. However some fringes can be seen due to interference between separate (002) diffracted beams. The general appearance of the dark field images show bone mineral crystals to have irregular shapes and sizes. The dimension measured perpendicular to the c-axial direction varies from approximately 5.0 nm to dimensions approaching that of the c-axial length in some instances. Since an (002) dark field image would still be produced if the crystal were rotated about its caxis, this would be consistent with the suggestion that bone mineral has a plate like form, with a thickness of approximately 5.0 nm. From many dark field micrographs only crystals which appeared well isolated were selected for measurement. The c-axial length distributions for rabbit, ox, and human bone have been analysed statistically and are summarized in Figures 6, 7 and 8. Using the same samples of bone as used for the microscopical work, the peak profiles of the (002) and
S.A. Jackson et al.: Bone Mineral Crystal Morphology
(004) diffracted beams were recorded. The measured values of peak width at half maximum intensity (B89 and peak position (20B), are summarized in Table 1, together with the calculated values of mean c-axial
RABBIT BONE
iiiiiiiiiiiii!
c
20
40
60
80
100
120
C-axial length (nm)
Min. 8. 3nm.
Max. Mean 1 S.D. no. Ohm. 32.6nm. 12.8nm.
Fig. 6. Histogram of c-axial length distribution in rabbit bone
OX BONE 70
60
50
40
7O 3O 20
~
0
20
40
60
80
I00
120
C-axial length' (nm) Min.
Max.
Mean
S.D.
&7nm.
113.0 nm.
36.2 nm.
19.0 nm.
Fig. 7. Histogram of c-axial length distribution in ox bone
S.A. Jackson et al.:Bone Mineral Crystal Morphology
221 significant difference, at the 5% level, from the values determined by line broadening.
HUMANBONE
?0 60
Discussion
50
The bright field electron microscopical results strongly suggest that bone mineral crystals are in the form of thin plates, approximately 5.0 nm thick, with an irregular shape and a variable maximum dimension. Whilst m a n y previous workers have assumed the largest dimensions of crystals imaged in the bright field to be the c-axial length, the fact that crystal orientation cannot be determined precludes any accurate measurement. Hence, to obtain quantitative information from electron micrographs, the bone mineral crystal outline has to be clearly defined and the crystal orientation known. The technique of dark field imaging solves these two problems simultaneously. F r o m m a n y dark field micrographs, c-axial measurements were taken for 3 bone types. For rabbit, ox and human bones the mean c-axial length was found to be 32.6 nm, 36.2 nm, and 32.4 nm respectively, the size distributions having standard deviations of 12.8 nm, 19.0 nm and 16.3 nm. High resolution bright field micrographs occasionally show images where the electron density and lattice resolution lines are not consistent over the entire crystal, and this could indicate structural imperfection which would cause lattice strain. Also, since the bone mineral crystals are so small, a large proportion o f unit cells per crystal reside on the surface, and will therefore be deformed. From the line broadening measurements, using a formulation which allowed for lattice strain, the mean c-axial lengths were calculated to be 33.2 rim, 36.1 nm and 32.2 nm for rabbit, ox, and h u m a n bones respec-
40 c
30 20 10 0
JJPJfJ 20
zlO
80
60
I00
120
C-axial length (nm) 6,0 rim.
I00.0 rim.
32,4 nm
16.3nm.
Fig. 8. Histogram of c-axial length distribution in human bone
length (t) and lattice strain factor (e), which contains both tensile and compressive components. To compare the dark field measurements with the X-ray results, from each set of dark field measurements, 10 subsets of data were randomly selected, each subset containing 25 measurements. The mean values for the c-axial length calculated from these subsets are normally distributed about the true mean. Hence, the average mean value calculated from the subsets can be compared with the X-ray measurements using Student's t test. Provided that lattice strain is allowed for in the calculation of mean c-axial length, each bone type measured by the dark field technique shows no
Table 1. Showing a summary of the X-ray diffraction measurements, and the calculated values of (t) the mean c-axial length and (~) the lattice strain factor Sample
B89(002) ~
B~ (004)~
20B(002)~
20B(004)~
Rabbit femur, heated to 800 ~ C for 2 h Rabbit femur, untreated
0.138 0.133 0.142 0.329 0.317 0.323 0.325 0.323 0.313 0.335 0.342 0.329
0.238 0.217 0.229 0.500 0.481 0.513 0.525 0.500 0.494 0.513 0.488 0.519
25.923 25.931 25.925 25.920 25.920 25.910 25.975 25.950 25.963 25.981 26.000 25.994
53.208 53.221 53.242 53.213 53.225 53.225 53.250 53.260 53.240 53.260 53.250 53.260
Ox femur, ~ untreated Human femur, untreated
t (nm)
e (%)
m
m
m
Mean 33.2 SD 1.1
Mean 0.61 SD 0.05
Mean 36.1 SD 1.1
Mean 0.64 SD 0.05
Mean 32.2 SD 1.1
Mean 0.61 SD 0.07
222 tively, which, when c o m p a r e d to the dark field measurements show no significant difference at the 5% level. The lattice strain factor (e) was calculated, and it was found that the m a x i m u m tensile or compressive component had a value o f approximately 0.3% in all samples. This finding is important since it means that the crystals are basically straight in the c-axial direction, and the d a r k field images are therefore of the whole crystal, and not just a small section o f a large curved crystal. In the past, bright field estimations o f c-axial length have tended to be 2 or 3 times longer than those obtained by the X - r a y methods; p r o b a b l y because only the larger crystals could be seen in isolation, together with the fact that lattice strain has not always been allowed for. Furthermore, the inaccuracies involved with bright field images invalidates any comparison between the two methods. Also the two methods have tended to be used in isolation, with different types of bone, making comparison of separate investigations difficult. The dark field technique has enabled consistent measurements o f m e a n c-axial length to be produced using both microscopical and X - r a y methods. It has also enabled the crystal size distribution to be measured within a particular sample.
References Ascenzi, A., Bonucci, E., Steve-Bocciarelli, D.: Fine structure of bone mineral in different experimental conditions. IV European Regional Conf. on Electron Microscopy. Rome, pp. 431-433 (1968)
S.A. Jackson et al.: Bone Mineral Crystal Morphology CarlstrSm, D.: X-ray crystallographic studies on apatites and calcified structures. Acta Radiologica. Supp. 121, 33-37 (1955) Carlstr6m, D., Finean, J.B.: X-ray diffraction studies on the ultrastructure of bone. Biochim. Biophys. Acta 13, 183-191 (1954) EngstrSm, A.: Aspects of the molecular structure of bone. In: The biochemistry and physiology of bone (G.H. Bourne, ed.). New York: Vol. 1, (2nd edition), pp. 237-257 (1972) Engstr6m, A., Finean, J.B.: The low angle x-ray diffraction of bone. Nature 171,564 (I953) Fernandez-Moran, J., EngstrSm, A.: Electron microscopy and xray diffraction of bone. Biochim. Biophys. Acta 23, 260-264 (1957) Jackson, S.A.: The morphology of bone mineral. PhD. Thesis, University of Surrey (1976) Johansen, D.M.D., Parks, H.F.: Electron microscopic observations on the 3-dimensional morphology of apatite crystallites of human dentine and bone. J. Biophys. Biochem. Cytol. 7 (No. 4), 743-745 (1960) Myers, H.M., EngstrSm, A.: A note on the organisation of hydroxyapatite in calcified tissues. Exp. Cell. Res. 40, 182-185 (1965) Robinson, R.A.: An electron microscopic study of the crystalline inorganic component of bone and its relationship to the organic matrix. J. Bone and Joint Surg. 34, 389434 (1952) Robinson, R.A., Watson, M.L.: Collagen-crystal relationships in bone as seen in the electron microscope. Anat. Res. 114, 383410 (1952) Steve-Bocciarelli, D.: Morphology of crystallites in bone. Calc. Tiss. Res. 5,261-269 (1969) Stuhler, R.: Uber den Feinbau des Knochens, Eine RSntgen - Feinstruktur Untersuchung. Fortschr. Gebiete RSntgenstrahln. 57, 231-234 (1938) Warren, B.E., Biscoe, J.: The structure of silica glass by x-ray diffraction studies. J. Am. Ceram. Soc. 21, 49-54 (1938) Received July lO/Revised December 15~Accepted December 21, 1977
Calcif. Tiss. Res. 25, 217-222 (1978)
9 by Springer-Verlag 1978
The Morphology of Bone Mineral Crystals S.A. Jackson, A.G. Cartwright, and D. Lewis Department of Mechanical Engineering,Universityof Surrey,Guildford,Surrey,England
Summary. Electron microscopical observations of the size and shape of bone mineral crystallites have not been in complete agreement with X-ray diffraction findings. The two prevalent viewpoints consider bone mineral crystals to be either rod, or plate like in habit. There appears to be agreement that the smallest dimension of the crystals is about 5 nm, but there is discrepancy in the reported c-axial lengths. The method of dark field imaging is used to obtain a quantitative measurement of the c-axial length distribution in rabbit, ox and human bone: mean c-axial lengths 32.6 nm, 36.2 nm and 32.4 nm, respectively, show no significant difference at the 5% level to the mean c-axial length measured by X-ray line broadening. Both bright and dark field images strongly suggest that bone mineral has a plate like form. Reasons for past discrepancies are discussed.
Key words: Bone mineral - - Electron microscopy - X-ray diffraction - - Dark field.
Introduction The determinations of bone mineral morphology in the past have employed two major techniques, namely Xray diffraction, and electron microscopy. With respect to the former, using the method of line broadening, Carlstr6m and Glas (1959) succeeded in measuring the (002), (004), (008) and (130) diffraction profiles of fish bone. They calculated a c-axial length of about 60 nm, a width of 5 rim, and concluded that the line broadening was due to both lattice size and strain effects. Myers and Engstr6m (1965), using calcified fowl tendon, Send requests for offprints to S.A. Jackson, Departmentof Medical
Physics, UniversityHospital, NottinghamNG7 2UH, England
calculated a c-axial dimension of 35 nm, and a width of 5 nm. Sttihler (1938), using various types of bone, estimated the c-axial length to lie between 3 nm and 29 nm, and Carlstr6m (1955), also using various types of bone, calculated a mean c-axial length of 23 _+ 2 nm. Using the low angle scattering of X-rays, Engstr6m and Finean (1953), found that the crystals in various bones were of uniform size, elongated along the c-axis and measuring 21 nm x 7.5 rim. Carlstr6m and Finean (1954) confirmed these findings using fish bone, which is a highly oriented structure. They concluded that the crystals were probably rod shaped, having dimensions of 22 n m • 6.5 rim. Engstr6m (1972) using the same technique and type of bone calculated a c-axial dimension of 33-35 nm x 35 nm. Using electron microscopy, Robinson (1952) measured an average crystallite size in human bone of 50 nm x 25 nm x 10 nm with the log direction parallel to the c-axis. In a later investigation, Robinson and Watson (1952), concluded that the mature crystal length was 35 nm-40 nm, with a width of about the same value. The thickness of these plates was estimated to lie between 2.5 nm and 5 nm. It has been suggested by Fernandez-Moran and Engstr6m (1957), that the observed plates are actually lateral aggregations of rods. However, this does not account for the difference in contrast observed with the microscope between the two types of structures; the rod like structure appearing more dense than the plate like objects. Steve-Bocciarelli (1969) made a study of ox bone with the electron microscope, and by using a tilting specimen stage, he was able to compare images of the same crystallite taken at +_20~ to the normal position. He arrived at the following conclusions: (a) Rod like images of the crystallites were actually edge views of plate like structures.
0008-0594/78/0025-0217501.20
218 (b) T h e plates m e a s u r e d 2 - 4 n m in thickness and up to 70 n m in length and width. (c) T h e plates s h o w e d no well defined profile, but s o m e t i m e s an angle o f 90 ~ c o u l d be o b s e r v e d . T h e s e o b s e r v a t i o n s are c o r r o b o r a t i o n o f w o r k by J o h a n s e n a n d P a r k s (1960), who, basically, arrived at the s a m e conclusions. In s u m m a r y , m i c r o s c o p i c a l o b s e r v a t i o n s on the size and shape o f b o n e mineral crystallites h a v e not been in a g r e e m e n t with X - r a y diffraction findings. T h e t w o p r e v a l e n t v i e w p o i n t s c o n s i d e r b o n e mineral to be either rod like or plate like in habit. T h e r e a p p e a r s to be a g r e e m e n t t h a t the smallest d i m e n s i o n o f the crystals is a b o u t 5 nm, but there is d i s c r e p a n c y in the r e p o r t e d caxial lengths; X - r a y diffraction p r o d u c i n g values o f a b o u t 20 n m - 3 5 nm, whilst m i c r o s c o p i c a l estimations tend to be 2 or 3 times larger. Whilst v a r i o u s h y p o theses h a v e been suggested to explain this d i s c r e p a n c y , e x p e r i m e n t a l e v i d e n c e is still i n c o n c l u s i v e in delineating both the size and shape o f b o n e m i n e r a l crystallites. In an a t t e m p t to explain p a s t discrepancies, and h e n c e obtain a m o r e a c c u r a t e k n o w l e d g e o f b o n e mineral m o r p h o l o g y , b o t h X - r a y diffraction and bright field electron m i c r o s c o p y h a v e been used t o g e t h e r with the d a r k field i m a g i n g technique, w h i c h to date, has n o t been applied to this particular problem.
Materials and Methods Initially, all techniques were developed using cortical bone, cut from the midfemoral region of 6-month-old female Dutch rabbits. Later, cortical bone from the midfemoral regions of a 2-year-old ox and a 65-year-old human male were used. The human sample was obtained at postmortem, and death was not caused by a primary bone disorder. In experiments to determine the effect of specimen preparation on bone mineral, Johansen and Parks (1960), Ascenzi et al. (1968), and Jackson (1976), have concluded that there are no significant differences (at the 5% level), in the mean size and composition of bone mineral crystals between fixed and unfixedunembedded samples. Consequently, unfixed unembedded bone has been used for this work. The bone samples, approximately 1 mm square, were cemented into an epoxy resin block for support in the LKB4800A ultramicrotome jig. The samples were oriented such that sections were cut in a plane, parallel to the periosteal surface. Using glass knives, the samples were trimmed to approximately 0.1 mm square, then sectioned using a diamond knife with a cutting speed of 5 cm/s and a feed of 60 nm-80 nm between cutting strokes. Distilled water was used as a flotation medium, the sections being picked up immediately on 400 mesh copper grids, and allowed to dry on filter paper before being examined. Microscopical examination was performed using a JEOL.100b electron microscope operated at 80 kV. The conventional observation of a biological specimen is by bright field imaging, where the final image is formed by a beam of electrons transmitted through tbe specimen. The dark field method uses diffracted beams to form the image, and this is achieved by tilting the main beam so that the selected diffracted beam passes through the objective aperture, as
S.A. Jackson eta[.: Bone Mineral Crystal Morphology
Brightfield b,earn
Darkfield
A&0
--
"
beam
Fig. 1. Dark field image formation. The main beam is tilted to allow the desired diffracted beam to pass through the objective aperture
shown in Figure 1. By using the (002) diffracted beam, (the (002) planes being perpendicular to the c-axis in hydroxyapatite), the dark field imaged crystals have their c-axes in the plane of the section, and in a direction within that plane determined by the position of the objective aperture on the diffraction pattern, see Figure 2. A 10 /lm diameter aperture was reqaired to isolate the (002) beam from its neighbours in order to reduce interference effects. After electronoptical rotation had been allowed for, approximately 300 measurements of c-axial length per sample were taken from many micrographs, of crystals which appeared well defined and isolated. The measurements were analysed statistically. Beef catalase was used as a magnification standard. In the X-ray diffraction studies the peak profiles of the (002) and (004) diffracted beams were recorded in triplicate using a standard X-ray diffraetometer and nickel filtered copper radiation. Instrumental broadening of the peak profiles were measured using a sample of rabbit bone powder which had been heated to 800 ~ C for two h, and containing crystals with c-axial lengths well in excess of 100 nm, which produced negligible broadening. To calculate the
Fig. 2. Diffraction pattern and objective aperture images superimposed. The (002) diffracted beams pass through the aperture to form the dark field image
S.A. Jackson et al.: Bone Mineral Crystal Morphology
219
mean c-axial length (t) and the lattice strain factor (e) from the measured peak profiles the method described below was adopted. The correction of Warren (1938), was used to allow for instrumental broadening, i.e., (B~)Z = (/~)2 + (~)z
(1)
where B-~= measured broadening = broadening by crystal = instrumental broadening (~), the peak broadening due to the crystal, was assumed to have components due to both crystal size and crystal strain, where fl89size =
0.92
Results
In the bright field m i c r o g r a p h , F i g u r e 3, b o n e mineral crystals c a n be seen as b o t h l o n g thin d a r k objects, a n d less dense irregular plate like objects. Since an electron m i c r o g r a p h is a t w o d i m e n s i o n a l p r o j e c t i o n o f a three d i m e n s i o n a l object, this c o u l d m e a n t h a t the rod like images are plates viewed on edge, or t h a t the two types of structures are actually present. C o l l a g e n c a n n o t be
(2)
t Cos 0B
strain = 2e tan 0 B
(3)
= X-ray wavelength = c-axiallength 0 B= Bragg angle e = Lattice strain, including both tensile and compressive components. Using the Warren formula for the addition of Gaussian peak s: t
(f189 = ~ size)2 + ~ strain)Z,
(4)
and combining (2) (3) and (4) yields (fl89 Cos
zO~=
+ 4e2 Sin z0~
(5)
The (004) reflection is a second order reflection from the (002) planes (i.e., n = 2 in the Bragg Law 2d Sin 0 B= n2), hence by measuring both the ~ values of the (002) and (004) peak profiles two simultaneous equations can be produced by substitution in Equation (5), there by enabling t and ~ to be calculated.
Fig. 4. High magnification bright field micrograph. Original magnification 350,000x. (100) lattice plane resolution is indicated
Fig. 3. Bright field electron micrograph of rabbit femur. Showing both rod and plate like objects Original magnification 35,000x
Fig. 5. Dark field image formed by using (002) diffracted beams. Original magnification 75,000x. The c-axial direction is indicated, and interference fringes are arrowed
220
seen in this micrograph since unstained collagen is electron transparent. Because of the dense nature of fully calcified bone matrix, single crystals are difficult to isolate, as their outlines tend to merge and overlap. The high magnification micrograph, Figure 4, shows many fringe patterns due mainly to (100) lattice plane resolution. (The (100) planes having a spacing of 0.8169 nm.) The resolved lines can be seen to extend over distances of a few tens of nanometres, indicating that crystals of this size are in fact single crystals, and not composed of aggregations of smaller units. The crystals themselves are almost impossible to isolate, making any form of size measurement difficult. Probably the most important observation to be drawn here, is the fact that even if crystals could be isolated for measurement, the orientation of crystal axes cannot be determined in any bright field image, and so measurements have little relation to the crystal lattice. However, both the problem of isolation and orientation have been overcome by using the method of dark field imaging as described earlier. Figure 5 shows a dark field image of a section of rabbit femur, formed using the (002) diffracted beam, the c-axial direction is indicated. It is known from previous X-ray work that the (002) planes of bone mineral are perpendicular to the longest dimension of the crystal, which is the c-axial direction, and therefore dark field images produced using the (002) beams are images of crystals whose c-axes are in the plane of the section. It is fortunate that the (002) diffracted beam is strong and well isolated. This has meant, by using a 10 #m diameter objective aperture, pure (002) dark field images could be produced, and thereby, reducing Moir6 fringe effects from interfering diffracted beams in overlapping crystals. However some fringes can be seen due to interference between separate (002) diffracted beams. The general appearance of the dark field images show bone mineral crystals to have irregular shapes and sizes. The dimension measured perpendicular to the c-axial direction varies from approximately 5.0 nm to dimensions approaching that of the c-axial length in some instances. Since an (002) dark field image would still be produced if the crystal were rotated about its caxis, this would be consistent with the suggestion that bone mineral has a plate like form, with a thickness of approximately 5.0 nm. From many dark field micrographs only crystals which appeared well isolated were selected for measurement. The c-axial length distributions for rabbit, ox, and human bone have been analysed statistically and are summarized in Figures 6, 7 and 8. Using the same samples of bone as used for the microscopical work, the peak profiles of the (002) and
S.A. Jackson et al.: Bone Mineral Crystal Morphology
(004) diffracted beams were recorded. The measured values of peak width at half maximum intensity (B89 and peak position (20B), are summarized in Table 1, together with the calculated values of mean c-axial
RABBIT BONE
iiiiiiiiiiiii!
c
20
40
60
80
100
120
C-axial length (nm)
Min. 8. 3nm.
Max. Mean 1 S.D. no. Ohm. 32.6nm. 12.8nm.
Fig. 6. Histogram of c-axial length distribution in rabbit bone
OX BONE 70
60
50
40
7O 3O 20
~
0
20
40
60
80
I00
120
C-axial length' (nm) Min.
Max.
Mean
S.D.
&7nm.
113.0 nm.
36.2 nm.
19.0 nm.
Fig. 7. Histogram of c-axial length distribution in ox bone
S.A. Jackson et al.:Bone Mineral Crystal Morphology
221 significant difference, at the 5% level, from the values determined by line broadening.
HUMANBONE
?0 60
Discussion
50
The bright field electron microscopical results strongly suggest that bone mineral crystals are in the form of thin plates, approximately 5.0 nm thick, with an irregular shape and a variable maximum dimension. Whilst m a n y previous workers have assumed the largest dimensions of crystals imaged in the bright field to be the c-axial length, the fact that crystal orientation cannot be determined precludes any accurate measurement. Hence, to obtain quantitative information from electron micrographs, the bone mineral crystal outline has to be clearly defined and the crystal orientation known. The technique of dark field imaging solves these two problems simultaneously. F r o m m a n y dark field micrographs, c-axial measurements were taken for 3 bone types. For rabbit, ox and human bones the mean c-axial length was found to be 32.6 nm, 36.2 nm, and 32.4 nm respectively, the size distributions having standard deviations of 12.8 nm, 19.0 nm and 16.3 nm. High resolution bright field micrographs occasionally show images where the electron density and lattice resolution lines are not consistent over the entire crystal, and this could indicate structural imperfection which would cause lattice strain. Also, since the bone mineral crystals are so small, a large proportion o f unit cells per crystal reside on the surface, and will therefore be deformed. From the line broadening measurements, using a formulation which allowed for lattice strain, the mean c-axial lengths were calculated to be 33.2 rim, 36.1 nm and 32.2 nm for rabbit, ox, and h u m a n bones respec-
40 c
30 20 10 0
JJPJfJ 20
zlO
80
60
I00
120
C-axial length (nm) 6,0 rim.
I00.0 rim.
32,4 nm
16.3nm.
Fig. 8. Histogram of c-axial length distribution in human bone
length (t) and lattice strain factor (e), which contains both tensile and compressive components. To compare the dark field measurements with the X-ray results, from each set of dark field measurements, 10 subsets of data were randomly selected, each subset containing 25 measurements. The mean values for the c-axial length calculated from these subsets are normally distributed about the true mean. Hence, the average mean value calculated from the subsets can be compared with the X-ray measurements using Student's t test. Provided that lattice strain is allowed for in the calculation of mean c-axial length, each bone type measured by the dark field technique shows no
Table 1. Showing a summary of the X-ray diffraction measurements, and the calculated values of (t) the mean c-axial length and (~) the lattice strain factor Sample
B89(002) ~
B~ (004)~
20B(002)~
20B(004)~
Rabbit femur, heated to 800 ~ C for 2 h Rabbit femur, untreated
0.138 0.133 0.142 0.329 0.317 0.323 0.325 0.323 0.313 0.335 0.342 0.329
0.238 0.217 0.229 0.500 0.481 0.513 0.525 0.500 0.494 0.513 0.488 0.519
25.923 25.931 25.925 25.920 25.920 25.910 25.975 25.950 25.963 25.981 26.000 25.994
53.208 53.221 53.242 53.213 53.225 53.225 53.250 53.260 53.240 53.260 53.250 53.260
Ox femur, ~ untreated Human femur, untreated
t (nm)
e (%)
m
m
m
Mean 33.2 SD 1.1
Mean 0.61 SD 0.05
Mean 36.1 SD 1.1
Mean 0.64 SD 0.05
Mean 32.2 SD 1.1
Mean 0.61 SD 0.07
222 tively, which, when c o m p a r e d to the dark field measurements show no significant difference at the 5% level. The lattice strain factor (e) was calculated, and it was found that the m a x i m u m tensile or compressive component had a value o f approximately 0.3% in all samples. This finding is important since it means that the crystals are basically straight in the c-axial direction, and the d a r k field images are therefore of the whole crystal, and not just a small section o f a large curved crystal. In the past, bright field estimations o f c-axial length have tended to be 2 or 3 times longer than those obtained by the X - r a y methods; p r o b a b l y because only the larger crystals could be seen in isolation, together with the fact that lattice strain has not always been allowed for. Furthermore, the inaccuracies involved with bright field images invalidates any comparison between the two methods. Also the two methods have tended to be used in isolation, with different types of bone, making comparison of separate investigations difficult. The dark field technique has enabled consistent measurements o f m e a n c-axial length to be produced using both microscopical and X - r a y methods. It has also enabled the crystal size distribution to be measured within a particular sample.
References Ascenzi, A., Bonucci, E., Steve-Bocciarelli, D.: Fine structure of bone mineral in different experimental conditions. IV European Regional Conf. on Electron Microscopy. Rome, pp. 431-433 (1968)
S.A. Jackson et al.: Bone Mineral Crystal Morphology CarlstrSm, D.: X-ray crystallographic studies on apatites and calcified structures. Acta Radiologica. Supp. 121, 33-37 (1955) Carlstr6m, D., Finean, J.B.: X-ray diffraction studies on the ultrastructure of bone. Biochim. Biophys. Acta 13, 183-191 (1954) EngstrSm, A.: Aspects of the molecular structure of bone. In: The biochemistry and physiology of bone (G.H. Bourne, ed.). New York: Vol. 1, (2nd edition), pp. 237-257 (1972) Engstr6m, A., Finean, J.B.: The low angle x-ray diffraction of bone. Nature 171,564 (I953) Fernandez-Moran, J., EngstrSm, A.: Electron microscopy and xray diffraction of bone. Biochim. Biophys. Acta 23, 260-264 (1957) Jackson, S.A.: The morphology of bone mineral. PhD. Thesis, University of Surrey (1976) Johansen, D.M.D., Parks, H.F.: Electron microscopic observations on the 3-dimensional morphology of apatite crystallites of human dentine and bone. J. Biophys. Biochem. Cytol. 7 (No. 4), 743-745 (1960) Myers, H.M., EngstrSm, A.: A note on the organisation of hydroxyapatite in calcified tissues. Exp. Cell. Res. 40, 182-185 (1965) Robinson, R.A.: An electron microscopic study of the crystalline inorganic component of bone and its relationship to the organic matrix. J. Bone and Joint Surg. 34, 389434 (1952) Robinson, R.A., Watson, M.L.: Collagen-crystal relationships in bone as seen in the electron microscope. Anat. Res. 114, 383410 (1952) Steve-Bocciarelli, D.: Morphology of crystallites in bone. Calc. Tiss. Res. 5,261-269 (1969) Stuhler, R.: Uber den Feinbau des Knochens, Eine RSntgen - Feinstruktur Untersuchung. Fortschr. Gebiete RSntgenstrahln. 57, 231-234 (1938) Warren, B.E., Biscoe, J.: The structure of silica glass by x-ray diffraction studies. J. Am. Ceram. Soc. 21, 49-54 (1938) Received July lO/Revised December 15~Accepted December 21, 1977
