Biochimica et Biophysica Acta, 393 (1975) 531-541
© Elsevier Scientific Publishing Company, Amsterdam - - Printed in The Netherlands BBA 37072 IN VIVO A N D IN V I T R O A G I N G OF C O L L A G E N E X A M I N E D U S I N G AN ISOMETRIC MELTING TECHNIQUE
T. W. MITCHELL and B. J. RIGBY C.S.I.R.O., Division of Textile Physics, 338 Blaxland Road, Ryde, N.S.W. 2112 (Australia)
(Received December 12th, 1974)
SUMMARY 1. In vivo and in vitro aging of tendon from rat tail, kangaroo tail and human wrist tendon was examined by the technique of isometric melting, in physiological saline. 2. For all these collagens, two mechanisms of structure stabilisation can be distinguished in the melting curves. One of these involves co-valent cross-linking as judged by its increasing stability to heat and acid pH, while the second appears to involve only secondary interactions. 3. The time rate of the first process is slow in vivo; rat tendon up to 2 years does not show it, but it is present in 6-year-old human tendon. However, its in vitro rate is markedly dependent upon the free oxygen content of the physiological saline. At an oxygen concentration of 300 nmol/ml, the in vitro aging rate is about 30 times the in vivo rate for rat tail tendon, and about 20 times for both kangaroo tail tendon and human wrist tendon. At a concentration of 60 nmol/ml (which is about the same as normal arteriovenous blood difference) in vitro aging proceeds close to the in vivo rate.
INTRODUCTION The various changes with age in the physical and chemical properties of collagenous tissues (e.g. decrease in solubility) are attributed to alterations in the type of co-valent cross-linking present in newly synthesized collagen. In particular, Bailey and Robins [1] have proposed that the initial cross-links in rat tail tendon and skin are thermally labile aldimine bonds which, as the animal ages, are converted to as yet unknown thermally stable bonds. The total number of bonds is thought to be constant, with just the ratio of thermally stable to labile bonds increasing with age. We have used the technique of isometric melting to examine some aspects of in vivo and in vitro aging in rat tail tendon and human wrist tendon. This technique allows one to determine the rate of aging in terms of the increase in thermal stability of the sample. The particular effects which are measured are (i) the contractile stress developed by the sample as it melts and (ii) the temperature at which the maximum stress is reached. Also, the shape of the temperature vs stress curve gives information on different aging mechanisms which appear with time. An important parameter in
532 this kind of experiment is the rate at which the sample is heated. This can affect the absolute value of the measurements listed above, although the general behaviour is unaffected. An appendix is included showing the dependence of the melting curve characteristics upon rate of heating. MATERIALS AND METHODS Tail tendons from two rats aged, respectively, 50 and 690 days at death were washed in 0.9 ~ NaCl solution. Two batches were then formed from each, one to be maintained at 20 °C and the other at 36 °C in 0 . 9 ~ NaCI saturated with air. The measured dissolved oxygen contents were 310 and 250 nmol/ml at 20 and 36 °C, respectively, which are close to the values for air-saturated solutions. The pH of each solution was 5.9 and, even though the solutions were unbuffered, this value remained steady for weeks: the effect of buffering ions is to be examined separately. As well, four samples of human wrist tendon were examined and some kangaroo tail tendon from an animal of unknown age. The human tendon came from persons aged at death; 1 day, 5.5 years, 24 years and 54 years. The sample is held at each end by passing the tendon through a hole in a stainless steel block into which is then pushed a plastic plug. The top end is attached to a Statham force transducer connected with a chart recorder. The sample is kept wet throughout the mounting procedure and then immersed in the test solution contained in a double-walled cell. This solution is stirred magnetically and heated at a controlled rate by a water bath. Unless otherwise stated, the rate of heating was 1 °C/rain : however, an appendix is attached showing the effect of rate of heating over the range 0.05-13 °C/min. All stress determinations were based upon the diameter of the tendon when immersed in 0.9 ~ NaC1. An experiment is begun by extending the mounted sample until a force is just detectable on the recorder. The temperature is then increased. RESULTS A few general remarks can be made about the stress vs temperature curves before entering into details. Rat tail tendon, when melted no more than 1 or 2 days after removal from the animal (i.e. no in vitro aging), always gave a simple curve, no matter what age the animal was at death. By this, we mean a curve in which the stress rises to a maximum and then turns downward as the temperature is increased. Such behaviour is exemplified by curve A in Fig. 1. The other type of behaviour is shown by curve C in Fig. 1. Here the stress rises as before, then enters a second stage in which the rate of increase becomes slower; eventually the stress begins to increase again with temperature until, finally, a maximum is reached and the stress begins to fall. As will be seen in the appendix, these two types of behaviour are independent of rate of heating, i.e. they are not artefacts due to experimental conditions, but reflect absolute differences between the structures exhibiting them. In samples with high in vivo aging or in young samples with large in vitro aging, the stress developed in the second process is sometimes still increasing at boiling point, i.e. no limit can be observed. This second type of behaviour appears to be exclusive to in vitro aging for rat tail tendon, but it can be seen in wrist tendon from a human aged 5.5 years. In
533 C _ x 10 6
1G IZ
B
~8 uJ
~ 6 /. 2 0
60
713
80
91)
TEMPERATURE ('C}
Fig. 1. Isometric melting curves of rat tail tendon from an animal killed at 50 days. The tendon was in vitro aged in 0.9% NaCI at 20 °C containing 310 nmol/ml of oxygen for (A) 115 days, (B) 256 days and (C) 359 days. The behaviour was almost identical to that of rat tail tendon which had been aged for the same periods in saline at 36 °C containing 250 nmol/ml of oxygen. Fig. 2, we h a v e idealised the m e l t i n g b e h a v i o u r a n d defined a n u m b e r o f c u r v e c h a r a c t e r i s t i c s to be u s e d in t h e f o l l o w i n g discussion. Fig. 1 d o e s n o t s h o w t h e m e l t i n g c u r v e for t h e t e n d o n fresh f r o m the 5 0 - d a y - o l d r a t (i.e. n o in v i t r o a g i n g ) since the stress d e v e l o p e d d u r i n g m e l t i n g (0.01 • 106 d y n e s . c m -2) was t o o s m a l l for t h e scale in t h e figure. F o r t e n d o n f r o m the 2 3 - m o n t h - o l d
SM
$2 Sl
TS r~
T2
TM
TEMPERATURE
Fig. 2. Diagram illustrating the main features of the stress temperature curve, when a tendon is melted under isometric conditions. Two kinds of behaviour can be distinguished. In the first, the stress begins to rise at the familiar shrinkage temperature, Ts, and increases linearly with temperature until a temperature just below 7"1is reached. Here there is a decrease in the rate of increase with temperature, until the stress reaches a maximum and begins to decrease as indicated by the dotted line. In the second type of melting, the stress does not drop after T1, but continues to increase at a constant rate until a temperature T2, at which the rate begins to increase again, ultimately reaching a maximum at temperature TM. St, Sz and Su are the stresses corresponding to T~, T2 and TM, respectively.
534 rat, again with no in vitro aging, the stress was 1.4-l0 6 d y n e - c m - 2 : yet, as curve A in Fig. 1 shows, the maximum stress attained for the younger rat tendon has exceeded this value after only 115 days of in vitro aging. It is clear that rat tail tendon ages much more rapidly in vitro than in vivo, as far as the present experiments are concerned. This is shown clearly in Fig. 3, where SM for the collagen of both rats is plotted against absolute time. The in vivo stress produced by melting (approximated by SO does not contribute greatly to the change in SM caused by in vitro aging. The rates of in vivo and in vitro aging, calculated from the above data and Fig. 3, are 0.8.106 dyne. cm -2. y e a r - 1and 30.106 dyne. cm-2. year- 1, respectively. There is in fact, as Fig. 3 shows, a transition region between the steady-state in vivo and in vitro rates of aging. It is most marked in the tendon from the 50-day-old rat. The values quoted above are the steady-state values and it should be remembered that they correspond to a heating rate of 1 °C/min. 30-x 106
/ g
20
~r n
100
2OO
3OO
~00
5OO 6OO DAYS
7O0
B00
900
I000 I~00
Fig. 3. The maximum stress, SM, as defined in Fig. 2 for rat tail tendon plotted against absolute time after birth of rat. 0, 50-day-old rat; ×, 23-month-old rat; K denotes death of animal. ( -- ) in vivo aging, ( --) vastly increased rate of in vitro aging. The greater rate of in vitro aging is due presumably to the increased free oxygen in saline compared with in vivo conditions. The manner in which the curve characteristics defined in Fig. 2 change with time in vitro are shown in Figs 4 and 5. With respect to the temperatures Ts, 7"1, T2 and T~ it is seen (Fig. 4) that the familiar shrinkage temperature Ts and T~ increase slowly in comparison to T2 and TM. In one year, Ts and T1 have risen by about 3 °C, whereas TM has risen by about 27 °C. The transition region between the two types of melting behaviour defined in Fig. 2 and given by (T2 -- T1), also approaches time independence after about 150 days with a value of about 10 °C, but TM goes on increasing. The stresses $1, $2 and SM vary with time as shown in Fig. 5. The interesting point is that $1 increases only slowly when compared with $2 and SM. This behaviour, taken with the information in Fig. 1, is consistent with the notion that there are a t least two different processes taking place during aging, as mentioned earl i er. For tendon taken from the rat killed at 23 months and stored at two temperatures, 20 and 36 °C, there was little difference between the melting curves.
535
90
/ x
700 60
s6
!
i
~F" + - - ' ~ +
2~o z.,~ ~o
T_s
do ,~o 1~
Co ~
_
I
~
,~o L~o 2½0 2rod 2~o 2~o a~o a~o
IN VITRO AGING TIME IN DAYS
Fig. 4. The change with in vitro aging of the temperatures Ts, T1, Tz and TM as defined in Fig. 2.
The sample, from a rat aged 23 months at death, was tail tendon which had been stored in 0.9~ NaCI containing 310 nmole/ml oxygen, at 20 °C. 2~ -x 106 2z
2~
SM
2C 18
~0
S2
8 6 C 2
0 ~ 20 ~0 do ido ~0 I~0 i~0 ~0 260 ~2 2~0 ~0 2~0 3~0 3~0 IN VITRO AGING TIME IN DAYS
Fig. 5. The change with in vitro aging of the stresses $1, $2 and SM defined in Fig. 2. Sample and conditions as in Fig. 4 caption.
Human tendon H u m a n wrist tendon, when melted isometrically, shows the same curve characteristics as does r a t tail tendon. E x c e p t for the 1-day-old sample, which p r o d u c e d a stress o f only 0.3. l06 d y n e . c m -2, the curves for samples o f in vivo age 5.5, 24 a n d 54 years, respectively, are given in Fig. 6. A l s o included in Fig. 6 are melting curves for in vitro aged tendons f r o m 5.5- a n d 54-year-old samples. These were aged for
536
-x 106
20
16
14
/
CI
uJ z 10 w
C
2 0
60
7O
I 80 9O TEMPERATURE (*C)
I 100
Fig. 6. Isometric melting curves of human wrist tendon in 0.9 ~ NaCI. The curves marked A, B and C were for samples which had no in vitro aging, while those marked A 1 and C 1 had 7 and 4 months in vitro aging, respectively, in saline containing 3 ] 0 nmol/ml of oxygen at 20 °C. A, 54 ),cars at death; B, 24 )'ears; C, 5.5 ),cars. Since A and A 1 arc almost identical, it appears that the aging process has been completed by 54 years or less in human tendon.
periods of 4 and 7 months, respectivley, in saline containing 310 nmol/ml of oxygen: they will be discussed later. The l-day-old sample exhibited the first type of melting curve as defined by curve A in Fig. l, i.e. there was no second rise at 78-80 °C as is clearly seen for the older specimens. Only the 5.5-year-old sample shows a maximum stress, which persisted even at boiling point. In spite of the much greater thermal stability of the human wrist tendon, several of the curve characteristics have absolute values in good agreement with rat tail tendon values. Thus $1, which was found to increase only slowly with in vitro aging in rat and to have a value less than 2.106 dyne. cm -2 before the second stage of aging became apparent, ranged from 0.3.106 to 1.7.106 dyne. c m - 2 for human wrist tendon. Since the results for human wrist tendon were for zero in vitro aging, these figures show good agreement. (7"1 -- Ts) has the value 3 °C and (T2 -- T1)---- l l °C, while the corresponding values for rat are 3 and i0 °C. A rough estimate of the rate of in vivo aging using the maximum stress attained for l day and 5.5-year-old samples was 1.2.106 dyne. cm-2. year - 1 which is in reasonable agreement with 0.8.106 dyne. cm -2. year -1 for rat tail tendon. For the 5.5-year-old sample, the in vitro rate of aging was 20. l 0 6 dyne.cm -Eyear-1 which, while not as rapid as for rat tail tendon is much greater than the in vivo rate (see curves C and C t in Fig. 6). The curve A ~ in Fig. 6 refers to tendon from the 54-year-old source which had been aged in vitro for 7 months; its melting behaviour is almost identical with that for the in vivo aged sample (A). Apparently the process was complete by 54 years or less. Table I, summarises data for rat tail tendon, human wrist tendon and kangaroo
537 TABLE I Values of some of the characteristics of an isometric melting curve as defined in Fig. 2 for the tendon of rat and kangaroo tail and human wrist. Also included are rates of aging in terms of the increase in SM. The in vitro aging was carried out in saline containing 310 nmol/ml of oxygen at 20 °C. There was no significant difference between the in vitro aging rates at 20 and 36 °C. Apparently the increased temperature has compensated for the lower oxygen concentration (approx. 250 nmol/ml). It is assumed (see text) that in vivo aging has taken place at 36 °C in the presence of approx. 90 nmol/ml of oxygen. Sample and aging condition
T I - - Ts Tz -- Tt
TM St (dyne'cm -2 × 106) Aging rate in vivo Aging rate in vitro dyne. cm -2. year- 1 (× 106)
Rat tail tendon
Kangaroo Human wrist tendon tail tendon
In vivo
23 months ?
In vitro
11 months
6 months
3
3
11 90 3.0
13 81 1.6
0.8 30
-20
54 years 0
24 years 0
5.5 years 0
3 3 3 11 11 12 >100 >100 94 1.7 1.6 1.3
9 months foetus 0 2 63 0.3
1.2
20
tail tendon. The age o f the k a n g a r o o at d e a t h was n o t known, b u t it can be seen that for the collagens o f these three animals which have very similar a m i n o acid c o m p o sition, the melting curve characteristics are also very similar. In this w o r k we have n o t t a k e n into account the effect o f collagen t u r n o v e r and growth. Thus in m o s t samples, p a r t i c u l a r l y those o f low in vivo age, a p r o p o r t i o n o f the collagen molecules in the tendons w o u l d have been laid d o w n within the last 50 days before death. However, this p r o p o r t i o n w o u l d be small a n d soluble, and we do n o t expect t h a t it w o u l d be responsible for a significant p a r t o f the results presented. DISCUSSION The changes in the isometric melting p a t t e r n with in vivo a n d in vitro aging are consistent with the general hypothesis that aging involves co-valent cross-linking. If the isometric melting is carried o u t in HC1 at p H 1, the stability increases with aging which is again indicative o f co-valent cross-linking. However, the fact that aging results in increased t h e r m a l stability raises a n u m b e r o f possibilities. Thus, the cross-link density c o u l d increase with aging time, or the density could r e m a i n c o n s t a n t b u t the existing cross-links b e c o m e m o r e t h e r m a l l y stable with time. Either o f the a b o v e situations could occur for one type o f b o n d ; if there were m o r e t h a n one type o f cross-link involved, b o t h processes c o u l d take place i n d e p e n d e n t l y for each type o f bond. Bailey a n d R o b i n s [1], in a s u m m a r y o f recent work, f a v o u r the view that there is one type o f link a n d t h a t its density r e m a i n s constant. However, this cross-link is g r a d u a l l y converted to a t h e r m a l l y stable f o r m with time. T h e initial b o n d s are t h o u g h t to be labile aldimines derived f r o m lysine a n d h y d r o x y l y s i n e residues, which are c o n v e r t e d in an u n k n o w n w a y to an u n k n o w n t h e r m a l l y stable bond. This process
538 is thought to operate to maturation: changes in cross-linking subsequent to maturation are not well understood. When we interpret our results in terms of the cross-linking picture just given the first points which come to mind are (i) the two types of melting behaviour and (ii) the rapidly increased rate of in vitro aging. Under (i), we can conclude that there are two regions of the structure with different thermal stabilities. In terms of our method their stabilities are separated by about 10 °C. The first structure begins to melt around 60 °C and gives rise to an ultimate stress ($1) of about 2.106 dyne-cm -2. The second structure begins to melt around 70 °C and appears to give rise to a stress (SM) which in this work shows no limit for large in vitro aging time. For rat tail tendon, only the first type of melting was observed in vivo. In this work 23 months was the greatest in vivo age, but Boros-Farkas and Everitt [2] used rats up to 36 months and an examination of their results showed that the maximum stress developed was 1.5.106 dyne. cm -2, i.e. no second-type melting behaviour had occurred. The second appeared with in vitro aging. For human wrist tendon the second type had appeared after 5.5 years in vivo aging and could appear earlier. So it would seem that the second type of melting is not exclusive to in vitro aging, but will appear in in vivo aging given time enough. It is reasonable to compare rat tail tendon and human wrist tendon since their amino acid compositions are almost identical. These two melting processes could be due to two different cross-linking mechanisms, or the first (characterised by $1) may not be due to co-valent cross-linking, but to the formation of increasingly ordered crystalline regions. There is low and high angle X-ray evidence for such an occurrence [3, 4]. Referring now to point (ii), the increased rate of in vitro aging compared to in vivo could have an explanation in the increased amount of dissolved oxygen in the saline used for storing, over the amount available in vivo. At 36 °C for example, there were 250 nmol/ml of oxygen in the saline solution, whereas the difference between the free oxygen content of arterial and venous blood according to Stone, W. E. (in ref. 5) is about 93 nmol/ml. Furthermore, other data (Rigby, B. J. and Robinson, M. S., unpublished) show that for rat tail tendon which had been aged in vitro in saline containing the same amount of dissolved oxygen as in normal blood, aging, as measured by the isometric melting technique, proceeded at the same rate as it would in vivo. This assumes, of course, that free oxygen could be involved in the cross-linking reactions. In this connection it has been reported by Tsurufuji and Ogata [6] that the maturation of neutral salt-soluble collagen into insoluble collagen is interrupted by replacing oxygen by nitrogen during incubation. If free oxygen can take part in the in vivo cross-linking of collagen, then the process could go to completion in animals with long life spans. Apparently it does not go to completion in rat tail tendon. As pointed out by Bailey and Robins [1] it is not clear why there is a need for the cross-links to become more stable with time. In fact, they may not need to do so and according to the above considerations the increase in stability may be an unavoidable consequence of the natural availability of free oxygen. We have already pointed out that, over the normal life-span of the rat, its tail tendon does not exhibit the increased thermal stability marked by the appearance of the second type of melting behaviour. But in human wrist tendon, which has a similar amino acid composition and so presumably a similar cross-linking system, the second type of melting behaviour has become apparent by 5.5 years. It would be of interest to examine the melting behaviour of a range of mammals with
539 similar collagens, but of widely differing life-spans. If the melting pattern changed in a similar way with time it would be evidence of a cross-linking process occurring with its own intrinsic rate, perhaps as a function of the available free oxygen. Finally, we make some comment upon the linear region of the stress vs temperature curve defined by (T1 -- Ts) in Fig. 2. This turns out to be 3 °C, not only for rat tail tendon with zero and non-zero in vitro aging, but for human wrist tendon with in vivo aging over the range 1 day to 54 years and for kangaroo tail tendon (Table I). We have earlier referred to this region of the melting curve as signifying the existence of a portion of the tendon in which there is very ordered packing of the tropo-collagen molecules. When these crystallites melt, the molecules melt cooperatively and so the range of melting is similar to the range observed by single molecules in dilute solution, viz. 3 °C. We do not consider that these parts of the structure are involved with co-valent cross-links (or at least not the type associated with appearance of the second kind of melting, defined by (TM -- T2)), since neither (T1 -- Ts) nor the stress involved with this melting, $1, appears to be much affected by in vivo aging or by the increased free oxygen that is available during in vitro aging.
LU
o~56
B0
I 70
I 80
I
9O
TEMPERATURE (*C)
Fig. 7. Isometric melting curves of rat tail tendon in 0.9 ~ NaCI at two rates of heating. The rat was 23 months at death and its tendon had been aged in vitro for 520 days. A, 0.05 °C/rnin; B, 4 °C/min. The arrow on curve A, indicates that the sample broke.
540 These crystallites may in fact be able to pack regularly because thcy are frec of bulky cross-links. The interactions would be predominantly van der Waals, hydrogen bond and electrostatic. APPENDIX
Effect of rate of heating upon the melting curve As might be expected, the stresses $1, $2 and SM and the temperatures Ts, T1, T2 and T~ are dependent upon the rate of heating during the isometric melting experiment. The effect upon some of these curve characteristics was examined over the range 0.05-13 °C/min. In general, (i) an increase in the heating rate increases each of the stresses and temperatures listed above and (ii) the changes with rate of heating become less marked with in vitro aging. These two results are expected because, as the time over which the experiment takes place is reduced, there is less time for stress-relaxation processes to reduce the stress and, as in vitro aging proceeds, the increases in thermal stability of the cross-link network again reduces stress-relaxation. Fig. 7 shows two melting curves at widely differing rates performed with one sample of rat tail tendon and Table II summarises data in which both the rate of
TABLE II IN VITRO AGING IN DAYS Some characteristics of the isometric melting curve of rat tail tendon in 0.9 % NaCl, made at two rates of heating and for two in vitro aging times. The units for the stresses S~, $2 and SM are dynec m - z )< 10 6.
Ts T1 7"2 TM $1 $2 SM
Heating rate at 44 days, (°C/rain)
Heating rate at 520 days, (°C/min)
0.17
4
0.17
4
55.8 58.2 59.8 60.8 0.5 0.6 0.9
62.4 65.5 72.9 76.9 3.8 5.7 8.0
60.4 63.9 76.6 97.0 2.6 8.3 21.2
62.7 65.7 78.7 >96 2.4 9.0 33.0
heating and in vitro aging have been altered. They illustrate the general conclusions made above. Fig. 7 shows clearly the second meltng process and, because of the large difference in rate of heating, indicates that it is rate invariant. Table II again shows the invariance of (T1 -- Ts) ~ 3 °C, this time with rate of heating. This is further support for the idea that the first melting region (which is represented by (7"1 -- Ts)) is due to well ordered crystallites. Finally, the increase in stress with increase in heating rate is dominated by the second melting region, i.e. in the region (T2 -~ TM).
541 REFERENCES 1 Bailey, A. J. and Robins, S. P. (1973) Front. Matrix Biol. 1, 130-156 2 Boros-Farkas, M. and Everitt, A. V. (1967) Gerontologia 13, 37-49 3 Kratky, O., Lauer, M., Ratzenhofer, M. and Sekora, A. (1962) in Collagen, (Ramanathan, N., ed.), pp. 227-232, Interscience, New York 4 Feitelberg, S. and Kaunitz, P. E. (1949) Biochim. Biophys. Acta 3, 155-160 5 Brobeck, J. R., ed. (1973) Physiological Basis of Medicine, 9th edn, Sect. 6, Williams and Wilkins, Baltimore 6 Tsurufuji, S. and Ogata Y. (1965) Biochim. Biophys. Acta 104, 193-199
© Elsevier Scientific Publishing Company, Amsterdam - - Printed in The Netherlands BBA 37072 IN VIVO A N D IN V I T R O A G I N G OF C O L L A G E N E X A M I N E D U S I N G AN ISOMETRIC MELTING TECHNIQUE
T. W. MITCHELL and B. J. RIGBY C.S.I.R.O., Division of Textile Physics, 338 Blaxland Road, Ryde, N.S.W. 2112 (Australia)
(Received December 12th, 1974)
SUMMARY 1. In vivo and in vitro aging of tendon from rat tail, kangaroo tail and human wrist tendon was examined by the technique of isometric melting, in physiological saline. 2. For all these collagens, two mechanisms of structure stabilisation can be distinguished in the melting curves. One of these involves co-valent cross-linking as judged by its increasing stability to heat and acid pH, while the second appears to involve only secondary interactions. 3. The time rate of the first process is slow in vivo; rat tendon up to 2 years does not show it, but it is present in 6-year-old human tendon. However, its in vitro rate is markedly dependent upon the free oxygen content of the physiological saline. At an oxygen concentration of 300 nmol/ml, the in vitro aging rate is about 30 times the in vivo rate for rat tail tendon, and about 20 times for both kangaroo tail tendon and human wrist tendon. At a concentration of 60 nmol/ml (which is about the same as normal arteriovenous blood difference) in vitro aging proceeds close to the in vivo rate.
INTRODUCTION The various changes with age in the physical and chemical properties of collagenous tissues (e.g. decrease in solubility) are attributed to alterations in the type of co-valent cross-linking present in newly synthesized collagen. In particular, Bailey and Robins [1] have proposed that the initial cross-links in rat tail tendon and skin are thermally labile aldimine bonds which, as the animal ages, are converted to as yet unknown thermally stable bonds. The total number of bonds is thought to be constant, with just the ratio of thermally stable to labile bonds increasing with age. We have used the technique of isometric melting to examine some aspects of in vivo and in vitro aging in rat tail tendon and human wrist tendon. This technique allows one to determine the rate of aging in terms of the increase in thermal stability of the sample. The particular effects which are measured are (i) the contractile stress developed by the sample as it melts and (ii) the temperature at which the maximum stress is reached. Also, the shape of the temperature vs stress curve gives information on different aging mechanisms which appear with time. An important parameter in
532 this kind of experiment is the rate at which the sample is heated. This can affect the absolute value of the measurements listed above, although the general behaviour is unaffected. An appendix is included showing the dependence of the melting curve characteristics upon rate of heating. MATERIALS AND METHODS Tail tendons from two rats aged, respectively, 50 and 690 days at death were washed in 0.9 ~ NaCl solution. Two batches were then formed from each, one to be maintained at 20 °C and the other at 36 °C in 0 . 9 ~ NaCI saturated with air. The measured dissolved oxygen contents were 310 and 250 nmol/ml at 20 and 36 °C, respectively, which are close to the values for air-saturated solutions. The pH of each solution was 5.9 and, even though the solutions were unbuffered, this value remained steady for weeks: the effect of buffering ions is to be examined separately. As well, four samples of human wrist tendon were examined and some kangaroo tail tendon from an animal of unknown age. The human tendon came from persons aged at death; 1 day, 5.5 years, 24 years and 54 years. The sample is held at each end by passing the tendon through a hole in a stainless steel block into which is then pushed a plastic plug. The top end is attached to a Statham force transducer connected with a chart recorder. The sample is kept wet throughout the mounting procedure and then immersed in the test solution contained in a double-walled cell. This solution is stirred magnetically and heated at a controlled rate by a water bath. Unless otherwise stated, the rate of heating was 1 °C/rain : however, an appendix is attached showing the effect of rate of heating over the range 0.05-13 °C/min. All stress determinations were based upon the diameter of the tendon when immersed in 0.9 ~ NaC1. An experiment is begun by extending the mounted sample until a force is just detectable on the recorder. The temperature is then increased. RESULTS A few general remarks can be made about the stress vs temperature curves before entering into details. Rat tail tendon, when melted no more than 1 or 2 days after removal from the animal (i.e. no in vitro aging), always gave a simple curve, no matter what age the animal was at death. By this, we mean a curve in which the stress rises to a maximum and then turns downward as the temperature is increased. Such behaviour is exemplified by curve A in Fig. 1. The other type of behaviour is shown by curve C in Fig. 1. Here the stress rises as before, then enters a second stage in which the rate of increase becomes slower; eventually the stress begins to increase again with temperature until, finally, a maximum is reached and the stress begins to fall. As will be seen in the appendix, these two types of behaviour are independent of rate of heating, i.e. they are not artefacts due to experimental conditions, but reflect absolute differences between the structures exhibiting them. In samples with high in vivo aging or in young samples with large in vitro aging, the stress developed in the second process is sometimes still increasing at boiling point, i.e. no limit can be observed. This second type of behaviour appears to be exclusive to in vitro aging for rat tail tendon, but it can be seen in wrist tendon from a human aged 5.5 years. In
533 C _ x 10 6
1G IZ
B
~8 uJ
~ 6 /. 2 0
60
713
80
91)
TEMPERATURE ('C}
Fig. 1. Isometric melting curves of rat tail tendon from an animal killed at 50 days. The tendon was in vitro aged in 0.9% NaCI at 20 °C containing 310 nmol/ml of oxygen for (A) 115 days, (B) 256 days and (C) 359 days. The behaviour was almost identical to that of rat tail tendon which had been aged for the same periods in saline at 36 °C containing 250 nmol/ml of oxygen. Fig. 2, we h a v e idealised the m e l t i n g b e h a v i o u r a n d defined a n u m b e r o f c u r v e c h a r a c t e r i s t i c s to be u s e d in t h e f o l l o w i n g discussion. Fig. 1 d o e s n o t s h o w t h e m e l t i n g c u r v e for t h e t e n d o n fresh f r o m the 5 0 - d a y - o l d r a t (i.e. n o in v i t r o a g i n g ) since the stress d e v e l o p e d d u r i n g m e l t i n g (0.01 • 106 d y n e s . c m -2) was t o o s m a l l for t h e scale in t h e figure. F o r t e n d o n f r o m the 2 3 - m o n t h - o l d
SM
$2 Sl
TS r~
T2
TM
TEMPERATURE
Fig. 2. Diagram illustrating the main features of the stress temperature curve, when a tendon is melted under isometric conditions. Two kinds of behaviour can be distinguished. In the first, the stress begins to rise at the familiar shrinkage temperature, Ts, and increases linearly with temperature until a temperature just below 7"1is reached. Here there is a decrease in the rate of increase with temperature, until the stress reaches a maximum and begins to decrease as indicated by the dotted line. In the second type of melting, the stress does not drop after T1, but continues to increase at a constant rate until a temperature T2, at which the rate begins to increase again, ultimately reaching a maximum at temperature TM. St, Sz and Su are the stresses corresponding to T~, T2 and TM, respectively.
534 rat, again with no in vitro aging, the stress was 1.4-l0 6 d y n e - c m - 2 : yet, as curve A in Fig. 1 shows, the maximum stress attained for the younger rat tendon has exceeded this value after only 115 days of in vitro aging. It is clear that rat tail tendon ages much more rapidly in vitro than in vivo, as far as the present experiments are concerned. This is shown clearly in Fig. 3, where SM for the collagen of both rats is plotted against absolute time. The in vivo stress produced by melting (approximated by SO does not contribute greatly to the change in SM caused by in vitro aging. The rates of in vivo and in vitro aging, calculated from the above data and Fig. 3, are 0.8.106 dyne. cm -2. y e a r - 1and 30.106 dyne. cm-2. year- 1, respectively. There is in fact, as Fig. 3 shows, a transition region between the steady-state in vivo and in vitro rates of aging. It is most marked in the tendon from the 50-day-old rat. The values quoted above are the steady-state values and it should be remembered that they correspond to a heating rate of 1 °C/min. 30-x 106
/ g
20
~r n
100
2OO
3OO
~00
5OO 6OO DAYS
7O0
B00
900
I000 I~00
Fig. 3. The maximum stress, SM, as defined in Fig. 2 for rat tail tendon plotted against absolute time after birth of rat. 0, 50-day-old rat; ×, 23-month-old rat; K denotes death of animal. ( -- ) in vivo aging, ( --) vastly increased rate of in vitro aging. The greater rate of in vitro aging is due presumably to the increased free oxygen in saline compared with in vivo conditions. The manner in which the curve characteristics defined in Fig. 2 change with time in vitro are shown in Figs 4 and 5. With respect to the temperatures Ts, 7"1, T2 and T~ it is seen (Fig. 4) that the familiar shrinkage temperature Ts and T~ increase slowly in comparison to T2 and TM. In one year, Ts and T1 have risen by about 3 °C, whereas TM has risen by about 27 °C. The transition region between the two types of melting behaviour defined in Fig. 2 and given by (T2 -- T1), also approaches time independence after about 150 days with a value of about 10 °C, but TM goes on increasing. The stresses $1, $2 and SM vary with time as shown in Fig. 5. The interesting point is that $1 increases only slowly when compared with $2 and SM. This behaviour, taken with the information in Fig. 1, is consistent with the notion that there are a t least two different processes taking place during aging, as mentioned earl i er. For tendon taken from the rat killed at 23 months and stored at two temperatures, 20 and 36 °C, there was little difference between the melting curves.
535
90
/ x
700 60
s6
!
i
~F" + - - ' ~ +
2~o z.,~ ~o
T_s
do ,~o 1~
Co ~
_
I
~
,~o L~o 2½0 2rod 2~o 2~o a~o a~o
IN VITRO AGING TIME IN DAYS
Fig. 4. The change with in vitro aging of the temperatures Ts, T1, Tz and TM as defined in Fig. 2.
The sample, from a rat aged 23 months at death, was tail tendon which had been stored in 0.9~ NaCI containing 310 nmole/ml oxygen, at 20 °C. 2~ -x 106 2z
2~
SM
2C 18
~0
S2
8 6 C 2
0 ~ 20 ~0 do ido ~0 I~0 i~0 ~0 260 ~2 2~0 ~0 2~0 3~0 3~0 IN VITRO AGING TIME IN DAYS
Fig. 5. The change with in vitro aging of the stresses $1, $2 and SM defined in Fig. 2. Sample and conditions as in Fig. 4 caption.
Human tendon H u m a n wrist tendon, when melted isometrically, shows the same curve characteristics as does r a t tail tendon. E x c e p t for the 1-day-old sample, which p r o d u c e d a stress o f only 0.3. l06 d y n e . c m -2, the curves for samples o f in vivo age 5.5, 24 a n d 54 years, respectively, are given in Fig. 6. A l s o included in Fig. 6 are melting curves for in vitro aged tendons f r o m 5.5- a n d 54-year-old samples. These were aged for
536
-x 106
20
16
14
/
CI
uJ z 10 w
C
2 0
60
7O
I 80 9O TEMPERATURE (*C)
I 100
Fig. 6. Isometric melting curves of human wrist tendon in 0.9 ~ NaCI. The curves marked A, B and C were for samples which had no in vitro aging, while those marked A 1 and C 1 had 7 and 4 months in vitro aging, respectively, in saline containing 3 ] 0 nmol/ml of oxygen at 20 °C. A, 54 ),cars at death; B, 24 )'ears; C, 5.5 ),cars. Since A and A 1 arc almost identical, it appears that the aging process has been completed by 54 years or less in human tendon.
periods of 4 and 7 months, respectivley, in saline containing 310 nmol/ml of oxygen: they will be discussed later. The l-day-old sample exhibited the first type of melting curve as defined by curve A in Fig. l, i.e. there was no second rise at 78-80 °C as is clearly seen for the older specimens. Only the 5.5-year-old sample shows a maximum stress, which persisted even at boiling point. In spite of the much greater thermal stability of the human wrist tendon, several of the curve characteristics have absolute values in good agreement with rat tail tendon values. Thus $1, which was found to increase only slowly with in vitro aging in rat and to have a value less than 2.106 dyne. cm -2 before the second stage of aging became apparent, ranged from 0.3.106 to 1.7.106 dyne. c m - 2 for human wrist tendon. Since the results for human wrist tendon were for zero in vitro aging, these figures show good agreement. (7"1 -- Ts) has the value 3 °C and (T2 -- T1)---- l l °C, while the corresponding values for rat are 3 and i0 °C. A rough estimate of the rate of in vivo aging using the maximum stress attained for l day and 5.5-year-old samples was 1.2.106 dyne. cm-2. year - 1 which is in reasonable agreement with 0.8.106 dyne. cm -2. year -1 for rat tail tendon. For the 5.5-year-old sample, the in vitro rate of aging was 20. l 0 6 dyne.cm -Eyear-1 which, while not as rapid as for rat tail tendon is much greater than the in vivo rate (see curves C and C t in Fig. 6). The curve A ~ in Fig. 6 refers to tendon from the 54-year-old source which had been aged in vitro for 7 months; its melting behaviour is almost identical with that for the in vivo aged sample (A). Apparently the process was complete by 54 years or less. Table I, summarises data for rat tail tendon, human wrist tendon and kangaroo
537 TABLE I Values of some of the characteristics of an isometric melting curve as defined in Fig. 2 for the tendon of rat and kangaroo tail and human wrist. Also included are rates of aging in terms of the increase in SM. The in vitro aging was carried out in saline containing 310 nmol/ml of oxygen at 20 °C. There was no significant difference between the in vitro aging rates at 20 and 36 °C. Apparently the increased temperature has compensated for the lower oxygen concentration (approx. 250 nmol/ml). It is assumed (see text) that in vivo aging has taken place at 36 °C in the presence of approx. 90 nmol/ml of oxygen. Sample and aging condition
T I - - Ts Tz -- Tt
TM St (dyne'cm -2 × 106) Aging rate in vivo Aging rate in vitro dyne. cm -2. year- 1 (× 106)
Rat tail tendon
Kangaroo Human wrist tendon tail tendon
In vivo
23 months ?
In vitro
11 months
6 months
3
3
11 90 3.0
13 81 1.6
0.8 30
-20
54 years 0
24 years 0
5.5 years 0
3 3 3 11 11 12 >100 >100 94 1.7 1.6 1.3
9 months foetus 0 2 63 0.3
1.2
20
tail tendon. The age o f the k a n g a r o o at d e a t h was n o t known, b u t it can be seen that for the collagens o f these three animals which have very similar a m i n o acid c o m p o sition, the melting curve characteristics are also very similar. In this w o r k we have n o t t a k e n into account the effect o f collagen t u r n o v e r and growth. Thus in m o s t samples, p a r t i c u l a r l y those o f low in vivo age, a p r o p o r t i o n o f the collagen molecules in the tendons w o u l d have been laid d o w n within the last 50 days before death. However, this p r o p o r t i o n w o u l d be small a n d soluble, and we do n o t expect t h a t it w o u l d be responsible for a significant p a r t o f the results presented. DISCUSSION The changes in the isometric melting p a t t e r n with in vivo a n d in vitro aging are consistent with the general hypothesis that aging involves co-valent cross-linking. If the isometric melting is carried o u t in HC1 at p H 1, the stability increases with aging which is again indicative o f co-valent cross-linking. However, the fact that aging results in increased t h e r m a l stability raises a n u m b e r o f possibilities. Thus, the cross-link density c o u l d increase with aging time, or the density could r e m a i n c o n s t a n t b u t the existing cross-links b e c o m e m o r e t h e r m a l l y stable with time. Either o f the a b o v e situations could occur for one type o f b o n d ; if there were m o r e t h a n one type o f cross-link involved, b o t h processes c o u l d take place i n d e p e n d e n t l y for each type o f bond. Bailey a n d R o b i n s [1], in a s u m m a r y o f recent work, f a v o u r the view that there is one type o f link a n d t h a t its density r e m a i n s constant. However, this cross-link is g r a d u a l l y converted to a t h e r m a l l y stable f o r m with time. T h e initial b o n d s are t h o u g h t to be labile aldimines derived f r o m lysine a n d h y d r o x y l y s i n e residues, which are c o n v e r t e d in an u n k n o w n w a y to an u n k n o w n t h e r m a l l y stable bond. This process
538 is thought to operate to maturation: changes in cross-linking subsequent to maturation are not well understood. When we interpret our results in terms of the cross-linking picture just given the first points which come to mind are (i) the two types of melting behaviour and (ii) the rapidly increased rate of in vitro aging. Under (i), we can conclude that there are two regions of the structure with different thermal stabilities. In terms of our method their stabilities are separated by about 10 °C. The first structure begins to melt around 60 °C and gives rise to an ultimate stress ($1) of about 2.106 dyne-cm -2. The second structure begins to melt around 70 °C and appears to give rise to a stress (SM) which in this work shows no limit for large in vitro aging time. For rat tail tendon, only the first type of melting was observed in vivo. In this work 23 months was the greatest in vivo age, but Boros-Farkas and Everitt [2] used rats up to 36 months and an examination of their results showed that the maximum stress developed was 1.5.106 dyne. cm -2, i.e. no second-type melting behaviour had occurred. The second appeared with in vitro aging. For human wrist tendon the second type had appeared after 5.5 years in vivo aging and could appear earlier. So it would seem that the second type of melting is not exclusive to in vitro aging, but will appear in in vivo aging given time enough. It is reasonable to compare rat tail tendon and human wrist tendon since their amino acid compositions are almost identical. These two melting processes could be due to two different cross-linking mechanisms, or the first (characterised by $1) may not be due to co-valent cross-linking, but to the formation of increasingly ordered crystalline regions. There is low and high angle X-ray evidence for such an occurrence [3, 4]. Referring now to point (ii), the increased rate of in vitro aging compared to in vivo could have an explanation in the increased amount of dissolved oxygen in the saline used for storing, over the amount available in vivo. At 36 °C for example, there were 250 nmol/ml of oxygen in the saline solution, whereas the difference between the free oxygen content of arterial and venous blood according to Stone, W. E. (in ref. 5) is about 93 nmol/ml. Furthermore, other data (Rigby, B. J. and Robinson, M. S., unpublished) show that for rat tail tendon which had been aged in vitro in saline containing the same amount of dissolved oxygen as in normal blood, aging, as measured by the isometric melting technique, proceeded at the same rate as it would in vivo. This assumes, of course, that free oxygen could be involved in the cross-linking reactions. In this connection it has been reported by Tsurufuji and Ogata [6] that the maturation of neutral salt-soluble collagen into insoluble collagen is interrupted by replacing oxygen by nitrogen during incubation. If free oxygen can take part in the in vivo cross-linking of collagen, then the process could go to completion in animals with long life spans. Apparently it does not go to completion in rat tail tendon. As pointed out by Bailey and Robins [1] it is not clear why there is a need for the cross-links to become more stable with time. In fact, they may not need to do so and according to the above considerations the increase in stability may be an unavoidable consequence of the natural availability of free oxygen. We have already pointed out that, over the normal life-span of the rat, its tail tendon does not exhibit the increased thermal stability marked by the appearance of the second type of melting behaviour. But in human wrist tendon, which has a similar amino acid composition and so presumably a similar cross-linking system, the second type of melting behaviour has become apparent by 5.5 years. It would be of interest to examine the melting behaviour of a range of mammals with
539 similar collagens, but of widely differing life-spans. If the melting pattern changed in a similar way with time it would be evidence of a cross-linking process occurring with its own intrinsic rate, perhaps as a function of the available free oxygen. Finally, we make some comment upon the linear region of the stress vs temperature curve defined by (T1 -- Ts) in Fig. 2. This turns out to be 3 °C, not only for rat tail tendon with zero and non-zero in vitro aging, but for human wrist tendon with in vivo aging over the range 1 day to 54 years and for kangaroo tail tendon (Table I). We have earlier referred to this region of the melting curve as signifying the existence of a portion of the tendon in which there is very ordered packing of the tropo-collagen molecules. When these crystallites melt, the molecules melt cooperatively and so the range of melting is similar to the range observed by single molecules in dilute solution, viz. 3 °C. We do not consider that these parts of the structure are involved with co-valent cross-links (or at least not the type associated with appearance of the second kind of melting, defined by (TM -- T2)), since neither (T1 -- Ts) nor the stress involved with this melting, $1, appears to be much affected by in vivo aging or by the increased free oxygen that is available during in vitro aging.
LU
o~56
B0
I 70
I 80
I
9O
TEMPERATURE (*C)
Fig. 7. Isometric melting curves of rat tail tendon in 0.9 ~ NaCI at two rates of heating. The rat was 23 months at death and its tendon had been aged in vitro for 520 days. A, 0.05 °C/rnin; B, 4 °C/min. The arrow on curve A, indicates that the sample broke.
540 These crystallites may in fact be able to pack regularly because thcy are frec of bulky cross-links. The interactions would be predominantly van der Waals, hydrogen bond and electrostatic. APPENDIX
Effect of rate of heating upon the melting curve As might be expected, the stresses $1, $2 and SM and the temperatures Ts, T1, T2 and T~ are dependent upon the rate of heating during the isometric melting experiment. The effect upon some of these curve characteristics was examined over the range 0.05-13 °C/min. In general, (i) an increase in the heating rate increases each of the stresses and temperatures listed above and (ii) the changes with rate of heating become less marked with in vitro aging. These two results are expected because, as the time over which the experiment takes place is reduced, there is less time for stress-relaxation processes to reduce the stress and, as in vitro aging proceeds, the increases in thermal stability of the cross-link network again reduces stress-relaxation. Fig. 7 shows two melting curves at widely differing rates performed with one sample of rat tail tendon and Table II summarises data in which both the rate of
TABLE II IN VITRO AGING IN DAYS Some characteristics of the isometric melting curve of rat tail tendon in 0.9 % NaCl, made at two rates of heating and for two in vitro aging times. The units for the stresses S~, $2 and SM are dynec m - z )< 10 6.
Ts T1 7"2 TM $1 $2 SM
Heating rate at 44 days, (°C/rain)
Heating rate at 520 days, (°C/min)
0.17
4
0.17
4
55.8 58.2 59.8 60.8 0.5 0.6 0.9
62.4 65.5 72.9 76.9 3.8 5.7 8.0
60.4 63.9 76.6 97.0 2.6 8.3 21.2
62.7 65.7 78.7 >96 2.4 9.0 33.0
heating and in vitro aging have been altered. They illustrate the general conclusions made above. Fig. 7 shows clearly the second meltng process and, because of the large difference in rate of heating, indicates that it is rate invariant. Table II again shows the invariance of (T1 -- Ts) ~ 3 °C, this time with rate of heating. This is further support for the idea that the first melting region (which is represented by (7"1 -- Ts)) is due to well ordered crystallites. Finally, the increase in stress with increase in heating rate is dominated by the second melting region, i.e. in the region (T2 -~ TM).
541 REFERENCES 1 Bailey, A. J. and Robins, S. P. (1973) Front. Matrix Biol. 1, 130-156 2 Boros-Farkas, M. and Everitt, A. V. (1967) Gerontologia 13, 37-49 3 Kratky, O., Lauer, M., Ratzenhofer, M. and Sekora, A. (1962) in Collagen, (Ramanathan, N., ed.), pp. 227-232, Interscience, New York 4 Feitelberg, S. and Kaunitz, P. E. (1949) Biochim. Biophys. Acta 3, 155-160 5 Brobeck, J. R., ed. (1973) Physiological Basis of Medicine, 9th edn, Sect. 6, Williams and Wilkins, Baltimore 6 Tsurufuji, S. and Ogata Y. (1965) Biochim. Biophys. Acta 104, 193-199
