close

Вход

Забыли?

вход по аккаунту

?

Collagen turnover in the adult femoral mid-shaft Modeled from anthropogenic radiocarbon tracer measurements.

код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 133:808–816 (2007)
Collagen Turnover in the Adult Femoral Mid-Shaft:
Modeled From Anthropogenic Radiocarbon
Tracer Measurements
Robert E.M. Hedges,1* John G. Clement,2 C. David L. Thomas,2 and Tamsin C. O’Connell1,3
1
Research Laboratory for Archaeology, University of Oxford, Oxford OX1 3QY, UK
School of Dental Science, University of Melbourne, Victoria 3010, Australia
3
McDonald Institute for Archaeological Research, University of Cambridge, Cambridge CB2 3ER, UK
2
KEY WORDS
bone; isotope; diet;
14
C
ABSTRACT
We have measured the 14C content of
human femoral mid-shaft collagen to determine the dynamics of adult collagen turnover, using the sudden doubling and subsequent slow relaxation of global atmospheric 14C content due to nuclear bomb testing in the
1960s and 1970s as a tracer. 14C measurements were
made on bone collagen from 67 individuals of both sexes
who died in Australia in 1990-1993, spanning a range of
ages at death from 40 to 97, and these measurements
were compared with values predicted by an age-dependent turnover model. We found that the dataset could constrain models of collagen turnover, with the following
outcomes: 1) Collagen turnover rate of females decreases, on average, from 4%/yr to 3%/yr from 20 to 80
The 14C content of collagen extracted from samples of
mid-shaft femur has been measured using Accelerator
Mass Spectrometry (AMS), in order to estimate the rate
of collagen turnover in human bone. This is possible
because an individual’s collagen carbon isotopes are
derived from their diet, and dietary carbon is derived
from atmospheric carbon dioxide through plant photosynthesis. During the 1960s, the radiocarbon content of
the atmosphere underwent a sharp increase (by a factor
of nearly 2) over a few years as a result of nuclear bomb
testing, followed by a slow decrease. The extent to which
this pulse remains recorded in bone collagen over time is
a measure of the extent to which bone collagen turns over.
Information on bone turnover rates is important in
medical and physiological studies of human bone turnover. It is also useful in forensic dating (i.e., for postnuclear bomb test deaths), and in archaeological science.
Within archaeological science, quantification of turnover rate is important for the calibration of high precision (620–30 years) 14C dates on human bone, as well
as for palaeodietary interpretation of bone stable isotope
measurements. Sometimes the radiocarbon calibration
curve is very steep, so that the time period between collagen formation and the death of the individual can be
amplified into a significant radiocarbon difference. For
example, in dating the Tyrolean Iceman, differences in
the 14C ages of skin and of bone collagen were attributed
to turnover differences (Bonani et al., 1992). This issue
will become increasingly important as 14C measurements
become more precise. When a bone is sampled for isotopic measurements, it is essential to know the age
range of the individual being represented. For example,
if the collagen turnover rate has a value of 4%/yr in
C 2007
V
WILEY-LISS, INC.
years. Male collagen turnover rates average 1.5–3%/yr
over the same period. 2) For both sexes the collagen
turnover rate during adolescent growth is much higher
(5–15%/yr at age 10–15 years), with males having a significantly higher turnover rate than have females, by up
to a factor of 2. 3) Much of the variation in residual
bomb 14C in a person’s bone can be attributed to individual variation in turnover rate, but of no more than about
30% of the average values for adults. 4) Human femoral
bone collagen isotopically reflects an individual’s diet
over a much longer period of time than 10 years, including a substantial portion of collagen synthesised during
adolescence Am J Phys Anthropol 133:808–816, 2007.
C 2007
V
Wiley-Liss, Inc.
adults, then a sample from a 35 year old will contain a
>50% contribution of collagen from the period when that
person was a teenager (the period of greatest growth).
Interpretations of stable isotope values from collagen,
especially in a social context, can be misleading if data
thought to apply to adult behavior in reality apply to
adolescence. This is especially true for the important
topic of individual mobility, e.g., the movement of people
into a new locality at marriage, the Mesolithic populations in Brittany being a case in point (Schulting and
Richards, 2001).
PREVIOUS WORK
Most systematic research on bone collagen turnover
has been done either using labeling methods or via
markers of bone turnover. Labeling studies (using tetracycline, [13C]-proline, or similar) are direct measures of
bone synthesis, but are generally confined to its measurement on bone biopsies (e.g. the iliac crest) (Parfitt
et al., 1997; Babraj et al., 2002). Studies of bone turnover
Grant sponsor: NERC; Grant number: B/S/2001/00358.
*Correspondence to: Robert E.M. Hedges, Research Laboratory for
Archaeology, University of Oxford, Dyson Perrins Building, Oxford
OX1 3QY, UK. E-mail: robert.hedges@rlaha.ox.ac.uk
Received 30 May 2006; accepted 12 January 2007
DOI 10.1002/ajpa.20598
Published online 2 April 2007 in Wiley InterScience
(www.interscience.wiley.com).
COLLAGEN TURNOVER IN HUMAN FEMURS USING BOMB
markers (such as osteocalcin for bone formation, and urinary hydroxyproline, or pyridinium for resorption) primarily reflect alterations in the number of remodeling
sites (the activation frequency) of the whole body bone
pool, and tend to emphasize the faster turning-over material (Robins and New, 1997). Neither histomorphometry nor bone turnover markers are informative about element-specific turnover.
The radiocarbon atmospheric spike generated from nuclear bomb tests has been used by several groups as a
bone turnover label, although the scarce availability of
suitable bone has limited results. The original measurements in the 1970s by Stenhouse and Baxter (1979)
were on only two bones, but indicated a long residence
time of over 10 years. Little has been published since.
The University of Arizona AMS lab measured collagen
14
C of a series of finger bones, which showed a decrease
in turnover rate with age, but the results were never
published in full (Kalin et al., 1995). More recently, the
AMS lab in Vienna has analyzed a limited number of
bones, showing fast turnover of lipids, but generally confirming a slow turnover period of over 10 years for adult
bone collagen (Wild et al., 2000). Work by Geyh (2001) at
Hannover suggests that turnover substantially decreases
after the age of 19, to about 1.5%/yr. Ubelaker et al.
(2006) showed minimal uptake of 14C in two individuals
over the age of 30 years. In Oxford, collagen and carbonate measurements on three individuals suggest that the
turnover rate after the age of 20 is less than 5%/yr
(Hedges, unpublished data), with carbonate turning over
at approximately the same rate as collagen, a fact relevant for palaeodietary reconstruction. But to establish
definitively a turnover rate, a carefully chosen suite of a
large number of bones must be measured and the results
analyzed in the context of at least one suitable model.
Furthermore, the type and age of bone must be specified,
since turnover clearly varies substantially between adolescence and maturity, and also, it is believed, between
different bones.
THIS STUDY’S APPROACH
We used the sudden doubling and subsequent slow
relaxation of global atmospheric 14C content due to nuclear bomb testing in the 1960s and 1970s as a global
tracer. The form of the bomb 14C pulse (the ‘‘bomb
curve’’) for the Southern Hemisphere atmosphere
approximates to a rapid increase followed by a roughly
linear decline to near prebomb values by 2000 (see
Fig. 1) (Hua and Barbetti, 2004). Food of local terrestrial
origin has the same 14C content as the atmosphere of
the time, though delayed by up to a year (Hua and Barbetti, 2004). Newly synthesized collagen is labeled with
the current dietary 14C composition, and an individual’s
bone collagen 14C content at any time therefore represents an integrated record of collagen gain and loss during their lifetime, assuming that the collagen turnover
rate is equivalent to the collagen carbon turnover rate.
Such a record will include bone growth, together with
the osteoclastic loss of collagen and its remodeling
through osteoblastic replacement with fresh collagen,
with the 14C value of the collagen replacement varying
for each year according to the bomb curve.
We studied samples of mid-shaft femoral sections from
67 individuals ranging in age from 40 to 97, who died in
Australia between 1990 and 1993. For each individual,
the dates of birth and death, sex, height, and weight at
14
C
809
Fig. 1. Plot showing the 14C bomb curve in the Southern
Hemisphere. Data from Hua and Barbetti, 2004, and references
therein.
death were known. For each bone sample, we measured
the collagen 14C content and carbon and nitrogen stable
isotope ratios (13C/12C and 15N/14N).
Using all the available information and data, we
looked for correlations between all these variables, to
examine whether patterns might be evident and in order
to test some of the assumptions involved in modeling the
bone collagen turnover. Then we constructed models of
collagen turnover consistent with the 14C and age data,
as explained below.
MATERIALS AND METHODS
Samples were from the Melbourne Femur Collection,
and were taken from Australians who died in otherwise
normal health in 1990–1993. Ethical approval for this
study was given by Victorian Institute of Forensic Medicine’s in-house ethics committee, promulgated under
Australian National Health and Medical Research Council guidelines (approval number EC.10/2002).
From each femur, a cross-section of whole bone was
taken from a similar mid-shaft location (of between 0.02
and 0.85 g in mass), including the entire thickness of the
cortex. Collagen was extracted from each bone sample
for analysis by standard methods. Samples had been
stored and transported in ethanol. The samples were initially rinsed thrice in water for 4 h each, then defatted
by soaking in a mixture of methanol and chloroform (1:2,
v/v) for 8 h, with a minimum of 30 min in an ultrasonic
bath, then changing the solvent and repeating the
extraction. The samples were then demineralized in 0.5
M aqueous hydrochloric acid at 48C for 1–3 days (until
all the mineral had been removed), gelatinized in water
of pH 3 at 758C for 72 h, and the soluble collagen lyophilized. The dried collagen was then analyzed using continuous flow isotope ratio mass spectrometry for carbon
and nitrogen stable isotope composition, and for 14C content by AMS (Hedges et al., 1989; O’Connell et al., 2001;
Ramsey et al., 2004). Carbon and nitrogen stable isotope
ratios are given as d13C and d15N values, relative to
internationally defined scales, VPDB for carbon, and
American Journal of Physical Anthropology—DOI 10.1002/ajpa
810
R.E.M. HEDGES ET AL.
AIR for nitrogen, and are expressed in units of permil (%)
(Hoefs, 1997). Levels of 14C are expressed as the percentage of the atmospheric 14C value at 1950, i.e. ‘‘percent
Modern Carbon’’ (pMC) (Stuiver and Polach, 1977).
RESULTS
The results are reported in Table I, together with
known details for each subject. The 14C content for
males and females against year of birth are shown in
Figure 2.
The main features of the data are that:
There is no correlation between age, weight, or height
and the 14C content of bone collagen.
There is some significant correlation between carbon
and nitrogen stable isotope values and 14C content for
the whole cohort and for males, but not for females.
Regressing 14C against d13C and d15N using ANOVA:
for the whole cohort, R2 ¼ 0.182, p ¼ 0.002; for males,
R2 ¼ 0.283, p ¼ 0.008; for females, R2 ¼ 0.024, p ¼
0.683.
There is a significant correlation of 14C compared to
year of birth, d13C, and d15N for the entire sample and
for males, but not for females. Regressing 14C against
year of birth, d13C, and d15N using ANOVA: for the
whole cohort, R2 ¼ 0.306, p < 0.001; for males, R2 ¼
0.435, p ¼ 0.001; for females, R2 ¼ 0.149, p ¼ 0.166.
Males and females are significantly different in both
carbon and nitrogen stable isotope values. For d13C,
male mean ¼ 17.7%, female mean ¼ 18.1%, mean
difference ¼ 0.4 6 0.2%, independent samples t-test: 65
df, t ¼ 2.364, p ¼ 0.021. For d15N, male mean ¼ þ11.8%,
female mean ¼ þ11.4%, mean difference ¼ 0.3 6 0.1%,
paired samples t-test: 65 df, t ¼ 2.766, p ¼ 0.007.
The individual data points show a high degree of scatter (well beyond that expected from experimental
uncertainty).
THE TURNOVER MODEL
The basis of the model connects an assumed bone collagen ‘‘turnover rate template’’ with the radiocarbon
composition of bone collagen from a person born in year
19xx, where the bone was sampled in year 19yy (being,
in this case, the year of death).
The variation in radiocarbon within the individual
comes from the known atmospheric variation in 14C,
due to nuclear bomb testing, over the latter half of the
twentieth century. Atmospheric CO2 is incorporated
into terrestrial food chains within a year (tracked
through vegetation such as tree rings, and calculated
from large herbivore muscle protein turnover times). A
small additional delay of up to a few months can be
expected from human food chain distribution and storage systems, but atmospheric radiocarbon is incorporated into human diet within 2 years. For modeling
purposes, we used a value of 1 year. There are two possible complications to this. First, since equilibration of
atmospheric CO2 in the oceans is slow, marine organisms are liable to have lower 14C levels than have contemporary terrestrial organisms (the ‘‘marine reservoir
effect’’), therefore people consuming marine diets have
lower 14C levels than those consuming terrestrial diets.
Second, since bomb testing was predominantly a Northern Hemisphere phenomenon, and mixing between the
Northern and Southern parts of the atmosphere is
slow, the pattern of the 14C bomb curve is different
over time in the two hemispheres, with the Northern
Hemisphere curve having a larger and sharper peak in
14
C than that of the Southern Hemisphere, resulting in
a similar difference in the 14C levels in foods in each
hemisphere over the period of the bomb curve (Hua
and Barbetti, 2004).
Because of the complex form of the bomb curve (and
consequently the dietary 14C levels inferred from atmospheric values over time), and also because of the aim to
describe turnover as an age-dependent function, the calculation of the 14C level of bone collagen of each individual is carried out iteratively for each year from birth in
Y0 to death in YD.
Although the data we have measured refer to people
whose age at death is >40 years, the results come from
integrating diet and turnover throughout life, and therefore provide valid information for ages younger than 40
years. However, because of the effects of faster turnover
rate at younger ages, combined with the major part of
the bomb pulse being very early in life for those in our
study who died close to 40 years, the ability of the
results to define turnover rates at younger ages is
reduced. Correspondingly our estimates for the turnover
template parameters are broader for younger ages.
‘‘Turnover rate’’ is defined as the % of collagen lost
and replaced in 1 year from the defined volume for that
particular age, and given as %/yr.
The overall modeling process in detail
The model is derived using the following stepwise process:
(a) Start with the 14C composition during year-of-age 0–
1 (i.e., corresponding to bomb curve for year Y0).
(b) Update for year of age 1–2 years (i.e. Y1)
First, evaluate amount of collagen lost (with 14C composition corresponding to Y0), based on the assumed
value of turnover rate for year-of-age 0–1.
Second, evaluate amount of new collagen deposited.
(This will depend on a ‘‘bone collagen growth model,’’
which is explained later, but in the simplest case
(steady state), is equal to the amount lost.)
Third, evaluate the overall 14C composition for yearof-age 1–2 based on the weighted composition of the
14
C composition of the newly deposited collagen.
(c) Update for the next age year, 2–3 years (year Y2), by
the same process.
(d) Continue until year of death YD is reached, which is
therefore the predicted 14C composition for collagen
in bone of that year.
(e) Compare the modeled curve of values for individuals
born at different times (i.e., for a range of Y0) with
the measured data points.
This approach assumes that any reincorporation of
‘‘removed’’ collagen is negligible—a reasonable assumption given the much slower rate of bone collagen turnover compared with other body protein pools.
The turnover template
An explicit turnover rate (i.e., % collagen loss/gain per
year) is required for each year of life. This is derived
from a modeled ‘‘turnover template,’’ which is defined by
four (decreasing) turnover rate values at four time-
American Journal of Physical Anthropology—DOI 10.1002/ajpa
COLLAGEN TURNOVER IN HUMAN FEMURS USING BOMB
14
811
C
TABLE I. Subject details and results of analyses
VIFM
case no.
Sex
Year of
birth
3490/91
2422/91
4827/90
3487/90
3703/90
3651/90
3570/93
3889/93
15/91
3574/90
3232/90
2610/90
3658/93
2439/91
4303/90
1097/91
4264/90
3649/90
4412/90
3326/90
3237/90
3318/90
3258/90
4560/90
4223/90
3270/90
1506/91
344/91
1662/91
3289/90
2557/91
563/91
1310/91
2156/91
3482/91
3572/90
3004/90
625/91
3715/90
3521/93
3213/90
3265/90
4474/90
3210/90
2617/90
3234/90
3511/90
3253/90
3151/92
3728/90
2957/92
3305/92
815/93
3241/90
3463/90
3420/90
4343/90
3431/90
2977/92
3171/92
2962/92
3011/90
3582/90
3294/90
2700/90
347/91
3501/91
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
M
1894
1895
1900
1902
1903
1904
1905
1905
1907
1908
1909
1912
1912
1913
1915
1915
1916
1919
1922
1926
1928
1930
1933
1935
1938
1941
1941
1945
1946
1948
1948
1950
1950
1951
1951
1895
1897
1902
1903
1908
1910
1915
1918
1919
1920
1921
1922
1929
1931
1932
1932
1932
1933
1934
1936
1940
1941
1942
1942
1942
1942
1943
1945
1948
1950
1950
1951
Age at
death
(yr)
Height
(cm)
Weight
(kg)
Collagen
atomic C/N
ratio
Collagen
d13C (%)
Collagen
d15N (%)
97
96
90
88
87
86
88
88
84
82
81
78
81
78
75
76
74
71
68
64
62
60
57
55
52
49
50
46
45
42
43
41
41
40
40
95
93
89
87
85
80
75
72
71
70
69
68
61
61
58
60
60
60
56
54
50
49
48
50
50
50
47
45
42
40
41
40
153
142
165
155
155
152
151
159
157
155
147
169
142
161
158
152
160
153
147
172
157
160
158
172
161
171
164
164
164
169
168
146
160
174
159
175
172
164
167
177
165
187
177
165
172
178
174
182
173
174
174
174
173
180
161
176
170
178
185
184
156
178
175
177
178
164
179
41
39
53
54
32
69
52
50
39
48
38
58
49
68
49
63
45
46
56
64
69
55
68
60
59
62
57
74
71
72
51
35
52
81
55
58
56
51
55
62
57
82
54
80
62
72
64
98
79
86
53
83
71
70
71
79
69
84
87
95
73
78
66
71
85
51
88
3.2
3.2
3.2
3.3
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.4
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.1
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.3
3.2
3.2
3.2
3.2
3.2
3.2
3.1
3.1
3.2
3.2
3.4
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
18.1
18.0
17.5
19.4
17.5
19.2
19.0
17.8
17.8
17.5
19.5
18.0
18.1
18.8
17.8
18.2
19.3
17.1
18.5
18.4
18.2
17.7
18.1
18.4
17.7
17.9
18.5
18.5
17.0
18.0
17.7
17.9
17.6
17.8
18.1
17.9
17.9
17.4
18.1
17.9
19.1
18.8
17.7
18.3
16.9
17.7
18.2
18.2
17.7
17.1
18.4
18.0
18.0
17.3
17.9
18.2
17.4
17.5
18.0
19.2
17.6
17.1
17.4
14.5
15.4
17.8
17.5
11.7
11.0
11.7
10.8
11.6
10.8
11.0
11.6
11.2
11.3
10.1
11.6
11.0
10.7
12.5
11.7
11.2
11.9
11.1
11.4
10.7
11.8
11.8
10.8
12.2
11.9
11.5
12.2
11.8
11.6
11.6
11.1
12.3
11.2
11.4
11.3
11.8
11.4
11.8
11.6
10.9
11.7
12.5
11.8
11.9
11.9
11.7
11.4
10.9
12.3
11.9
11.4
11.9
12.4
12.2
11.4
13.1
12.2
11.4
11.5
11.5
12.1
12.6
11.5
11.6
11.5
11.4
Collagen
C (pMC)
Error
(61 sd)
OxA date
number
123.76
115.75
133.97
119.40
115.62
119.40
121.69
122.49
115.78
125.94
129.78
123.22
121.57
121.17
116.99
126.54
122.25
120.54
124.92
122.18
132.71
112.19
122.50
117.73
120.52
124.20
120.84
121.70
126.37
126.16
127.41
126.32
131.40
130.15
128.50
116.44
114.09
125.51
113.31
111.17
109.17
117.21
119.83
120.45
118.97
112.24
128.79
123.61
129.72
120.14
111.84
116.21
123.57
115.50
124.52
126.32
119.70
125.60
127.78
108.64
110.00
129.79
114.70
133.30
138.30
143.91
139.70
0.35
0.33
0.34
0.32
0.31
0.33
0.35
0.35
0.32
0.33
0.32
0.32
0.34
0.34
0.35
0.35
0.36
0.36
0.33
0.33
0.34
0.30
0.33
0.34
0.34
0.40
0.34
0.30
0.35
0.32
0.35
0.36
0.40
0.36
0.36
0.31
0.31
0.36
0.31
0.34
0.29
0.35
0.34
0.35
0.43
0.32
0.36
0.32
0.38
0.37
0.33
0.34
0.35
0.36
0.35
0.39
0.30
0.40
0.37
0.33
0.34
0.39
0.30
0.40
0.40
0.38
0.38
OxA-V-2098-7
OxA-V-2096-55
OxA-V-2093-56
OxA-V-2093-48
OxA-V-2093-53
OxA-V-2093-52
OxA-V-2098-17
OxA-V-2098-19
OxA-V-2093-57
OxA-V-2093-51
OxA-V-2093-40
OxA-V-2093-37
OxA-V-2098-18
OxA-V-2096-56
OxA-V-2080-52
OxA-V-2096-53
OxA-V-2080-51
OxA-V-2080-48
OxA-V-2093-55
OxA-V-2093-47
OxA-V-2093-42
OxA-V-2093-46
OxA-V-2093-44
OxA-V-2080-54
OxA-V-2080-50
OxA-V-2123-18
OxA-V-2096-54
OxA-V-2123-21
OxA-V-2080-56
OxA-V-2093-45
OxA-V-2096-57
OxA-V-2080-55
OxA-V-2123-42
OxA-V-2080-57
OxA-V-2098-8
OxA-V-2093-50
OxA-V-2093-38
OxA-V-2096-52
OxA-V-2093-54
OxA-V-2098-16
OxA-V-2093-39
OxA-V-2080-45
OxA-V-2080-53
OxA-V-2094-20
OxA-V-2080-42
OxA-V-2093-41
OxA-V-2093-49
OxA-V-2093-43
OxA-V-2098-11
OxA-V-2080-49
OxA-V-2098-10
OxA-V-2098-14
OxA-V-2098-15
OxA-V-2080-44
OxA-V-2080-47
OxA-V-2080-46
OxA-V-2123-20
OxA-V-2123-41
OxA-V-2098-12
OxA-V-2099-21
OxA-V-2098-13
OxA-V-2080-43
OxA-V-2123-19
OxA-V-2123-17
OxA-V-2123-16
OxA-V-2098-9
OxA-V-2098-6
14
American Journal of Physical Anthropology—DOI 10.1002/ajpa
812
R.E.M. HEDGES ET AL.
remodeling drift, we considered three different types of
growth model.
Fig. 2. Plot of
for all subjects.
14
C content of collagen versus age at death
points, two of which are fixed, and two of which can
vary: at age 0 (age of birth), at age between 10 and 20
(adolescence), at age between 19 and 30 (cessation of
growth), and at age 100 (age by which all subjects are
dead). There are therefore six adjustable parameters (4
turnover rate values and 2 varying time points, adolescence and cessation of growth), with linear interpolation
to evaluate the turnover rate at any given age. Although
more complex than a single analytic function with adjustable parameters, this approach enables a close relationship between changes of turnover rate and biological
growth to be maintained.
The bone collagen growth model
Modeling the average age and 14C composition of surviving collagen in a specific volume of adult mid-shaft femur can be complicated by the fact that the quantity of
collagen in the measured volume of the bone element
may not have remained constant over time. This can
happen for two main reasons: a) growth of the skeleton
(i.e., remodeling drift), and b) loss of bone during osteoporosis. In (a), the absolute position of the juvenile bone
is absent from the volume defined by the adult sample,
and gradually occupies an increasing fraction of the
sampled volume with increasing age, due to remodeling
drift. In (b), there is collagen loss per sampled volume of
bone element over time as a result of bone resorption due
to osteoporosis. However, studies on intracortical porosity
show that bone loss with age does not necessarily result
in a decrease in collagen content per volume of sampled
cortical bone material: intracortical femoral porosity can
but does not necessarily increase with age (Thomas
et al., 2005). Samples from very old individuals can have
comparable porosity to that seen in young adult specimens, which is what we observe in this sample set (Stein
et al., 1999). Therefore, we discount the problem of bone
loss as a result of osteoporosis in the model.
To find out how sensitive the predicted turnover rate
template might be to variation in collagen content due to
(i) The collagen amount in the sampled volume of bone
is constant throughout life.
This is appropriate for the samples used here,
except that it does not include changes in bone geometry during skeletal growth.
(ii) Within one individual, collagen amount per unit volume depends on bone geometry. Here, we account
for remodeling drift, and assume the compact femoral mid-shaft starts, at year-of-age 0–1 as a cylindrical shell, with inner and outer radii of R1A and
R2A, and length LA. R1, R2, and L all grow to a
final value of R1D, R2D and LD, which also defines
the sampled volume. Clearly there is a ‘‘mismatch’’
between the volume of the two cylindrical shells at
birth and at the completion of skeletal growth, and
given that there is negligible recycling of bone collagen, dietary 14C is only represented at all in the
sampled bone, at age X, once R2X > R1D. The quantity of collagen represented can therefore be evaluated for each age according to growth tables—the
effect is similar to a time-shifted version of model (i)
(Garn, 1970).
(iii) Collagen amount is constant per unit volume of
bone throughout life, after the onset of adolescence,
taken as a minimum age of 10 years. This ‘‘jump’’
model is a simplified version of (ii), in which the collagen amount per unit volume is presumed to be
constant (as with (i)) but only from age 10 years,
once bone geometry has undergone the most radical
changes with growth.
The models can be compared for their different predictions on 14C values for age-at-death in YD according to
the overall process (above). In practice there is little difference between the three models, because turnover and
growth due to remodeling drift is rapid for age <15
years, and all the data tested here are for outcomes at
age >40 years. Therefore much of the detailed change
resulting from juvenile skeletal growth has been erased.
When considering the predicted collagen turnover rate
templates for the three models, the templates differ at
younger ages (<15 years), as is to be expected, but show
that there is little difference for ages greater than about
20 years (Fig. 3).
Comparing the turnover model with the data
Given datasets for 1) the 14C content of the diet for
each calendar year, 2) the six variable parameters of the
turnover template (4 turnover rate values and 2 varying
time points), and 3) any change in collagen amount in
the sampled volume for each year of age, it is possible to
generate a curve of predicted 14C as pMC against date of
birth (from 1895 to 1970) for individuals sampled in
1990–1993. This curve is then compared with the plot of
individual data points for individual collagen 14C content
against date of birth (or, the same thing, age at death).
The variable parameters for the turnover template are
adjusted to maximize the fit (i.e., minimize the rootmean-square value of the differences) between the data
points and the curve.
The parameters for the turnover template which is the
approximate best fit to the dataset are shown in Table II
and in Figure 4, as well as those for the templates which
American Journal of Physical Anthropology—DOI 10.1002/ajpa
COLLAGEN TURNOVER IN HUMAN FEMURS USING BOMB
14
813
C
TABLE II. Best fit values for the parameters of the turnover
rate template
Time-point (yr)
Males
Turnover
rate at
this age
(fraction/yr)
0.50 6 0.20
Birth
0a
Adolescence
17 6 1 0.137 6 0.013
Cessation
25 6 2 0.030 6 0.015
of growth
Death definitely 100a 0.015 6 0.007
occurred
a
Females
Turnover
rate at
this age
(fraction/yr)
0a
15 6 1
19 6 1
0.50 6 0.20
0.060 6 0.040
0.041 6 0.010
100a
0.030 6 0.010
Fixed value.
Fig. 3. Least-squares fitted mean of the turnover rate
obtained with the different growth models. See section ‘The
bone collagen growth model’.
correspond to the envelopes of the range of data
(including 90% of the data points). These were derived
using the ‘‘jump’’ model (iii) outlined above, and can be
used to generate a curve of predicted collagen 14C as a
function of age for males and females (Figs. 5 and 6).
While this optimizing procedure cannot define a unique
set of values for the turnover template parameters, it
does define the space for which a reasonable fit between
model and data is possible. This is illustrated in the way
the template parameters change for fits to the upper and
lower extremes (i.e. including 90% of points) of the data
(Figs. 5 and 6). Those points which fall outside these
upper and lower limits are taken as outliers, and are
excluded from the interpretation. Despite the scatter of
data, the ‘‘fit’’ is quite constrained; for example, no-one
after the age of about 25 years can have a turnover rate
faster than about 5%/yr and be consistent with any of
the data points.
Assumptions of the models
The models are based on a number of assumptions,
which may or may not be valid.
1. The 14C content of an individual’s diet only depends
on the year in which it was consumed. There are two
potential weaknesses in this approach; however, in practice, they do not seem to matter, partly because the overall fit of model with data points is statistical.
One way is people moving from the Northern Hemisphere to the Southern Hemisphere, especially during
the 1960–1970s, when the 14C abundance significantly
differed between the two hemispheres. Such a move
would give an individual ‘‘too high’’ a 14C value (compared to the model prediction). In fact two or three data
points do seem to have outlying values that might be
due to this effect.
The second way is that an individual regularly consumed a high proportion of marine fish, as a result of
both the marine 14C reservoir effect and the delay and
reduction in uptake of the bomb pulse into marine surface waters. However, the consumption of marine fish
Fig. 4. Comparison of the modelled turnover rates with
upper and lower boundaries (2s) for both adult females and
adult males.
will also be registered as an increase in the stable isotope ratios of both carbon and nitrogen (i.e., increase in
d13C and d15N) of the consumer protein (Schoeninger et
al., 1983). Although collagen stable isotope values can be
elevated for reasons other than marine protein intake,
our data clearly show that individuals with anomalously
depleted 14C values do not have higher stable carbon
and nitrogen isotope values, and vice versa, so we are
confident that all individuals consumed insufficient marine food to affect their 14C values.
2. All individuals have the same turnover template.
Since the models can only be validated by fitting to a series of data points for a range of dates of birth, the
method fails without this assumption. In fact, the spread
of data points we measured may indicate that this
assumption is only approximately true. However, from
the model, the fit of the predicted curve of 14C vs. date of
birth when the date of sampling is 1990–1993 is a very
sensitive function of the modeled turnover template.
Therefore, although individual turnover variation will
generate noise in the fit between data and model, an
overall statistical fit is still apparent and does define a
fairly narrow range of turnover rates across the sampled
population.
3. The turnover rate is identical throughout the volume
of the sample taken. This can hardly be true during
American Journal of Physical Anthropology—DOI 10.1002/ajpa
814
R.E.M. HEDGES ET AL.
Fig. 5. Plot of the measured 14C content of collagen versus
age at death for males together with the best-fit turnover rate
and its upper and lower boundaries (2s). Outliers were excluded
from the evaluation.
Fig. 6. Plot of the measured 14C content of collagen versus
age at death for females together with the best-fit turnover rate
and its upper and lower boundaries (2s). Outliers were excluded
from the evaluation.
growth (osteoclast activity must be greater in endosteal
bone), and need not be true subsequently, due to the pattern resulting from surface remodeling. However, since
we are taking averages over the volume sampled, and
since growth does not make much difference to the
derived turnover rates (see under [4]), this is not an important issue. In principle, it would be possible to look at
this issue more specifically (e.g., to compare endosteal
and periosteal adult bone values in the same sample).
4. That a valid model for bone growth with implications for collagen net uptake can be applied to the volume
of the sample. At its most obvious, the periosteal part of
the bone sample will have been formed at a different age
than the endosteal part, and neither of the ages will be
at birth. Subsequent history for each location will be
somewhat different. As discussed under model (ii) above,
because these data refer to adult bone, effects of juvenile
growth (<15 years) are all but erased for individuals
aged over 40 years. However, interindividual differences
will influence the turnover model.
30%/yr at age 10–15), although this value is probably
rather dependent on the geometrical growth pattern of
the bone. During adolescence, male collagen turnover
rate is much greater than the female rate, by up to a factor of 2. This difference might be an artifact of different
growth rates affecting the choice of collagen growth
model, so this conclusion should be regarded as tentative. The modeled turnover template cannot be expected
to be so reliable for both males and females for ages <15
years or so, as due to the high turnover rate for ages
<15, far less collagen of that age is represented in samples from older subjects in this study. For younger subjects, where the peak atmospheric 14C content occurred
between 5 and 15 years of age (i.e. aged 30–40 years in
1990), the residual 14C signal in bone collagen in 1990 is
indeed higher, but, being the balance of two larger
effects (higher dietary 14C for longer, but in tissues
which are being turned over at a faster rate), its accuracy is compromised. To evaluate this better, one would
need to measure samples of individuals’ bones taken in
the 1970–1980s.
Entering adulthood, the collagen turnover rate of
females is 4%/yr at age 20 years, and decreases fairly
steadily to 3%/yr at 80 years. The male collagen turnover
rate changes on average from 3%/yr at age 25 years to
1.5%/yr at 80 years. So, in contrast to the pattern modeled during adolescence, males have a lower turnover
rate than females when older.
The results presented here, particularly the conclusion
that males have a lower turnover rate than females
when older but also that they have a higher turnover
rate when younger, are consistent with findings based on
the geometry and porosity distribution of the mid-shaft
femur, which were derived from studying the same
sample set (Feik et al., 1996, 1997; Stein et al., 1999;
Thomas et al., 2000). The turnover rates for mature
adults are also similar to those quoted in the literature
which are derived from activity rates of the bone remodeling unit (BMU), of about 3–4%/yr for cortical bone
(Manolagas, 2000).
DISCUSSION
Given the scatter (see Fig. 2), at first sight it would
seem hopeless to represent the data by a single curve.
However, in practice, the range of the parameters for
turnover templates that generate curves consistent with
the data is rather small. Any template outside the limitations of upper and lower bounds would not generate
data that lies within the data scatter that we observe.
We interpret this scatter, which is well beyond measurement uncertainty, to be due to individual variation in
collagen turnover rate. The magnitude of this variation
is determined by the upper and lower bounds to the template parameters and is included in Table II.
From the fitting of the turnover templates to the data,
we derive values of the turnover rates for males and
females at different ages (Table II and Fig. 4). For both
sexes, the predicted collagen turnover rate during adolescent growth is much higher than in later life (10–
American Journal of Physical Anthropology—DOI 10.1002/ajpa
COLLAGEN TURNOVER IN HUMAN FEMURS USING BOMB
As discussed earlier, the modeled ‘‘turnover template’’
is a linear function defined by four (decreasing) turnover
rate values at four time-points, two fixed (birth and certain death), and two variable (adolescence and cessation
of growth), with linear interpolation to evaluate the
turnover rate at any given age. As well as considering
the turnover rate values generated for different ages, we
can consider the ages that the best fit model produces
for ‘‘adolescence’’ (taken as between 10 and 20 years),
and ‘‘cessation of growth’’ (between 19 and 30 years).
The best fit values for ‘‘adolescence’’ are 17 6 1 years for
males and 15 6 1 years for females, whilst for ‘‘cessation
of growth,’’ they are 25 6 2 years for males, and 19 6 1
years for females. Given that these values are derived
solely from the model, it is interesting that they approximate to physiologically meaningful time points. By age
17 years for males and 15 years for females, the bone
matrix is completely formed but as yet incompletely
mineralized (height velocity drops to <1 cm/yr: Tanner
et al., 1966). Ages 25 years in males and 19 years in
females corresponds to skeletal maturity. Thus, we are
confident that the best fit parameters are not simply an
artifact of the model.
14
C
815
Fig. 7. The fraction of annual diet that is represented in the
mid-shaft femoral collagen of a hypothetical individual at age
35, as calculated using two different turnover rate templates,
‘fast’ and ‘slow’ (as discussed in the text). These ‘fast’ and ‘slow’
turnover curves are also plotted, on the right-hand axis.
Residence time of collagen in femoral bone
Bone growth and turnover rates determine how much
of the collagen synthesized at any particular year of life
remains in the bone at death. Using the results obtained
here, it is possible to generate curves which indicate the
fraction of collagen in the sample that was synthesized
in any given year of an individual’s life. The smoothed
curves are constrained by the four time-points and four
turnover rates defined by the turnover template discussed above. Figure 7 shows annual fractional collagen
synthesis representation for a hypothetical individual
aged 35 years according to two possible turnover templates derived from the model (‘‘fast’’ and ‘‘slow,’’ also
plotted), which are the lower and upper limits of the
best fit modeled turnover template (Table II). The turnover rates of these templates differ by up to a factor of
two, mainly in adulthood. Such differences in turnover
rates of mid-shaft femurs are, according to our results,
quite possible in a normal population, and certainly
could be expected within the same individual for bones
of different type. Figure 7 shows that normal biological
variation in turnover rate can lead to quite large differences in life history representation of collagen, particularly for pre-adult and post-adult stages, and especially
for adults dying between 30 and 40 years. The model
indicates that for a collagen turnover template at the
modeled lower limit, 40–50% of the mid-shaft femoral
collagen of a person dying at age 35 years was synthesized prior to the age of 20 years, while for a ‘‘fast’’ turnover template, the same fraction could be reduced to 20–
25% (calculated by integrating the area under the
curve). The contribution of sub-adult collagen is still
high for individuals in the age range of 30–40 yr. However, for turnover rates of 2%/yr after age 25 (e.g. for
some males), even at age 50 years there is a 40% representation of collagen synthesized at age less than 25
years. This wide envelope of residence time means that
human femoral bone collagen isotopically reflects an
individual’s diet over a much longer period of time than
10 years, including a substantial portion of collagen synthesized during adolescence.
CONCLUSIONS
1.
2.
3.
4.
5.
6.
7.
The dataset from mid-shaft femurs can be used to
derive turnover models which are quite well determined after the age of about 20 years.
The collagen turnover rate of females decreases, on
average, from 4 to 3%/yr from 20 to 80 years.
The male collagen turnover rate changes on average
from 3%/yr to 1.5%/yr from 20 to 80 years.
For both sexes the collagen turnover rate during adolescent growth is much higher (10–30%/yr at age
10–15 years), although this value is rather dependent on the geometrical growth pattern of the bone.
Before the age of 20 years, the male collagen turnover rate is much greater than female turnover rate,
by up to a factor of 2. This finding needs strengthening with more younger data and critical modeling.
Much of the variation in remaining bomb 14C in a
person’s bone can be attributed to individual variation in turnover rate of no more than about 30% of
the average values.
Human femoral bone collagen isotopically reflects an
individual’s diet over a much longer period of time
than 10 years, including a substantial portion of collagen synthesized during adolescence.
ACKNOWLEDGMENTS
We thank the mortuary staff and the staff of the Donor Tissue Bank of the Victorian Institute of Forensic
Medicine for the collection of material. We are grateful
to Sherie Blackwell and Judith McNaughtan, School of
Dental Science, University of Melbourne, for specimen
preparation and sample transfer to Oxford. We are
indebted to the staff of the Oxford Radiocarbon Accelerator Unit for the radiocarbon dating, and to Peter Ditchfield of the Research Laboratory for Archaeology, Oxford,
for help with isotopic analyses. Many thanks to Alan
Boyde for helping the project to get off the ground, to
American Journal of Physical Anthropology—DOI 10.1002/ajpa
816
R.E.M. HEDGES ET AL.
Bob Dewar, McDonald Institute for Archaeological
Research, University of Cambridge, for assistance with
statistical interpretation of the data, and to Nigel Loveridge for helpful discussions and suggestions.
LITERATURE CITED
Babraj J, Cuthbertson DJ, Rickhuss P, Meier-Augenstein W,
Smith K, Bohe J, Wolfe RR, Gibson JNA, Adams C, Rennie
MJ. 2002. Sequential extracts of human bone show differing collagen synthetic rates. Biochem Soc Trans 30:61–
65.
Bonani G, Ivy SD, Niklaus TR, Suter M, Housley RA, Bronk
CR, van Klinken GJ, Hedges REM. 1992. Altersbestimmung
von milligrammproben der Oetztaler Gletscherleiche mit der
Beschleunigermassenspectrometrie-methode (AMS). In: Höpfel
F, Platzer W, Spindler K, editors. Der mann im Eis, Band i.
Innsbruck, Austria: Eigenverlag der Universität Innsbruck.
p 108–116.
Feik SA, Thomas CDL, Clement JG. 1996. Age trends in remodeling of the femoral midshaft differ between the sexes.
J Orthopaedic Res 14:590–597.
Feik SA, Thomas CDL, Clement JG. 1997. Age-related changes
in cortical porosity of the midshaft of the human femur.
J Anat 191:407–416.
Garn SM. 1970. The earlier gain and the later loss of cortical bone,
in nutritional perspective. Springfield, Illinois: CC Thomas.
Geyh MA. 2001. Bomb radiocarbon dating of animal tissue and
hair. Radiocarbon 43:723–730.
Hedges REM, Law IA, Bronk CR, Housley RA. 1989. The
Oxford accelerator mass spectrometry facility: technical developments in routine dating. Archaeometry 31:99–113.
Hoefs J. 1997. Stable isotope geochemistry. Berlin: Springer.
Hua Q, Barbetti M. 2004. Review of tropospheric bomb C-14
data for carbon cycle modeling and age calibration purposes.
Radiocarbon 46:1273–1298.
Kalin RM, Burns K, Jull AJT. 1995. Studies of residence time of
14
C in human bones: an application of AMS to forensic science. In: 209th annual meeting, American Chemical Society,
Anaheim, California.
Manolagas SC. 2000. Birth and death of bone cells: basic regulatory mechanisms and implications for the pathogenesis and
treatment of osteoporosis. Endocr Rev 21:115–137.
O’Connell TC, Healey MA, Hedges REM, Simpson AHW. 2001.
Isotopic comparison of hair, bone and nail: modern analyses.
J Archaeol Sci 28:1247–1255.
Parfitt AM, Han Z, Palnitkar S, Rao D, Shih M, Nelson D.
1997. Effects of ethnicity and age or menopause on osteoblast
function, bone mineralization, and osteoid accumulation in
iliac bone. J Bone Miner Res 12:1864–1873.
Ramsey CB, Higham T, Bowles A, Hedges R. 2004. Improvements to the pretreatment of bone at Oxford. Radiocarbon
46:155–163.
Robins SP, New SA. 1997. Markers of bone turnover in relation
to bone health. Proc Nutr Soc 56:903–914.
Schoeninger MJ, DeNiro MJ, Tauber H. 1983. Stable nitrogen isotope ratios of bone collagen reflect marine and terrestrial components of prehistoric human diet. Science 220:1381–1383.
Schulting RJ, Richards MP. 2001. Dating women and becoming
farmers: new palaeodietary and AMS dating evidence from
the Breton Mesolithic cemeteries of Téviec and Hoëdic.
J Anthropol Archaeol 20:314–344.
Stein MS, Feik SA, Thomas CDL, Clement JG, Wark JD. 1999.
An automated analysis of intracortical porosity in human femoral bone across age. J Bone Miner Res 14:624–632.
Stenhouse MJ, Baxter MS. 1979. The uptake of bomb 14C in
humans. In: Berger R, Suess HE, editors. Radiocarbon dating.
Berkeley: University of California Press. p 324–341.
Stuiver M, Polach H. 1977. Reporting of C-14 data—discussion.
Radiocarbon 19:355–363.
Tanner JM, Whitehouse RH, Takaishi M. 1966. Standards from
birth to maturity for height, weight, height velocity and weight
velocity; British children 1965. Arch Dis Child 41:613–635.
Thomas CDL, Feik SA, Clement JG. 2005. Regional variation of
intracortical porosity in the midshaft of the human femur:
age and sex differences. J Anat 206:115–125.
Thomas CDL, Stein MS, Feik SA, Wark JD, Clement JG. 2000.
Determination of age at death using combined morphology
and histology of the femur. J Anat 196:463–471.
Ubelaker DH, Buchholz BA, Stewart JEB. 2006. Analysis of artificial radiocarbon in different skeletal and dental tissue
types to evaluate date of death. J Forensic Sci 51:484–488.
Wild EM, Arlamovsky KA, Golser R, Kutschera W, Priller A,
Puchegger S, Rom W, Steier P, Vycudilik W. 2000. C-14 dating
with the bomb peak: an application to forensic medicine. Nucl
Instrum Methods Phys Res B 172:944–950.
American Journal of Physical Anthropology—DOI 10.1002/ajpa
Документ
Категория
Без категории
Просмотров
0
Размер файла
246 Кб
Теги
adults, measurements, anthropogenic, turnover, femoral, mid, modeler, radiocarbon, trace, shaft, collagen
1/--страниц
Пожаловаться на содержимое документа