Bone, 13, 4433446, (1992) Printed in the USA. All rights reserved
Copyright
8756-3282192 $5.00 + .OO 0 1992 Pergamon Press Ltd.
Do Different Fluorochrome Labels Give Equivalent Histomorphometric Information? T.C.SUN,’ S. MORI,’ J. ROPER,’ C. BROWN,t T. HOOSER’ and D. B. BURR’ ’ Departments of Anatomy and Orthopedic Surgery Biomechanics and Biomaterials Research Center, Indiana Vniversiv, Indianapolis, IN 46202, U.S.A. * Department of Orthopedic Surgery, The University of the Ryukyus, Japan Address for correspondence and reprints: David B. Burr, Department Medicine, Indianapolis, IN 46202, U.S.A.
of Anatomy,
Abstract
School of
ferent fluorochromes might give different information, resulting in incorrect interpretations or conclusions. The purpose of the present work is to test these commonly used fluorochromes to determine whether differences exist in their measured mineralizing surface (MS) or mineral apposition rate (MAR) that could affect the calculation of bone formation rate (BFR). A second goal is to define the methodological or biological mechanisms underlying any measured differences.
Substances that bind calcium are given to determine where and how fast bone is forming. Several vital dyes are used (tetracycline, calcein, alizarin, xylenol), but it is not known whether the histomorphometric results they provide are equivalent. This work tests whether different fluorocbrome labels give the same results when they are quantitatively measured. Twelve-week-old rats (n = 58) were divided into six groups and given double labels IP of calcein, tetracycline HCI, alizarin complexone, or xylenol using a 1-7-1 scheme. Two other groups received either calcein followed by tetracycline, or tetracycline followed by calcein. Our results show that (a) tetracycline hydrochloride leads to a significant underestimation of mineralizing surface when given as the second label, probably because of its weak fluorescence; (b) there were no differences among any of the non-tetracycline labels for any of the histomorphometric measurements, (c) there was no evidence of osteoblast suppression with any label; and (d) there was no evidence that tetracycline caused increased osteoblast resting periods. Key Words: Tetracycline-Bone chromes-Histomorphometry~steoblast.
MS 259, 635 Bamhill Drive, Indiana University,
Materials and Methods Fifty-eight, 12-week-old Sprague-Dawley rats (Harlan), each weighing approximately 250 g, were randomly divided into six groups and double-labeled intraperitoneally with either one fluorochrome or two different fluorochromes, using a 1-7-l schedule: (10 rats) calcein (Sigma Chemical Co) 10 mg/kg. (10 rats) tetracycline hydrochloride (Sigma) 30 mgikg. (10 rats) alizarin complexone (Sigma) 25 mglkg. (9 rats) xylenol orange (Sigma) 90 mg/kg. (9 rats) first label: calcein; second label: tetracycline. Dosages were the same as for Groups 1 and 2. Group 6: (10 rats) fust label: tetracycline; second label: calcein. Dosages were the same as for Groups 1 and 2.
Group Group Group Group Group
remodeling-Fluoro-
1: 2: 3: 4: 5:
Three days after the second labeling, the rats were sacrificed using established animal care protocols, using ethyl ether (reagent A.C.S. UN1 155 Fisher Scientific) followed by exsanguination and cervical dislocation. The tibias were harvested and fixed for five days in 10% buffered formalin, then transferred to 70% ETOH. Five transverse sections each 50 micrometers thick were cut from the unembedded tibia1 mid-shaft with a .22 mm diamond wire Histosaw (DDM-P 216; Medim, Germany). The sections were viewed at 156.25~ magnification using fluorescence microscopy with a narrow-band, blue-light filter block (450490 nm barrier, 510 nm excitation filters). The fluorescence-labeled periosteal and endocortical surfaces were digitized at 150X using a drawing tube and a digitizing board interfaced to an IBM Model 25 computer. Bioquant software was used to collect the following histomorphometric parameters from up to three sections of each rat tibia. MS/BS (mm/mm): Mineralizing surface (total perlosteal or endocortical bone surface referent, designated as MS/BS.Ps or MS/BS.Ec, respectively). Mineralizing surface was calculated as
Introduction Substances that bind to calcium are given to animals and people to determine from biopsies where and how fast bone is forming (Milch et al. 1957; Pa&t et al. 1987; Frost 1969). There are several different kinds of these fluorochrome dyes, and various regimens have been used in histomorphometric studies. In our laboratory we use tetracycline, calcein, alizarin complexone, and xylenol orange. These fluorochromes show different colors and have the advantage of giving sequential information when multiple labels are used. However, it is not known whether the histomorphometric results they provide are equivalent. It is possible that they are incorporated into bone differently, that they are viewed differently because their intensities are not the same, or that uptake varies because dosage differs (Hattner et al. 1977). Also, they may alter osteoblast activity or recruitment (Simmons et al. 1981) or inhibit mineralization rate (Simmons et al. 1983). For any of these reasons, histomorphometric data based on dif443
T. C. Sun et al.: Are fluorochrome
444
Table I. Mean and
SD of the bone forming
I
Gr.
Gr. 2 Tetracycline
Calcein MAR.Ps MAR.Ec MSIBS .Ps MS/BS.Ec
1.63 2.07 0.42 0.37
BFRPs BFR.Ec
0.78 ” 0.33 0.78 2 0.36
2 + k lr
parameters
0.35 0.59 0.16 0.12
1.56 2.22 0.23 0.26
rt 2 ” 2
0.48 0.43 0.10 0.07
0.39 ? 0.25 0.60 2 0.16
labels equivalent’?
in the six groups” Cr. 3 Alizarin
1.70 2.10 0.36 0.36
Gr. 4 Xylenol
2 0.25 i- 0.33 -t- 0.18 2 0.12
0.62 2 0.31 0.74 + 0.25
1.78 2.43 0.48 0.31
2 t ? rt
Gr. 5 Calceinketracycline
0.27 0.43 0.10 0.10
1.88 2.34 0.29 0.35
0.86 i 0.27 0.76 -+ 0.29
k ? 2 2
0.24 0.22 0.07 0.09
0.55 t 0.18 0.81 ? 0.21
Gr. 6 Tetracyclinekalcein 1.64 2.03 0.36 0.35
2 lr 2 ”
0.61 0.39 0.15 0.14
0.63 ” 0.30 0.73 ?I 0.44
Pb 0.59 0.22 0.002 0.23
0.02 0.79
“MAR, microns/day; MS/BS, mm/mm; BPR, microns/day. bProbability of differences among the six groups by ANOVA.
double-labeled surface plus one-half single-labeled surface, as defined by Parfitt et al. ( 1987) and Martin ( 1989). MAR (microns/day): Mineral apposition rate on the periosteal (MAR.Ps) or endocortical (MAR.Ec) surface was calculated as the area enclosed by the double-labeled surface, divided by the length of the double-labeled surface. BFR, surface referent (microns/day): Bone formation rate on the periosteal (BFR.Ps) or endocortical (BFR.Ec) surface. BFR = MSIBS x MAR. DL/SL: Total double-labeled surface divided by total singlelabeled surface. The six groups were statistically compared for each parameter using ANOVA (Systat, Inc.), followed by post-hoc Tukey tests. When two different fluorochromes were used (Groups 5 and 6), the lengths of the two labels were measured and compared using a paired t-test.
Results No significant differences were found in MAR (Table I) or in DL/SL ratio (Table II) among the six groups on either the periosteal or endocortical surfaces. However, the periosteal DL/SL ratios from Groups 2 (tetracycline) and 5 (calceinitetracycline) were less than half of the pooled mean of the other groups (pooled mean = 3.07 ? 2.15; Group 2 mean = 1.15 ? 1.15: Group 5 mean = 1.44 ? 1.15). The MS/BS.Ps and BFR.Ps were significantly different among the six groups @ = 0.002 andp = 0.02, respectively). Tukey tests showed that BFR.Ps in Group 2 (tetracycline) was significantly less than that in Group 4 (xylenol; p < 0.02); MS/BS.Ps in Group 2 was significantly less than that in Groups 1 (calcein) and 4 (xylenol; p < .03 and p < 0.005, respectively); and MS/BS.Ps in Group 5 (calceini tetracycline) was significantly less than that in Group 4 (p < 0.05). A trend to lower MS/BS.Ec and BFR.Ec in Group 2 was observed even though statistical significance could not be demonstrated for the endocortical surface. Power calculations show 87% power for MS/BS.Ec and 61% power for BFR.Ec to detect an effect of the same magnitude as that noted on the periosteal surface. These power calculations suggest that it is unlikely that the endocortical surface in Group 2 would demonstrate the mag-
nitude of difference in MS/BS found on the periosteal surface, even with a larger sample size. When calcein was used as the first label and tetracycline used as the second label (Group 5), tetracycline showed significantly less MYBS than calcein on both periosteal and endocortical surfaces (p < 0.001) (Fig. I). No difference was found when they were used in a reversed order (tetracycline followed by calcein. Group 6).
Discussion The problem that different fluorochromes could give different histomorphometric results has been recognized recently by Parffit et al. (1991). They found that the oxytetracycline label was always shorter than the demethylchlortetracycline label, regardless of which was used first, suggesting some intrinsic pharmacokinetic differences between the two compounds. We were unable to detect any significant differences in MAR among the different fluorochrome label groups. This indicates that these fluorochromes do not affect the individual vigor of fully differentiated osteoblasts. Our data also showed, however, that the periosteal MSiBS and BFR calculated from Group 2, which was given a double tetracycline label, was significantly less than those in several other groups (Table I). MSlBS and BFR were not significantly less on the endocottical surfaces in Group 2. When used as the second label following calcein (Group 5), tetracycline showed shorter labeled surface at both periosteum and endosteum (Fig. 1). This could lead to the conclusion that tetracycline inhibits the differentiation or recruitment of osteoblasts, which would have the effect of reducing the total tetracycline-labeled surface. Under closer inspection, however, our data do not support this explanation. When tetracycline was used as the first label and calcein as the second label (Group 6), tetracycline had no negative effect on calcein-labeled MS. The length of the calceinlabeled surface was identical to that in Group 5, in which calcein was given first. Tetracycline only showed less labeled surface when it was given as the last label. Interestingly, the DLlSL ratio shows that less double label was detected in those groups in which tetracycline was given as the sole fluorochrome (Group 2)
Table II. Ratio of double- to single-labeled surface
p-eriosteum endosteum “Probability
Gr. 1
cr. 2
Gr. 3
Cr. 4
Gr. 5
Gr. 6
PA
2.8 ” 1.96 3.42 ” 2.10
1.15 t 1.15 2.56 + 1.90
2.75 2 2.57 3.52 k 2.48
3.01 f 1.51 2.52 2 0.68
1.44 k 1.15 2.43 2 1.31
3.71 2 4.38 4.3 2 2.75
0.18 0.27
of differences
among the six groups using ANOVA.
T. C. Sun et al.: Are fluorochrome
60
labels equivalent?
445
-
MSlBS %
TT
50
Cl 1st label 2nd label
MS/BS.Ps I
MS/BS.Ec
1 .Calcein P.Tetracycline
MS/BS.Ps 1 1
MSIBS.Ec
l.Tetracycline 2.Calcein
/
Fig. 1. Surfaces labeled with tetracycline and calcein in Group 5 (calcein followed by tetracycline) and Group 6 (tetracycline followed by calcein). Periosteal (MS/BS.Ps) and endocortical (MS/BS.Ec) mineralizing surfaces are shown. The shaded bar represents the second label administered in each case. When calcein is given first, there is a significant @ = ,001) reduction in measured tetracycline-labeled MSlBS on both periosteal and endocortical surfaces. When tetracycline is given first, there is no difference in measured MS/BS between tetracycline and calcein. Error bars represent the sample standard deviation.
or given last (Group 5) (Table II). The DL/SL ratio was less than half the pooled mean of the other groups. The DUSL ratio was not less when tetracycline was given as the first label followed by calcein. Thus, the lower MS/BS and BFR of tetracycline in Groups 2 and 5 is most likely not because it fails to become incorporated into the bone (either due to some intrinsic pharmacokinetic property or due to osteoblastic suppression), but rather because some methodological factor prevents its accurate measurement when it is the second label. The data that suggest that tetracycline interferes with bone mineralization has been primarily derived from growing animals (see Harris et al. 1968 and Simmons et al. 1983 for reviews), or from bone rudiments in culture (Saxen 1965). In some cases, extremely high dosages (Harris et al. 1968) or prolonged daily treatment with tetracycline (Simmons et al. 1983) was required to demonstrate significant inhibition of mineralization. Simmons et al. (1981) demonstrated increased labeled surface following the administration of a second tetracycline-HCl label. There is no data showing that tetracycline interferes with mineralization or osteoblast development in adult skeletons at the dosages used for experimental or clinical study. In fact, Harris et al. ( 1968), using several different tetracycline compounds, showed that mineralization deficits could only be detected following either oral or IV administration of tetracyclines in dosages exceeding 60 mg/kg per day. This is twice the dosage used in our study. These data also rule out the possibility of label escape (Schwartz & Reeker 1982; Frost 1983) or a change in the ONOFF state of bone-forming activity (Frost 1980; Parfitt et al. 1977). The finding that the MAR of the tested dyes was no different among the six groups would exclude a more negative effect of tetracycline on the ON-OFF state of bone formation as the explanation of its shorter labeled surface. Moreover, the DUSL ratio in our study is several times higher than that calculated from data collected on trabecular surfaces of the caudal vertebrae in 12-week-old Wistar-Lewis rats by Baron et al.
(1984). The use of relatively thick sections in our study should lead to an underestimation of the DL/SL ratio, but in fact our values are higher than those reported by other investigators on a variety of different surfaces in several different animals, including humans (Martin 1989). This observation suggests that there were not significant periods of inactivity in bone formation during the interlabel period in our study, and supports the conclusions of Ott (1989), who found no interruption of mineralization following triple labeling of bone. The yellow fluorescence of tetracycline-HCl is weak. It tends to be masked by or mixed with the nonspecific green fluorescence that is always present in fluorescence microscopy of bone. When tetracycline-HCl is used as the second label following calcein, its lower visibility becomes a significant problem, resulting in underestimation of the MS/BS measurement. In part, this may be due to the strong intensity of the calcein label, but also is related to the difficulty of detecting the pale yellow tetracycline fluorescence when it is along the edge of the bone. When given as the first label followed by calcein, tetracycline is easier to recognize, delineate, and separate from background fluorescence. The strong calcein fluorescence is less difficult to detect than tetracycline fluorescence when it is close to the edge of the bone. Another possible explanation for the shorter length of the second tetracycline label could be that recently deposited tetracycline is more susceptible to chelation during formalin fixation. This explanation is unlikely for several reasons. First, there is no evidence that five days’ fixation with properly buffered formalin will remove surface calcium to which tetracycline is bound. Buffered formalin is neutral, not acidic. Second, studies comparing labeled surfaces of bone that was either preserved in alcohol or fixed in cold neutral buffered formalin for up to a week show that neutral buffered formalin does not reduce labeled surface length (Garetto, personal communication). Third, if the formalin were to chelate surface calcium during fixation, then the
446
T. C. Sun et al.: Are fluorochrome
label in Group 6 should be significantly shorter than the tetracycline label in this group. This is not the case (Fig. I). Therefore, the idea that the second label is more easily removed by formalin fixation cannot be an explanation for the reduced length of the second tetracycline label. We were not able to detect any differences in MAR. MSIBS, or BF’R when comparing alizarin complexone-labeled animals to animals from the other labeling groups. This conflicts with the results of Harris et al. (1968), who demonstrated a direct inhibition in mineralization of up to 60% by alizarin Red S. They noted reduced crystal formation, inhibition of crystal growth. and disturbances in collagen synthesis or aggregation (Harris et al. 1964). Our results show, however, that alizarin-complexone provides histomorphometric results comparable to those of calcein and xylenol. calcein
Frost. H
M
Tetracycline-based
T‘r.\.\u~ Re&. 3:21 l-237; Fro\t.
H. M.
Merub. Fro\t.
H. M
Boca Raton.
formation
FL: CRC
W.
Hattner.
R. S.,
tetracychne Norman,
bone-forming
of the labehng
“escape error ”
Techniques and interpretation
E. The rn viva inhibition Jr. Surg. 46A:493-508; 1964
U.: Radin,
1.: Hamblen,
D. L.: Haywood,
C/in. Orthop. RF/. Res. 61:52%60;
llnlcki,
L. P.; Hedge,
to the detectability Schaefer,
H. C.
The dose-response
Cobum.
I. W.:
ot
1968. relationship
of labeled osteons by fluorescence K.;
of bone
E. A. The inhibition
rn wvo.
A. W.:
centers
Press: 1983; pp. 133-142.
Red S. J. Bone
H.; Lavorgna,
ossification
III lam&u
Correction
D. F.: Friberg,
by Alirarin
Culcif.
1980
Bone histomorphumerry
ed.
W. H.: Travis,
Hams,
and “off”
RPI. Rex. 2S:l67~170;
Bone histomorphometry: R. R..
analyala of bone remodeling.
1969.
Resttng seams. “On”
Bone DIS.
Reeker,
Harm,
histological
labels equivalent’?
DeLuca,
ot
microscopy.
H. F.;
Fraser.
D
;
Vitumin D. biomechanica/, chemicul and clinrcul aspects related to calcium metabolism. New York: de Gruyter; Grigoleit.
H. G.; Herrath,
D. V.,
eds.
1977; pp. 377-380. Martin,
Conclusion
R. B. Label
section thickness.
Tetracycline hydrochloride, with its weak fluorescence and lower visibility, causes an underestimation of MS/BS and BFR when used as the final label. When used as the first label followed by other labels, its estimation of MS/BS and BFR is not significantly different than that given by other fluorochrome labels. The use of tetracycline-HCI has no significant effect on MAR. Calcein, alizarin complexone, and xylenol orange provide comparable estimates of MAR, MS/BS, and BFR when used as sequential labels. There was no evidence that tetracycline caused suppression of osteoblasts or increased periods of osteoblastic inactivity.
Milch,
R. A.;
Rall,
escape theory
D. P.; Tobie.
Natl. Cancer Inst. 19:87-91: Ott.
S. M.
Bone
A. M.:
fo‘ormation fails
Villanueva,
H. Classification tetracychne P. J ed. Parfitt,
Meunier.
J. E. Bone localizanon
:
“on-off”
to show
Drezner.
A. R.; Crouch.
Evidence
M.
K.:
M.
M.:
Paris: Armour
Glorieux.
F. H.;
A
M..
1.. Villanueva,
A. R.;
length between demethylchlortetracycline the interpretation
of mineralization. Montagu; Kanis,
Malluche,
H.:
nomenclaNomencla-
1987.
Shih,
M.
S. Difference
and oxytetracycline:
of bone histomorphometric
and
Meunter.
1977; pp. 229-310.
I. A.;
Histomorphometry
J Bone Min. Res 2:595%610:
Foldes,
C. H. E.. Duncan.
use of cell morphology
R. R. Bone histomorphometry
ture, symbols and units. Report of the ASBMR
Parfitt,
in postmenopausal
1989.
Mathews,
for intermittency
P. I.; Ott. S. M.; Reeker,
ture Committee.
J
of the tetracyclines.
pattern
l):S415;
of osteoid seams by combined
labeling.
and
1957.
Bone histomorphometry.
A. M
The effects of resting periods
1989.
J. Bone Min. Res. 4(Suppl.
osteoporosis. Parfitt,
revtsited:
Bone 10:255-264;
data. C&if.
in label
Implications Tissue
for
Int. 48:74-
77: 1991. Schwartz.
Acknowledgments:
This work was completed
with support from NIH grant T35 HL07584-07, “Short-Term Training for Students in Health and NIH grants R55 AR40655 and ROI AR39708. The Professions”; authors appreciate the help of Drs. Charles Turner, Jae Rho, Tomoaki Yoshikawa, Alonza Perry, and Mark Seifert, and of Lisa Pedigo for assistance during various portions of the experiment. The authors also extend their thanks to Dr. R. B. Martin, who read and commented on a prior draft of the manuscript.
M. P.; Reeker, R. R. The label escape error: Determination
bone-forming
surface in histologic
double labels
Metab.
Simmons,
Bone Dis.
D. I.: Teitelbaum.
of rabbit
osteoblasts
sections of bone measured
Rel. Res. 4:237-241;
S. L.: Rosenberg,
by tetracycline.
Metab.
R.; Tress,
becular
R.; Vignery,
bone morphology,
skeletal maturation.
A.
Ewdence
dynamic
of sequential
histomorphometry,
Anar. Rec. 208: 137-145:
1984.
remodelmg
m rat tra-
and changes during
active
1982.
G. D. Altered
metabolic
rhythm
Bone Dis. Rel. Res. 3:51-54:
1981. Simmons.
D
J.: Chang, S-L.;
Russell, J. E.; Grazman,
The effect of protracted tetracycline
C/in
Orthop. Rel. Res. 180:25>259;
B.; Webster,
D.; Oloff,
C
treatment on bone growth and maturation. 1983.
References Baron,
ofthe
by tetraycline
Dare Received: May 12, 1992 Dare Revised: July 20, 1992 Date Accepted: July 23, 1992
Copyright
8756-3282192 $5.00 + .OO 0 1992 Pergamon Press Ltd.
Do Different Fluorochrome Labels Give Equivalent Histomorphometric Information? T.C.SUN,’ S. MORI,’ J. ROPER,’ C. BROWN,t T. HOOSER’ and D. B. BURR’ ’ Departments of Anatomy and Orthopedic Surgery Biomechanics and Biomaterials Research Center, Indiana Vniversiv, Indianapolis, IN 46202, U.S.A. * Department of Orthopedic Surgery, The University of the Ryukyus, Japan Address for correspondence and reprints: David B. Burr, Department Medicine, Indianapolis, IN 46202, U.S.A.
of Anatomy,
Abstract
School of
ferent fluorochromes might give different information, resulting in incorrect interpretations or conclusions. The purpose of the present work is to test these commonly used fluorochromes to determine whether differences exist in their measured mineralizing surface (MS) or mineral apposition rate (MAR) that could affect the calculation of bone formation rate (BFR). A second goal is to define the methodological or biological mechanisms underlying any measured differences.
Substances that bind calcium are given to determine where and how fast bone is forming. Several vital dyes are used (tetracycline, calcein, alizarin, xylenol), but it is not known whether the histomorphometric results they provide are equivalent. This work tests whether different fluorocbrome labels give the same results when they are quantitatively measured. Twelve-week-old rats (n = 58) were divided into six groups and given double labels IP of calcein, tetracycline HCI, alizarin complexone, or xylenol using a 1-7-1 scheme. Two other groups received either calcein followed by tetracycline, or tetracycline followed by calcein. Our results show that (a) tetracycline hydrochloride leads to a significant underestimation of mineralizing surface when given as the second label, probably because of its weak fluorescence; (b) there were no differences among any of the non-tetracycline labels for any of the histomorphometric measurements, (c) there was no evidence of osteoblast suppression with any label; and (d) there was no evidence that tetracycline caused increased osteoblast resting periods. Key Words: Tetracycline-Bone chromes-Histomorphometry~steoblast.
MS 259, 635 Bamhill Drive, Indiana University,
Materials and Methods Fifty-eight, 12-week-old Sprague-Dawley rats (Harlan), each weighing approximately 250 g, were randomly divided into six groups and double-labeled intraperitoneally with either one fluorochrome or two different fluorochromes, using a 1-7-l schedule: (10 rats) calcein (Sigma Chemical Co) 10 mg/kg. (10 rats) tetracycline hydrochloride (Sigma) 30 mgikg. (10 rats) alizarin complexone (Sigma) 25 mglkg. (9 rats) xylenol orange (Sigma) 90 mg/kg. (9 rats) first label: calcein; second label: tetracycline. Dosages were the same as for Groups 1 and 2. Group 6: (10 rats) fust label: tetracycline; second label: calcein. Dosages were the same as for Groups 1 and 2.
Group Group Group Group Group
remodeling-Fluoro-
1: 2: 3: 4: 5:
Three days after the second labeling, the rats were sacrificed using established animal care protocols, using ethyl ether (reagent A.C.S. UN1 155 Fisher Scientific) followed by exsanguination and cervical dislocation. The tibias were harvested and fixed for five days in 10% buffered formalin, then transferred to 70% ETOH. Five transverse sections each 50 micrometers thick were cut from the unembedded tibia1 mid-shaft with a .22 mm diamond wire Histosaw (DDM-P 216; Medim, Germany). The sections were viewed at 156.25~ magnification using fluorescence microscopy with a narrow-band, blue-light filter block (450490 nm barrier, 510 nm excitation filters). The fluorescence-labeled periosteal and endocortical surfaces were digitized at 150X using a drawing tube and a digitizing board interfaced to an IBM Model 25 computer. Bioquant software was used to collect the following histomorphometric parameters from up to three sections of each rat tibia. MS/BS (mm/mm): Mineralizing surface (total perlosteal or endocortical bone surface referent, designated as MS/BS.Ps or MS/BS.Ec, respectively). Mineralizing surface was calculated as
Introduction Substances that bind to calcium are given to animals and people to determine from biopsies where and how fast bone is forming (Milch et al. 1957; Pa&t et al. 1987; Frost 1969). There are several different kinds of these fluorochrome dyes, and various regimens have been used in histomorphometric studies. In our laboratory we use tetracycline, calcein, alizarin complexone, and xylenol orange. These fluorochromes show different colors and have the advantage of giving sequential information when multiple labels are used. However, it is not known whether the histomorphometric results they provide are equivalent. It is possible that they are incorporated into bone differently, that they are viewed differently because their intensities are not the same, or that uptake varies because dosage differs (Hattner et al. 1977). Also, they may alter osteoblast activity or recruitment (Simmons et al. 1981) or inhibit mineralization rate (Simmons et al. 1983). For any of these reasons, histomorphometric data based on dif443
T. C. Sun et al.: Are fluorochrome
444
Table I. Mean and
SD of the bone forming
I
Gr.
Gr. 2 Tetracycline
Calcein MAR.Ps MAR.Ec MSIBS .Ps MS/BS.Ec
1.63 2.07 0.42 0.37
BFRPs BFR.Ec
0.78 ” 0.33 0.78 2 0.36
2 + k lr
parameters
0.35 0.59 0.16 0.12
1.56 2.22 0.23 0.26
rt 2 ” 2
0.48 0.43 0.10 0.07
0.39 ? 0.25 0.60 2 0.16
labels equivalent’?
in the six groups” Cr. 3 Alizarin
1.70 2.10 0.36 0.36
Gr. 4 Xylenol
2 0.25 i- 0.33 -t- 0.18 2 0.12
0.62 2 0.31 0.74 + 0.25
1.78 2.43 0.48 0.31
2 t ? rt
Gr. 5 Calceinketracycline
0.27 0.43 0.10 0.10
1.88 2.34 0.29 0.35
0.86 i 0.27 0.76 -+ 0.29
k ? 2 2
0.24 0.22 0.07 0.09
0.55 t 0.18 0.81 ? 0.21
Gr. 6 Tetracyclinekalcein 1.64 2.03 0.36 0.35
2 lr 2 ”
0.61 0.39 0.15 0.14
0.63 ” 0.30 0.73 ?I 0.44
Pb 0.59 0.22 0.002 0.23
0.02 0.79
“MAR, microns/day; MS/BS, mm/mm; BPR, microns/day. bProbability of differences among the six groups by ANOVA.
double-labeled surface plus one-half single-labeled surface, as defined by Parfitt et al. ( 1987) and Martin ( 1989). MAR (microns/day): Mineral apposition rate on the periosteal (MAR.Ps) or endocortical (MAR.Ec) surface was calculated as the area enclosed by the double-labeled surface, divided by the length of the double-labeled surface. BFR, surface referent (microns/day): Bone formation rate on the periosteal (BFR.Ps) or endocortical (BFR.Ec) surface. BFR = MSIBS x MAR. DL/SL: Total double-labeled surface divided by total singlelabeled surface. The six groups were statistically compared for each parameter using ANOVA (Systat, Inc.), followed by post-hoc Tukey tests. When two different fluorochromes were used (Groups 5 and 6), the lengths of the two labels were measured and compared using a paired t-test.
Results No significant differences were found in MAR (Table I) or in DL/SL ratio (Table II) among the six groups on either the periosteal or endocortical surfaces. However, the periosteal DL/SL ratios from Groups 2 (tetracycline) and 5 (calceinitetracycline) were less than half of the pooled mean of the other groups (pooled mean = 3.07 ? 2.15; Group 2 mean = 1.15 ? 1.15: Group 5 mean = 1.44 ? 1.15). The MS/BS.Ps and BFR.Ps were significantly different among the six groups @ = 0.002 andp = 0.02, respectively). Tukey tests showed that BFR.Ps in Group 2 (tetracycline) was significantly less than that in Group 4 (xylenol; p < 0.02); MS/BS.Ps in Group 2 was significantly less than that in Groups 1 (calcein) and 4 (xylenol; p < .03 and p < 0.005, respectively); and MS/BS.Ps in Group 5 (calceini tetracycline) was significantly less than that in Group 4 (p < 0.05). A trend to lower MS/BS.Ec and BFR.Ec in Group 2 was observed even though statistical significance could not be demonstrated for the endocortical surface. Power calculations show 87% power for MS/BS.Ec and 61% power for BFR.Ec to detect an effect of the same magnitude as that noted on the periosteal surface. These power calculations suggest that it is unlikely that the endocortical surface in Group 2 would demonstrate the mag-
nitude of difference in MS/BS found on the periosteal surface, even with a larger sample size. When calcein was used as the first label and tetracycline used as the second label (Group 5), tetracycline showed significantly less MYBS than calcein on both periosteal and endocortical surfaces (p < 0.001) (Fig. I). No difference was found when they were used in a reversed order (tetracycline followed by calcein. Group 6).
Discussion The problem that different fluorochromes could give different histomorphometric results has been recognized recently by Parffit et al. (1991). They found that the oxytetracycline label was always shorter than the demethylchlortetracycline label, regardless of which was used first, suggesting some intrinsic pharmacokinetic differences between the two compounds. We were unable to detect any significant differences in MAR among the different fluorochrome label groups. This indicates that these fluorochromes do not affect the individual vigor of fully differentiated osteoblasts. Our data also showed, however, that the periosteal MSiBS and BFR calculated from Group 2, which was given a double tetracycline label, was significantly less than those in several other groups (Table I). MSlBS and BFR were not significantly less on the endocottical surfaces in Group 2. When used as the second label following calcein (Group 5), tetracycline showed shorter labeled surface at both periosteum and endosteum (Fig. 1). This could lead to the conclusion that tetracycline inhibits the differentiation or recruitment of osteoblasts, which would have the effect of reducing the total tetracycline-labeled surface. Under closer inspection, however, our data do not support this explanation. When tetracycline was used as the first label and calcein as the second label (Group 6), tetracycline had no negative effect on calcein-labeled MS. The length of the calceinlabeled surface was identical to that in Group 5, in which calcein was given first. Tetracycline only showed less labeled surface when it was given as the last label. Interestingly, the DLlSL ratio shows that less double label was detected in those groups in which tetracycline was given as the sole fluorochrome (Group 2)
Table II. Ratio of double- to single-labeled surface
p-eriosteum endosteum “Probability
Gr. 1
cr. 2
Gr. 3
Cr. 4
Gr. 5
Gr. 6
PA
2.8 ” 1.96 3.42 ” 2.10
1.15 t 1.15 2.56 + 1.90
2.75 2 2.57 3.52 k 2.48
3.01 f 1.51 2.52 2 0.68
1.44 k 1.15 2.43 2 1.31
3.71 2 4.38 4.3 2 2.75
0.18 0.27
of differences
among the six groups using ANOVA.
T. C. Sun et al.: Are fluorochrome
60
labels equivalent?
445
-
MSlBS %
TT
50
Cl 1st label 2nd label
MS/BS.Ps I
MS/BS.Ec
1 .Calcein P.Tetracycline
MS/BS.Ps 1 1
MSIBS.Ec
l.Tetracycline 2.Calcein
/
Fig. 1. Surfaces labeled with tetracycline and calcein in Group 5 (calcein followed by tetracycline) and Group 6 (tetracycline followed by calcein). Periosteal (MS/BS.Ps) and endocortical (MS/BS.Ec) mineralizing surfaces are shown. The shaded bar represents the second label administered in each case. When calcein is given first, there is a significant @ = ,001) reduction in measured tetracycline-labeled MSlBS on both periosteal and endocortical surfaces. When tetracycline is given first, there is no difference in measured MS/BS between tetracycline and calcein. Error bars represent the sample standard deviation.
or given last (Group 5) (Table II). The DL/SL ratio was less than half the pooled mean of the other groups. The DUSL ratio was not less when tetracycline was given as the first label followed by calcein. Thus, the lower MS/BS and BFR of tetracycline in Groups 2 and 5 is most likely not because it fails to become incorporated into the bone (either due to some intrinsic pharmacokinetic property or due to osteoblastic suppression), but rather because some methodological factor prevents its accurate measurement when it is the second label. The data that suggest that tetracycline interferes with bone mineralization has been primarily derived from growing animals (see Harris et al. 1968 and Simmons et al. 1983 for reviews), or from bone rudiments in culture (Saxen 1965). In some cases, extremely high dosages (Harris et al. 1968) or prolonged daily treatment with tetracycline (Simmons et al. 1983) was required to demonstrate significant inhibition of mineralization. Simmons et al. (1981) demonstrated increased labeled surface following the administration of a second tetracycline-HCl label. There is no data showing that tetracycline interferes with mineralization or osteoblast development in adult skeletons at the dosages used for experimental or clinical study. In fact, Harris et al. ( 1968), using several different tetracycline compounds, showed that mineralization deficits could only be detected following either oral or IV administration of tetracyclines in dosages exceeding 60 mg/kg per day. This is twice the dosage used in our study. These data also rule out the possibility of label escape (Schwartz & Reeker 1982; Frost 1983) or a change in the ONOFF state of bone-forming activity (Frost 1980; Parfitt et al. 1977). The finding that the MAR of the tested dyes was no different among the six groups would exclude a more negative effect of tetracycline on the ON-OFF state of bone formation as the explanation of its shorter labeled surface. Moreover, the DUSL ratio in our study is several times higher than that calculated from data collected on trabecular surfaces of the caudal vertebrae in 12-week-old Wistar-Lewis rats by Baron et al.
(1984). The use of relatively thick sections in our study should lead to an underestimation of the DL/SL ratio, but in fact our values are higher than those reported by other investigators on a variety of different surfaces in several different animals, including humans (Martin 1989). This observation suggests that there were not significant periods of inactivity in bone formation during the interlabel period in our study, and supports the conclusions of Ott (1989), who found no interruption of mineralization following triple labeling of bone. The yellow fluorescence of tetracycline-HCl is weak. It tends to be masked by or mixed with the nonspecific green fluorescence that is always present in fluorescence microscopy of bone. When tetracycline-HCl is used as the second label following calcein, its lower visibility becomes a significant problem, resulting in underestimation of the MS/BS measurement. In part, this may be due to the strong intensity of the calcein label, but also is related to the difficulty of detecting the pale yellow tetracycline fluorescence when it is along the edge of the bone. When given as the first label followed by calcein, tetracycline is easier to recognize, delineate, and separate from background fluorescence. The strong calcein fluorescence is less difficult to detect than tetracycline fluorescence when it is close to the edge of the bone. Another possible explanation for the shorter length of the second tetracycline label could be that recently deposited tetracycline is more susceptible to chelation during formalin fixation. This explanation is unlikely for several reasons. First, there is no evidence that five days’ fixation with properly buffered formalin will remove surface calcium to which tetracycline is bound. Buffered formalin is neutral, not acidic. Second, studies comparing labeled surfaces of bone that was either preserved in alcohol or fixed in cold neutral buffered formalin for up to a week show that neutral buffered formalin does not reduce labeled surface length (Garetto, personal communication). Third, if the formalin were to chelate surface calcium during fixation, then the
446
T. C. Sun et al.: Are fluorochrome
label in Group 6 should be significantly shorter than the tetracycline label in this group. This is not the case (Fig. I). Therefore, the idea that the second label is more easily removed by formalin fixation cannot be an explanation for the reduced length of the second tetracycline label. We were not able to detect any differences in MAR. MSIBS, or BF’R when comparing alizarin complexone-labeled animals to animals from the other labeling groups. This conflicts with the results of Harris et al. (1968), who demonstrated a direct inhibition in mineralization of up to 60% by alizarin Red S. They noted reduced crystal formation, inhibition of crystal growth. and disturbances in collagen synthesis or aggregation (Harris et al. 1964). Our results show, however, that alizarin-complexone provides histomorphometric results comparable to those of calcein and xylenol. calcein
Frost. H
M
Tetracycline-based
T‘r.\.\u~ Re&. 3:21 l-237; Fro\t.
H. M.
Merub. Fro\t.
H. M
Boca Raton.
formation
FL: CRC
W.
Hattner.
R. S.,
tetracychne Norman,
bone-forming
of the labehng
“escape error ”
Techniques and interpretation
E. The rn viva inhibition Jr. Surg. 46A:493-508; 1964
U.: Radin,
1.: Hamblen,
D. L.: Haywood,
C/in. Orthop. RF/. Res. 61:52%60;
llnlcki,
L. P.; Hedge,
to the detectability Schaefer,
H. C.
The dose-response
Cobum.
I. W.:
ot
1968. relationship
of labeled osteons by fluorescence K.;
of bone
E. A. The inhibition
rn wvo.
A. W.:
centers
Press: 1983; pp. 133-142.
Red S. J. Bone
H.; Lavorgna,
ossification
III lam&u
Correction
D. F.: Friberg,
by Alirarin
Culcif.
1980
Bone histomorphumerry
ed.
W. H.: Travis,
Hams,
and “off”
RPI. Rex. 2S:l67~170;
Bone histomorphometry: R. R..
analyala of bone remodeling.
1969.
Resttng seams. “On”
Bone DIS.
Reeker,
Harm,
histological
labels equivalent’?
DeLuca,
ot
microscopy.
H. F.;
Fraser.
D
;
Vitumin D. biomechanica/, chemicul and clinrcul aspects related to calcium metabolism. New York: de Gruyter; Grigoleit.
H. G.; Herrath,
D. V.,
eds.
1977; pp. 377-380. Martin,
Conclusion
R. B. Label
section thickness.
Tetracycline hydrochloride, with its weak fluorescence and lower visibility, causes an underestimation of MS/BS and BFR when used as the final label. When used as the first label followed by other labels, its estimation of MS/BS and BFR is not significantly different than that given by other fluorochrome labels. The use of tetracycline-HCI has no significant effect on MAR. Calcein, alizarin complexone, and xylenol orange provide comparable estimates of MAR, MS/BS, and BFR when used as sequential labels. There was no evidence that tetracycline caused suppression of osteoblasts or increased periods of osteoblastic inactivity.
Milch,
R. A.;
Rall,
escape theory
D. P.; Tobie.
Natl. Cancer Inst. 19:87-91: Ott.
S. M.
Bone
A. M.:
fo‘ormation fails
Villanueva,
H. Classification tetracychne P. J ed. Parfitt,
Meunier.
J. E. Bone localizanon
:
“on-off”
to show
Drezner.
A. R.; Crouch.
Evidence
M.
K.:
M.
M.:
Paris: Armour
Glorieux.
F. H.;
A
M..
1.. Villanueva,
A. R.;
length between demethylchlortetracycline the interpretation
of mineralization. Montagu; Kanis,
Malluche,
H.:
nomenclaNomencla-
1987.
Shih,
M.
S. Difference
and oxytetracycline:
of bone histomorphometric
and
Meunter.
1977; pp. 229-310.
I. A.;
Histomorphometry
J Bone Min. Res 2:595%610:
Foldes,
C. H. E.. Duncan.
use of cell morphology
R. R. Bone histomorphometry
ture, symbols and units. Report of the ASBMR
Parfitt,
in postmenopausal
1989.
Mathews,
for intermittency
P. I.; Ott. S. M.; Reeker,
ture Committee.
J
of the tetracyclines.
pattern
l):S415;
of osteoid seams by combined
labeling.
and
1957.
Bone histomorphometry.
A. M
The effects of resting periods
1989.
J. Bone Min. Res. 4(Suppl.
osteoporosis. Parfitt,
revtsited:
Bone 10:255-264;
data. C&if.
in label
Implications Tissue
for
Int. 48:74-
77: 1991. Schwartz.
Acknowledgments:
This work was completed
with support from NIH grant T35 HL07584-07, “Short-Term Training for Students in Health and NIH grants R55 AR40655 and ROI AR39708. The Professions”; authors appreciate the help of Drs. Charles Turner, Jae Rho, Tomoaki Yoshikawa, Alonza Perry, and Mark Seifert, and of Lisa Pedigo for assistance during various portions of the experiment. The authors also extend their thanks to Dr. R. B. Martin, who read and commented on a prior draft of the manuscript.
M. P.; Reeker, R. R. The label escape error: Determination
bone-forming
surface in histologic
double labels
Metab.
Simmons,
Bone Dis.
D. I.: Teitelbaum.
of rabbit
osteoblasts
sections of bone measured
Rel. Res. 4:237-241;
S. L.: Rosenberg,
by tetracycline.
Metab.
R.; Tress,
becular
R.; Vignery,
bone morphology,
skeletal maturation.
A.
Ewdence
dynamic
of sequential
histomorphometry,
Anar. Rec. 208: 137-145:
1984.
remodelmg
m rat tra-
and changes during
active
1982.
G. D. Altered
metabolic
rhythm
Bone Dis. Rel. Res. 3:51-54:
1981. Simmons.
D
J.: Chang, S-L.;
Russell, J. E.; Grazman,
The effect of protracted tetracycline
C/in
Orthop. Rel. Res. 180:25>259;
B.; Webster,
D.; Oloff,
C
treatment on bone growth and maturation. 1983.
References Baron,
ofthe
by tetraycline
Dare Received: May 12, 1992 Dare Revised: July 20, 1992 Date Accepted: July 23, 1992
