~' It;92 Elscvicf 5
geing of the human c ;tructural and biochel Nageena S. Malik ~, Stasia J. Moss -~, N Ne~ essar Ah and Keit Keith M. M / Oxford Research Unit, The Open University, Ox]ord (U.K.) Department of~"Ophthalmolo~,~ Ophthalmolo~,% Unit'ersit~ (Receiver Received 23 April 1 (Revised manuscript received 1'
Key words: X-ray diffraction; Glycation;
ors B,\:. All light~ r c s c ~ c d 0~]2
stroma: langes m a J. F u r t h
~, R i t a S.
~anqdant Sert,ice Eye Bat& ctol (U.K.)
)
t; Collagen
lrOlll the t i l e c~ corneal s t f O lrna l l a give gll lh and low angle X-ray diffraction patterns from fro~ information ab xmolecular spacing of the collagen molecules and an the mean interfibrillar spacing of the c respectively. intensity synchrotron source, from f [~ectively. X-ray data were collected, using a high into human cornea ap[ proximately physiological hydration. The spacings were w, measured as a function ruction of tissue age. Bet~ 90 years there is an increase in the cross-sectional area ar associated with each [] molecule in corneal apF~rox. 3.04 nm 2 to 3.46 nm 2, and an increase in scleral sq wox. 2.65 nm 2 to 3. collagen from appr¢ non cross-linking between bet collagen mol channges may be due to an increase in the extent of non-enzymic mea collagen glycation u using the thiobarbit age range. We have investigated this possibility by measuring andI the subsequent advanced glycation end-products (AGEs) using fluorescencl:e emission. The rt 'e shown an age-related increase in glycation and AGEs AG in both tissues. We haw have also demonstrate have the interfibrillar spacing of corneal collagen with increasing age which ma~y be related to c proteoglycan composition of the interfibriilar matrix.
Introduction
;oes progressive changes with Human collagen undergoes age including a decrease in solubility, elasticity and permeability, an increase in thermal stability, resision, and an accumulation of tance to enzymatic digestion, fluorescent and yellow pig ments [1]. The precise chemmsformations rmations are unknown, ical changes of these transform aave suggested that the physHowever, in vitro studies have restive cross-linking between ical changes involve progressive the collagen molecules. Collagen cross-links are initially formed from lysine ~s and there are two major and hydroxylysine residues groups of crosslinks: those resulting from nonenzymatitd hydroxylysine residues and cally glycosylated lysine and "tion of lysyl oxidase (oxidathose produced by the action hse-initiated cross-links, the tion) [2]. Unlike lysyl oxidase formation of sugar-derived cross-1
understood. The sug~~ar most commonly vivo cross-linking is g [ueose. is also k Nonenzymatic glycosylation glyc tion and involves the spontaneous addit a protein, such as ccollagen. An initial adduct rearranges to t~ form a more sl with other pr~ product which can react re end-prod~ to form advanced glycation gl ross-links a n d / o r fluorop these may be cross-lin in Fig. 1 [3]. AGEs accumulate on long-lived pr collagen [4] and are uusually quantified b cence (excitation 3X 370 nm/emission 4* close to the 360 n m / /460 nm fluorophor isolated and Monnier [1] have recently rec~ the fluorescence of the th reducible cross-li osylated derivatives (57C and their glycosylate cterized enzymically deri~ k, pyridinoline (295 nm a,GEs can be formed fror ~f Amadori products or (~
than the sugars from new functional groups amino groups to form luorescent adducts and vo, particularly in aged lld result in a decrease nks if sugars competed the same lysine and hydroxylsine residues. How, it has now been suggested that lysyl oxidase-ini'.d cross-linking results in collagen which has a 'er rate of turnover hence allowing greater stabiion of the fibrils by A G E cross-linking [6]. It is ible that such cross-linking leads to structural altions in the collagen, which in certain cases may nately result in the failure of the tissue function. 'he regular arrangement of collagen fibrils in the Leal stroma allows a low angle X-ray diffraction ern to be obtained, from which the mean interfiar separation may be calculated [7,8]. Similarly, the dng of collagen molecules within fibrils gives rise to o. I l l gh angle X-ray reflection from which the mean interm, rmolecular spacing can be determined [9]. In the present ent study, X-ray diffraction was used to measure chanlges in intermolecular and interfibrillar spacing in human tan corneal collagen with age and also to show any age-related related change in the intermolecular spacing of human scleral collagen. The 'he relative changes in the glycation of human corneal gen Leal and scleral collagen m were measured ~ured over the same age range. Fluorescen,ice measurements were also carried out in order to lookk for an age-related change in the level of A G E s in both corneal and scleral collagen. Materials and Methods
Corneal supply The cultured human cornea~ )rneas and sclera were obtained from the Corneal Tr~ ransplant Service Eye Bank, U.K. The corneoscleral discs cs were stored at 34°C in an
incul ME~ calf bon~ myci
lium comprised essentia le's salts and H e p e s buff ~M i_-glutamine, 24 m M :s ml-1 penicillin, 0.1 mg t~g m l - ] amphotericin B
X-ra T expr
n (H) of the corneal str~
ornea - dry weight of cornea
%
H=
weight of cornea
It is t structural parameters s~ brilh aolecular spacings are st dent hydration. Accordingly, mea', ~ere made, the corneal h~ set I the corneal discs into 1, dial} and immersing them in 1.5% poly(ethylene gl glycol) (PEG) cont~ NaCI. Equilibration was w~ carried out for The corneas were then the~ brought to room removed from the dialysis dial tubing, weigh~ in airtight cells. Specirr )eclmens were examinee synchrotron source at Daresbury, U.K., 8.2 and 7.2b. The stati stations were used eit! ously or consecutively, but in both cases obtained on the inteI interfibrillar spacing a molecular spacing of the 1 same specimer patterns were recorded on Caeverker graphic film (Caeverk '.aeverken, Strangnass, S, exposure times of 4 mi min (high angle) or 31 and camera lengths of 11 cm (high angle) angle). Bragg spacings were measured fr films after digitising the t patterns using X L Laser Microder crodensitometer (LKB Gaithersburg, MD) which wh produced linea the diffraction rings. The Bragg spacin verted to intermolecu nolecular or interfibrillal described elsewhere [11,12]. [1
,%
R-NH
2
(p r o t e i n )
+
O=C-H
I
HCOH [
HOCH
I HCOH HCOH
"
R-N=CH I HCOH
~ R-NH-CH I 2 C=O
l
I
HOCH
HOCH
I nCOH
I HCO
HCOH
HCOH
/
/
CML
-~ f l u o r o p h o r e s
protein
I CH20} product )
•
&
ral collagen hydrolysates re required to preparc corneas and sclera. Exoride for 3 days at 4°C wed by extraction in a 4 .15 M sodium chloride /ed more strongly bound ns [13] and eventually, ~stion with the enzyme papain (0.5 mg ml-1 citrate fer, pH 6) at 55°C for 24 h was performed to lbilise the collagen residue. ;ince the human cornea is so small, difficulty was ad in extracting sufficient collagen for analysis. ,ling corneas of similar ages was not very successful, ce acid hydrolysis prior to biochemical analysis was :ied out in an oven at ll0°C for 21 h using 6 M HC1 a final neutralisation step with 10 M N a O H was formed [14]. )barbituric acid (TBA) assay ~. microplate version of the T B A method by Parker ~ L i : 1LI. 1 , [15] was used to measure glycation. Mild hydrolysis of 0.1 ml of sample (corneal or scleral digest) with 0.1 ml of 0.5 M oxalic acid was performed at 124 KPa, °C for 1 h to release 5-hydroxymethyl furfural (5124°C H MIF). F The protein was precipitated by addition of 0.1 ml of 40% trichloroacetic acid and 0.2 ml of superant was removed after centrifugation at 10000 rpm natant for 10 min. The supernatant was transferred to a roNate to which 0.067 ml of 0.05M T B A (to pH oH 6 microplate ~ollowing with N a O H ) was added. Followir win incubation at 37°C for 35 min, the absorbancee at 450 nm was read. Fructose standards were used to produce a standard curve from which the 5-HMF cemcentrations of corneal and
sclei corr corr stan
:ould be read. The absol eral collagen digests were 5-HMF concentrations
Fluc F incc
Fay to measure A G E s : emission was measured spectrophotofluorometet ence emission measurem leral collagen (digested d on a papain blank) and I in 6M HC1 and neutrali )re zeroed on an H C l / mission at 440 nm was 0 nm [16].
896( ried ther gen Na( Flu~ exci Res~
X-n n Fig. 2A shows the increase in the ( spacing) 2 as a function functi of age. The ( spacing) 2 is proporti~ ~ortional to the crossassociated with each molecule which sponds to the area occupied o¢ by a mole~ dration was not set, there The corneal hydration hydrations between H = 1 . 8 and H = however, the area of the intermolecular not vary much above H = 2.5 [17]. The t the intermolecular unit ur cell increases w to a 7% growth is linear it corresponds co~ (birth) 3.28 + 0.07 nm 2 (birtl ~ to 3.52 + 0.12 nn hydration was In Fig. 2B the corneal corr cell again incr( The intermolecular unit u
5-
5-
+
+
+++
+
7~-4= v
+ "~ 3-
~
-~ +
O
+
q]
m
22o
O
E
E
0
o
I
lO
2'o
3'o
1
4o 20 Age (years
Fio 2. Chan~e in (intermnlecul~r ~nacin~
3
30
40
5'0
60
71(
Age ( y e a r s ) reason to suppose the increase as not set. This figure indicates
12-
+
+
lo-
>, -Ba-
+
t~
E
0
E la_ I
~
10
0
2'0
30
4t0
5'0
60
70
80
o
210
100
90
3=0
iO
510
Age (years)
Age (years)
dycation of corneal collagen wit
Fig. by tt
3. Change in (intermolecular spacing)2 of sclera] collagen with Although there is no reason to suppose the increase is linear,
An initial increase is seen in col ~asing age, the curve appears to
line of 'best fit' is plotted to indicate the trend in the data points.
Glycation
I n this case a straight line through the data corres p c rods to an increase of 14%, from 3.04 + 0.15 nm 2
:th) to 3.46 + 0.25 nm 2 (age 90). (birth) The intermolecular unit cell in scleral collagen also The reases as a function of age (Fig. 3), from 2.65 + 0.11 incre nm 2 (birth) to 3.19 + 0.18 nm 2 (age 90), this corresponds to a 20% increase. spacing) 2 of Fi~. 4 shows that the: (interfibrillar snacin~,) Fig. corneal collagen decreasess by about 15% from birth to fibrillar spacing) 2 is proporage 90 years. The (interfibrillar tal area associated with each tional to the cross-sectional fibril (the fibril unit cell).
In the corneal stroma, str the T B A as there is an initial mcrease inc~ in glycation age (Fig. 5). After the tt age of 40, how~ concentration, appears to level off. The q ated from Amadori product p increases fi nmol per mg dry weight weig of tissue over th contrast, scleral colla gen shows a progt in glycation across the th entire age rang~ 4.68 nmol per mg dry dl weight of tissue point on this graph appears to show an tration value of appr( ~rox. 9 n m o l / m g dry
10-
4000-
+
3500-
cE 3000-
+
++
250o(J
tJ a. 2000";" 1500L. 1000-
+
+
5000
I lo
20
3'o
t ,o
io
Age (yet Fig. 4. Change in (interfibrillar spacing)
I
I
I
I
I
0
I
I 20
I 30
I 40
I 50
i 0
Age (years) glycation of scleral collagen wit A linear increase is observed il 3rneal collagen (r, the li
S clea 0 at per
+
+
+ ~- -r-I-
Dis,
+
q
)-
3-
0
i 10
210
310
;0
510
60
710
80
Age (yeors) 7. Change in fluorescence of corneal collagen with age. An ease in the fluorescence is observed, but it is unclear whether the ease is linear and whether it continues beyond the age of about ears, the line of 'best fit' is plotted to indicate the upward trend in the data points.
highh v a l u e at an age o f approx. 80 years may represent an i n d i v i d u a l w i t h i m p a i r e d glucose tolerance, since ing is k n o w n to be associated w i t h i m p a i r e d c o n t r o l ageing of glucose [18]. Fig. 7 shows that there is an increase in fluorescence Fi of corneal collagen with age. However, the level of :s with age and it is therefore scatter in the data increases difficult to ascertain if thee increase in fluorescence is linear.
+ ~
O
+
~" 2 . 5 -
E .2,0-
E E
/./~+
O t'~ 1.5-
x 1"0-
"1" °I-
0.5i O.O 0 0 L,_ Fio
obtained in this study de lration, there is a decrea lr spacing within human also reached by Kanai con ~'lectron microscopical st [191 le fibril unit cell must eJ re& on in stromal thickness sim rrence of 'lakes' - areas 1: inc[ rils into which water ac( of ( contribute to light scatt whi suggested that proteol I atrix play a role in regul ext~ ing between the coil collagen fibrils [20], a size of the fibrils the themselves [21,22]. decrease in the interfibrillar intert spacing fot to the obserx sent study may be linked li~ decrease in the ratio of proteoglycan to and in turn to the res by Scott et al. [21] an, in swelling pressure within the fibrils [23 ing to compare the X-ray X results from end of the age range rang( studied. Even w] each end of the ran ge are equilibrate hydration, there is aa 14% increase in unit cell and a 15% ddecrease in the fibr overall ( age. Both lead to a significant s extrafibrillar space where most of the are situated. Collagen contains both lysyl oxidase The latter are know induced cross-links. ~I have been detected and quantified b~ the extent o f , cence. It is known that t] collagen, increases wit proteins, such as coil own that in vitro g13 has also been showi packing in t expansion of the molecular mol It seemed pos by more than 12% [24]. [~ that the in vivo exr)ansion of molecu may be related to an corneal collagen ma} crease in glycation. the TBA In our early investigations invesl digests cult to perform on corneal , was not effic papain digestion simply sirr ;xtract and secondly produce a soluble exl n n n l i n ~ several1 corneas cnrne of similar ages, and not reproducible, t ;n with concentrated HCl nuch so that eventually o ea (approx. 15 mg) was r for the
)-
3
R C'hnnee
10
20
30
in fluoresce.nee
I 40 5 Age (yeG cff s c l e r a l
+
gen, on the other hand rease across the age rang hs to about 2.6 arb units ( at age 8(1 years (Fig. 8).
+
tion of corneal collagen 0 years of age and then r significant change• A fluorescence of corneal catter in the data makes or not fluorescence inf age. Recent studies on s proteins with age [25] 'e also shown that the greatest changes occur in ly life• Figs• 6 and 8 represent changes in glycation ! fluorescence of scleral collagen. The results were prising in that although glycation and possibly the Lcentration of A G E s level off in the cornea, both Ltinue to increase in scleral collagen and signifiLtly so (Figs• 6 and 8). However, the absolute values glycation and fluorescence are much lower in sclecollagen compared with corneal collagen. This may icate that there are fewer available sites for glycaI in scleral collagen, and may also be related to rages in the rate of turnover of scleral collagen with '~. Generally, the turnover of collagen declines with a g e : [26] however there has been no comparison of the difl!erences in turnover between human cornea and sclera. Eikenberr Eikenberry et al. [9] concluded that some collagen lecules arq are pushed apart by sugar-derived cross-links molecules andt that the observed expansion represents the average', spacing of normal and cross-linked molecules. The current rent investigation suggests further that this average cin~ increases with age a~e as shown by bv Figs. Fi~s_ 2a, 2a. 2b and spacing 3. Although many investig~ations have shown that there in glycation of lens proteins are age-related changes in :n reported that the extent of [27-29] it has recently been glycation of some proteins including lens proteins, does not increase with age [3]. The discrepancy results from differences in assay procedures :edures and sample populalso subject to interference by tions. The T B A assay is also 6-carbon sugars [30] with acid-labile glycosidic bonds. The major source of theese sugars are the corneal he prolonged extraction in 4 proteoglycans, however, the M guanidine hydrochloridee should overcome this prob~teoglycans [13]. Incomplete lem by removing the proteo~ removal of proteoglycans or other contaminating substances results in an interf~ rference with the T B A assay of less than 10% since boroh~ orohydride reduction experi~ore than 90% of the colour meats have shown that more generated in the T B A assa, ;say is derived from non-enzymically glycosylated colla lagen [31] (N. Ahmed, N.S. Malik and A.J. Furth, un~published data). Under our T IB~ A A accax~ r o g . l t g e ~ n h o conditions therefore, the T taken to indicate an age-related product• Our results show an inc content of Amadori product up t
ugh the T B A assay does )f glycation it was suitable nce we required informat ycation at different ages. 1. [3] have proposed thai ng-lived proteins proba[ steady-state with respect lea correlates well with th gation but does not ac However, Patrick et al in the absence of chan ~es in glycation of protein ~e-related alterations in turnover, this may well
coll lute inve vari t glyc equ glu~ COI'I
scle sug
app abl~ con scle
• have suggested that n znerating free radicals aJ nt for the protein dama are exposed to sugar dence to suggest that th
autq des whe
the: occurs in vivo. Although the corneas corm used are preset human corneas' it is difficult to identi eases' of the cornea. For instance, co~ can result in a reduction reduct of oxygen to th in turn affects colla gen and can giv~ ageing effect. It is therefore t not surp graphs show a scatter of points, but in insured as far as po )ossible that any db were not included.
Acknowledgements The authors thank the Corneal Trm ~plying the corneas use Eye Bank for supplyi We are grateful for tthe assistance of D the Daresbury Synch mchrotron and to Nig for his help in the pro )roduction of graphs. to thank the Royal National N; Institute fo] Medical Research Co Council, T F C Frost C and the Wellcome Trust Tr for financial sL
References 1 Sell, D.R. and Monnier , V.M. (1989) Connectfi
19 77-92. 2 Ricard-Blum, S. and V Ville, G. (1989), Int. J. E 1189. 3 Patrick, J.S., Thorpe, S S.R. and Baynes, J.W. (
45, B18-23. a
/~Tichimntn
q
~ ~1 (1990) (1 ett al.
Collagen-glycation factor of arteriosclerosis, pp. ~ers B.V., Current status of pre c complications (Sakamoto, N., eds.). and Kaplan, N.O. (1982) Met
t. (1990) J. Clin. Invest. 86,
20 E
Woolgar, A.E. (19781 J. Mol.
21 S
n, S.D., Meek, K.M., Elliott, ol. Biol. 160, 593-607. ad, G. and Brodsky, G. (1987)
22 P F 23 b, L 24 "I J 25 S I 26 ~/ t~
.S., Blacik, L.S., Sittig, R.A., 13 fin, H.G. (1975) Exp. Eye Res. ord, C.R and Hughes. F.W. (
1
)90) in Storage of corneas for ransplantatlon tr~asty, I.).L., eo.~ Current l:urrent Ophthalmic upnmalmlc Sur :~urgery, .ondon. ~¢orthington, C.R. and Inouye, H. (1985) Int. J. Biol. Macromol. ', 2-8. (lug, H. and Alexander, L.E. (19741 X-ray diffraction procedures or polycrystalline and amorphous materials, Wiley and Sons, ~ew York. ~/all, R.S. (1990) PhD Thesis, Open University. ;cott, J.E. and Bosworth, T.R. (1990) Biochem. J. 491-497. ~arker, K.M., England, J.D., Da Costa, J., Hess, R . L and Goldtein, D.E. (1981) Clin. Chem. 27, 669-672. (ohn, R.R., Cerami, A. and Monnier, V.M. (19841 Diabetes 33, ~7-59. deek, K.M., Fullwood, N.J., Cooke, P.H., Elliott, G.F., Maurice, ).M., Quantock, A.J., Wall, R.S. and Worthington, C.R. (1991) 3iophys. J. 60, 467-474. -Iarris, M.I., Hadden, W.C., Knowler, W.C. and Bennett, P.H. (1987) Diabetes, 36, 523-534. 19 Kanai. (anai, A. and Kaufman, H.E. (1973) Ann. Ophthalmol. 5, 285292.
27 (_ ( 28 ( E 29 \ 3O S
Flint, M.H., Gillard, G.C. and • 1.
(19691 The Eye Davson, [t., ed 9-60(I. gad, G., Brodsky, B. and Eiker 3, 495-5(15. id Abraham• E.C. (1987) Ophth Garlick, P.J. and Millward, 15 mallian tissues and in the who Biomedical Press. ~hylack, LT., Jr., Tung, W.H. 2hem. 256, 5176-5180. Mazer, J.S., Chylack, L.T., Jr., ~4) J. Clin. Invest. 74, 1742-174 landesz-Vigo, J. and Cabezasaol. 66, 220-222. and Wieland, O.H. (1981) J. 1-87,
31 F ;win, I.G., Hopton, M., Frappel Turner, G. (19851, Diabetic Diab Medicine 2, 367-3~ 32 Hunt, J.g., Dean, R.T. and Wolff, S.P. (1988 205 212.
geing of the human c ;tructural and biochel Nageena S. Malik ~, Stasia J. Moss -~, N Ne~ essar Ah and Keit Keith M. M / Oxford Research Unit, The Open University, Ox]ord (U.K.) Department of~"Ophthalmolo~,~ Ophthalmolo~,% Unit'ersit~ (Receiver Received 23 April 1 (Revised manuscript received 1'
Key words: X-ray diffraction; Glycation;
ors B,\:. All light~ r c s c ~ c d 0~]2
stroma: langes m a J. F u r t h
~, R i t a S.
~anqdant Sert,ice Eye Bat& ctol (U.K.)
)
t; Collagen
lrOlll the t i l e c~ corneal s t f O lrna l l a give gll lh and low angle X-ray diffraction patterns from fro~ information ab xmolecular spacing of the collagen molecules and an the mean interfibrillar spacing of the c respectively. intensity synchrotron source, from f [~ectively. X-ray data were collected, using a high into human cornea ap[ proximately physiological hydration. The spacings were w, measured as a function ruction of tissue age. Bet~ 90 years there is an increase in the cross-sectional area ar associated with each [] molecule in corneal apF~rox. 3.04 nm 2 to 3.46 nm 2, and an increase in scleral sq wox. 2.65 nm 2 to 3. collagen from appr¢ non cross-linking between bet collagen mol channges may be due to an increase in the extent of non-enzymic mea collagen glycation u using the thiobarbit age range. We have investigated this possibility by measuring andI the subsequent advanced glycation end-products (AGEs) using fluorescencl:e emission. The rt 'e shown an age-related increase in glycation and AGEs AG in both tissues. We haw have also demonstrate have the interfibrillar spacing of corneal collagen with increasing age which ma~y be related to c proteoglycan composition of the interfibriilar matrix.
Introduction
;oes progressive changes with Human collagen undergoes age including a decrease in solubility, elasticity and permeability, an increase in thermal stability, resision, and an accumulation of tance to enzymatic digestion, fluorescent and yellow pig ments [1]. The precise chemmsformations rmations are unknown, ical changes of these transform aave suggested that the physHowever, in vitro studies have restive cross-linking between ical changes involve progressive the collagen molecules. Collagen cross-links are initially formed from lysine ~s and there are two major and hydroxylysine residues groups of crosslinks: those resulting from nonenzymatitd hydroxylysine residues and cally glycosylated lysine and "tion of lysyl oxidase (oxidathose produced by the action hse-initiated cross-links, the tion) [2]. Unlike lysyl oxidase formation of sugar-derived cross-1
understood. The sug~~ar most commonly vivo cross-linking is g [ueose. is also k Nonenzymatic glycosylation glyc tion and involves the spontaneous addit a protein, such as ccollagen. An initial adduct rearranges to t~ form a more sl with other pr~ product which can react re end-prod~ to form advanced glycation gl ross-links a n d / o r fluorop these may be cross-lin in Fig. 1 [3]. AGEs accumulate on long-lived pr collagen [4] and are uusually quantified b cence (excitation 3X 370 nm/emission 4* close to the 360 n m / /460 nm fluorophor isolated and Monnier [1] have recently rec~ the fluorescence of the th reducible cross-li osylated derivatives (57C and their glycosylate cterized enzymically deri~ k, pyridinoline (295 nm a,GEs can be formed fror ~f Amadori products or (~
than the sugars from new functional groups amino groups to form luorescent adducts and vo, particularly in aged lld result in a decrease nks if sugars competed the same lysine and hydroxylsine residues. How, it has now been suggested that lysyl oxidase-ini'.d cross-linking results in collagen which has a 'er rate of turnover hence allowing greater stabiion of the fibrils by A G E cross-linking [6]. It is ible that such cross-linking leads to structural altions in the collagen, which in certain cases may nately result in the failure of the tissue function. 'he regular arrangement of collagen fibrils in the Leal stroma allows a low angle X-ray diffraction ern to be obtained, from which the mean interfiar separation may be calculated [7,8]. Similarly, the dng of collagen molecules within fibrils gives rise to o. I l l gh angle X-ray reflection from which the mean interm, rmolecular spacing can be determined [9]. In the present ent study, X-ray diffraction was used to measure chanlges in intermolecular and interfibrillar spacing in human tan corneal collagen with age and also to show any age-related related change in the intermolecular spacing of human scleral collagen. The 'he relative changes in the glycation of human corneal gen Leal and scleral collagen m were measured ~ured over the same age range. Fluorescen,ice measurements were also carried out in order to lookk for an age-related change in the level of A G E s in both corneal and scleral collagen. Materials and Methods
Corneal supply The cultured human cornea~ )rneas and sclera were obtained from the Corneal Tr~ ransplant Service Eye Bank, U.K. The corneoscleral discs cs were stored at 34°C in an
incul ME~ calf bon~ myci
lium comprised essentia le's salts and H e p e s buff ~M i_-glutamine, 24 m M :s ml-1 penicillin, 0.1 mg t~g m l - ] amphotericin B
X-ra T expr
n (H) of the corneal str~
ornea - dry weight of cornea
%
H=
weight of cornea
It is t structural parameters s~ brilh aolecular spacings are st dent hydration. Accordingly, mea', ~ere made, the corneal h~ set I the corneal discs into 1, dial} and immersing them in 1.5% poly(ethylene gl glycol) (PEG) cont~ NaCI. Equilibration was w~ carried out for The corneas were then the~ brought to room removed from the dialysis dial tubing, weigh~ in airtight cells. Specirr )eclmens were examinee synchrotron source at Daresbury, U.K., 8.2 and 7.2b. The stati stations were used eit! ously or consecutively, but in both cases obtained on the inteI interfibrillar spacing a molecular spacing of the 1 same specimer patterns were recorded on Caeverker graphic film (Caeverk '.aeverken, Strangnass, S, exposure times of 4 mi min (high angle) or 31 and camera lengths of 11 cm (high angle) angle). Bragg spacings were measured fr films after digitising the t patterns using X L Laser Microder crodensitometer (LKB Gaithersburg, MD) which wh produced linea the diffraction rings. The Bragg spacin verted to intermolecu nolecular or interfibrillal described elsewhere [11,12]. [1
,%
R-NH
2
(p r o t e i n )
+
O=C-H
I
HCOH [
HOCH
I HCOH HCOH
"
R-N=CH I HCOH
~ R-NH-CH I 2 C=O
l
I
HOCH
HOCH
I nCOH
I HCO
HCOH
HCOH
/
/
CML
-~ f l u o r o p h o r e s
protein
I CH20} product )
•
&
ral collagen hydrolysates re required to preparc corneas and sclera. Exoride for 3 days at 4°C wed by extraction in a 4 .15 M sodium chloride /ed more strongly bound ns [13] and eventually, ~stion with the enzyme papain (0.5 mg ml-1 citrate fer, pH 6) at 55°C for 24 h was performed to lbilise the collagen residue. ;ince the human cornea is so small, difficulty was ad in extracting sufficient collagen for analysis. ,ling corneas of similar ages was not very successful, ce acid hydrolysis prior to biochemical analysis was :ied out in an oven at ll0°C for 21 h using 6 M HC1 a final neutralisation step with 10 M N a O H was formed [14]. )barbituric acid (TBA) assay ~. microplate version of the T B A method by Parker ~ L i : 1LI. 1 , [15] was used to measure glycation. Mild hydrolysis of 0.1 ml of sample (corneal or scleral digest) with 0.1 ml of 0.5 M oxalic acid was performed at 124 KPa, °C for 1 h to release 5-hydroxymethyl furfural (5124°C H MIF). F The protein was precipitated by addition of 0.1 ml of 40% trichloroacetic acid and 0.2 ml of superant was removed after centrifugation at 10000 rpm natant for 10 min. The supernatant was transferred to a roNate to which 0.067 ml of 0.05M T B A (to pH oH 6 microplate ~ollowing with N a O H ) was added. Followir win incubation at 37°C for 35 min, the absorbancee at 450 nm was read. Fructose standards were used to produce a standard curve from which the 5-HMF cemcentrations of corneal and
sclei corr corr stan
:ould be read. The absol eral collagen digests were 5-HMF concentrations
Fluc F incc
Fay to measure A G E s : emission was measured spectrophotofluorometet ence emission measurem leral collagen (digested d on a papain blank) and I in 6M HC1 and neutrali )re zeroed on an H C l / mission at 440 nm was 0 nm [16].
896( ried ther gen Na( Flu~ exci Res~
X-n n Fig. 2A shows the increase in the ( spacing) 2 as a function functi of age. The ( spacing) 2 is proporti~ ~ortional to the crossassociated with each molecule which sponds to the area occupied o¢ by a mole~ dration was not set, there The corneal hydration hydrations between H = 1 . 8 and H = however, the area of the intermolecular not vary much above H = 2.5 [17]. The t the intermolecular unit ur cell increases w to a 7% growth is linear it corresponds co~ (birth) 3.28 + 0.07 nm 2 (birtl ~ to 3.52 + 0.12 nn hydration was In Fig. 2B the corneal corr cell again incr( The intermolecular unit u
5-
5-
+
+
+++
+
7~-4= v
+ "~ 3-
~
-~ +
O
+
q]
m
22o
O
E
E
0
o
I
lO
2'o
3'o
1
4o 20 Age (years
Fio 2. Chan~e in (intermnlecul~r ~nacin~
3
30
40
5'0
60
71(
Age ( y e a r s ) reason to suppose the increase as not set. This figure indicates
12-
+
+
lo-
>, -Ba-
+
t~
E
0
E la_ I
~
10
0
2'0
30
4t0
5'0
60
70
80
o
210
100
90
3=0
iO
510
Age (years)
Age (years)
dycation of corneal collagen wit
Fig. by tt
3. Change in (intermolecular spacing)2 of sclera] collagen with Although there is no reason to suppose the increase is linear,
An initial increase is seen in col ~asing age, the curve appears to
line of 'best fit' is plotted to indicate the trend in the data points.
Glycation
I n this case a straight line through the data corres p c rods to an increase of 14%, from 3.04 + 0.15 nm 2
:th) to 3.46 + 0.25 nm 2 (age 90). (birth) The intermolecular unit cell in scleral collagen also The reases as a function of age (Fig. 3), from 2.65 + 0.11 incre nm 2 (birth) to 3.19 + 0.18 nm 2 (age 90), this corresponds to a 20% increase. spacing) 2 of Fi~. 4 shows that the: (interfibrillar snacin~,) Fig. corneal collagen decreasess by about 15% from birth to fibrillar spacing) 2 is proporage 90 years. The (interfibrillar tal area associated with each tional to the cross-sectional fibril (the fibril unit cell).
In the corneal stroma, str the T B A as there is an initial mcrease inc~ in glycation age (Fig. 5). After the tt age of 40, how~ concentration, appears to level off. The q ated from Amadori product p increases fi nmol per mg dry weight weig of tissue over th contrast, scleral colla gen shows a progt in glycation across the th entire age rang~ 4.68 nmol per mg dry dl weight of tissue point on this graph appears to show an tration value of appr( ~rox. 9 n m o l / m g dry
10-
4000-
+
3500-
cE 3000-
+
++
250o(J
tJ a. 2000";" 1500L. 1000-
+
+
5000
I lo
20
3'o
t ,o
io
Age (yet Fig. 4. Change in (interfibrillar spacing)
I
I
I
I
I
0
I
I 20
I 30
I 40
I 50
i 0
Age (years) glycation of scleral collagen wit A linear increase is observed il 3rneal collagen (r, the li
S clea 0 at per
+
+
+ ~- -r-I-
Dis,
+
q
)-
3-
0
i 10
210
310
;0
510
60
710
80
Age (yeors) 7. Change in fluorescence of corneal collagen with age. An ease in the fluorescence is observed, but it is unclear whether the ease is linear and whether it continues beyond the age of about ears, the line of 'best fit' is plotted to indicate the upward trend in the data points.
highh v a l u e at an age o f approx. 80 years may represent an i n d i v i d u a l w i t h i m p a i r e d glucose tolerance, since ing is k n o w n to be associated w i t h i m p a i r e d c o n t r o l ageing of glucose [18]. Fig. 7 shows that there is an increase in fluorescence Fi of corneal collagen with age. However, the level of :s with age and it is therefore scatter in the data increases difficult to ascertain if thee increase in fluorescence is linear.
+ ~
O
+
~" 2 . 5 -
E .2,0-
E E
/./~+
O t'~ 1.5-
x 1"0-
"1" °I-
0.5i O.O 0 0 L,_ Fio
obtained in this study de lration, there is a decrea lr spacing within human also reached by Kanai con ~'lectron microscopical st [191 le fibril unit cell must eJ re& on in stromal thickness sim rrence of 'lakes' - areas 1: inc[ rils into which water ac( of ( contribute to light scatt whi suggested that proteol I atrix play a role in regul ext~ ing between the coil collagen fibrils [20], a size of the fibrils the themselves [21,22]. decrease in the interfibrillar intert spacing fot to the obserx sent study may be linked li~ decrease in the ratio of proteoglycan to and in turn to the res by Scott et al. [21] an, in swelling pressure within the fibrils [23 ing to compare the X-ray X results from end of the age range rang( studied. Even w] each end of the ran ge are equilibrate hydration, there is aa 14% increase in unit cell and a 15% ddecrease in the fibr overall ( age. Both lead to a significant s extrafibrillar space where most of the are situated. Collagen contains both lysyl oxidase The latter are know induced cross-links. ~I have been detected and quantified b~ the extent o f , cence. It is known that t] collagen, increases wit proteins, such as coil own that in vitro g13 has also been showi packing in t expansion of the molecular mol It seemed pos by more than 12% [24]. [~ that the in vivo exr)ansion of molecu may be related to an corneal collagen ma} crease in glycation. the TBA In our early investigations invesl digests cult to perform on corneal , was not effic papain digestion simply sirr ;xtract and secondly produce a soluble exl n n n l i n ~ several1 corneas cnrne of similar ages, and not reproducible, t ;n with concentrated HCl nuch so that eventually o ea (approx. 15 mg) was r for the
)-
3
R C'hnnee
10
20
30
in fluoresce.nee
I 40 5 Age (yeG cff s c l e r a l
+
gen, on the other hand rease across the age rang hs to about 2.6 arb units ( at age 8(1 years (Fig. 8).
+
tion of corneal collagen 0 years of age and then r significant change• A fluorescence of corneal catter in the data makes or not fluorescence inf age. Recent studies on s proteins with age [25] 'e also shown that the greatest changes occur in ly life• Figs• 6 and 8 represent changes in glycation ! fluorescence of scleral collagen. The results were prising in that although glycation and possibly the Lcentration of A G E s level off in the cornea, both Ltinue to increase in scleral collagen and signifiLtly so (Figs• 6 and 8). However, the absolute values glycation and fluorescence are much lower in sclecollagen compared with corneal collagen. This may icate that there are fewer available sites for glycaI in scleral collagen, and may also be related to rages in the rate of turnover of scleral collagen with '~. Generally, the turnover of collagen declines with a g e : [26] however there has been no comparison of the difl!erences in turnover between human cornea and sclera. Eikenberr Eikenberry et al. [9] concluded that some collagen lecules arq are pushed apart by sugar-derived cross-links molecules andt that the observed expansion represents the average', spacing of normal and cross-linked molecules. The current rent investigation suggests further that this average cin~ increases with age a~e as shown by bv Figs. Fi~s_ 2a, 2a. 2b and spacing 3. Although many investig~ations have shown that there in glycation of lens proteins are age-related changes in :n reported that the extent of [27-29] it has recently been glycation of some proteins including lens proteins, does not increase with age [3]. The discrepancy results from differences in assay procedures :edures and sample populalso subject to interference by tions. The T B A assay is also 6-carbon sugars [30] with acid-labile glycosidic bonds. The major source of theese sugars are the corneal he prolonged extraction in 4 proteoglycans, however, the M guanidine hydrochloridee should overcome this prob~teoglycans [13]. Incomplete lem by removing the proteo~ removal of proteoglycans or other contaminating substances results in an interf~ rference with the T B A assay of less than 10% since boroh~ orohydride reduction experi~ore than 90% of the colour meats have shown that more generated in the T B A assa, ;say is derived from non-enzymically glycosylated colla lagen [31] (N. Ahmed, N.S. Malik and A.J. Furth, un~published data). Under our T IB~ A A accax~ r o g . l t g e ~ n h o conditions therefore, the T taken to indicate an age-related product• Our results show an inc content of Amadori product up t
ugh the T B A assay does )f glycation it was suitable nce we required informat ycation at different ages. 1. [3] have proposed thai ng-lived proteins proba[ steady-state with respect lea correlates well with th gation but does not ac However, Patrick et al in the absence of chan ~es in glycation of protein ~e-related alterations in turnover, this may well
coll lute inve vari t glyc equ glu~ COI'I
scle sug
app abl~ con scle
• have suggested that n znerating free radicals aJ nt for the protein dama are exposed to sugar dence to suggest that th
autq des whe
the: occurs in vivo. Although the corneas corm used are preset human corneas' it is difficult to identi eases' of the cornea. For instance, co~ can result in a reduction reduct of oxygen to th in turn affects colla gen and can giv~ ageing effect. It is therefore t not surp graphs show a scatter of points, but in insured as far as po )ossible that any db were not included.
Acknowledgements The authors thank the Corneal Trm ~plying the corneas use Eye Bank for supplyi We are grateful for tthe assistance of D the Daresbury Synch mchrotron and to Nig for his help in the pro )roduction of graphs. to thank the Royal National N; Institute fo] Medical Research Co Council, T F C Frost C and the Wellcome Trust Tr for financial sL
References 1 Sell, D.R. and Monnier , V.M. (1989) Connectfi
19 77-92. 2 Ricard-Blum, S. and V Ville, G. (1989), Int. J. E 1189. 3 Patrick, J.S., Thorpe, S S.R. and Baynes, J.W. (
45, B18-23. a
/~Tichimntn
q
~ ~1 (1990) (1 ett al.
Collagen-glycation factor of arteriosclerosis, pp. ~ers B.V., Current status of pre c complications (Sakamoto, N., eds.). and Kaplan, N.O. (1982) Met
t. (1990) J. Clin. Invest. 86,
20 E
Woolgar, A.E. (19781 J. Mol.
21 S
n, S.D., Meek, K.M., Elliott, ol. Biol. 160, 593-607. ad, G. and Brodsky, G. (1987)
22 P F 23 b, L 24 "I J 25 S I 26 ~/ t~
.S., Blacik, L.S., Sittig, R.A., 13 fin, H.G. (1975) Exp. Eye Res. ord, C.R and Hughes. F.W. (
1
)90) in Storage of corneas for ransplantatlon tr~asty, I.).L., eo.~ Current l:urrent Ophthalmic upnmalmlc Sur :~urgery, .ondon. ~¢orthington, C.R. and Inouye, H. (1985) Int. J. Biol. Macromol. ', 2-8. (lug, H. and Alexander, L.E. (19741 X-ray diffraction procedures or polycrystalline and amorphous materials, Wiley and Sons, ~ew York. ~/all, R.S. (1990) PhD Thesis, Open University. ;cott, J.E. and Bosworth, T.R. (1990) Biochem. J. 491-497. ~arker, K.M., England, J.D., Da Costa, J., Hess, R . L and Goldtein, D.E. (1981) Clin. Chem. 27, 669-672. (ohn, R.R., Cerami, A. and Monnier, V.M. (19841 Diabetes 33, ~7-59. deek, K.M., Fullwood, N.J., Cooke, P.H., Elliott, G.F., Maurice, ).M., Quantock, A.J., Wall, R.S. and Worthington, C.R. (1991) 3iophys. J. 60, 467-474. -Iarris, M.I., Hadden, W.C., Knowler, W.C. and Bennett, P.H. (1987) Diabetes, 36, 523-534. 19 Kanai. (anai, A. and Kaufman, H.E. (1973) Ann. Ophthalmol. 5, 285292.
27 (_ ( 28 ( E 29 \ 3O S
Flint, M.H., Gillard, G.C. and • 1.
(19691 The Eye Davson, [t., ed 9-60(I. gad, G., Brodsky, B. and Eiker 3, 495-5(15. id Abraham• E.C. (1987) Ophth Garlick, P.J. and Millward, 15 mallian tissues and in the who Biomedical Press. ~hylack, LT., Jr., Tung, W.H. 2hem. 256, 5176-5180. Mazer, J.S., Chylack, L.T., Jr., ~4) J. Clin. Invest. 74, 1742-174 landesz-Vigo, J. and Cabezasaol. 66, 220-222. and Wieland, O.H. (1981) J. 1-87,
31 F ;win, I.G., Hopton, M., Frappel Turner, G. (19851, Diabetic Diab Medicine 2, 367-3~ 32 Hunt, J.g., Dean, R.T. and Wolff, S.P. (1988 205 212.
