Exp. Eye Res. (2000) 70, 81–88Article No. exer.1999.0754, available online at http :www.idealibrary.com on

Formation of Hydroxyl Radicals in the Human Lens is Related to

the Severity of Nuclear Cataract

BRETT GARNERa*, MICHAEL J. DAVIESb ROGER J. W. TRUSCOTTa

a Australian Cataract Research Foundation, Department of Chemistry, University of Wollongong,

Wollongong NSW 2522, Australia and b EPR Group, The Heart Research Institute, 145 Missenden Rd,

Camperdown, NSW 2050, Australia

(Received Rochester 6 July 1999 and accepted 17 August 1999)

Recent studies have identified specific hydroxylated amino acid oxidation products which stronglysuggest the presence of hydroxyl radical (HO[)-damaged proteins in human cataractous lenses. In thepresent study, the ability of early stage (type II) and advanced (type IV) nuclear cataractous lenshomogenates to catalyse HO[ production in the presence of H

#O#

was investigated using electronparamagnetic resonance (EPR) spectroscopy with the free radical trap, 5,5-dimethyl-1-pyrroline-N-oxide(DMPO). Cataractous lens homogenates incubated with 1 m H

#O#

generated a distinct HO[ signal,which was significantly more intense in the nuclear region of the type IV compared to the type II lenses.The ability of individual lens nuclei and cortices to stimulate HO[ production was positively correlated.The DMPO-HO[ signal was competitively inhibited by ethanol, confirming that the DMPO-HO[ signalwas due to HO[ formation and not DMPO-OOH degradation. The metal ion chelator, diethylene-triaminepentaacetic acid, also inhibited HO[ formation, indicating that lenticular metal ions play a keyrole in HO[ formation. Cataractous lens homogenates also stimulated ascorbyl radical production, furthersuggesting the presence of redox-active metal ions in the tissue. Analysis of lenses for total Fe and Cu(using atomic absorption spectrometry) showed that the more advanced type IV lenses tended to havehigher Fe, but similar Cu, levels compared to the type II lenses. The levels of both metals were lower innon-cataractous lenses. These data support the hypothesis that transition metal-mediated HO[ productionmay play a role in the aetiology of age-related nuclear cataract. # 2000 Academic Press

Key words : cataract ; protein oxidation; hydrogen peroxide ; metal ion; hydroxyl radical ; ascorbylradical ; EPR spectroscopy.

1. Introduction

Cataract is the largest single cause of blindness in the

world and the surgery required to treat this disease is

a major financial burden (Kupfer, Underwood and

Gillen, 1994). Age-related nuclear cataract is charac-

terized by the progressive opacification of the lens and

this is accompanied by an increased coloration and

fluorescence of the major protein components of the

lens, the crystallins (Augusteyn, 1974; Lerman and

Borkman, 1976; Lerman et al., 1976). The color

associated with cataractous lenses ranges from yellow

to brown to almost black in the most severe cases. This

increase in color has formed the basis for a classi-

fication system which ranges from type I to type IV

(Pirie, 1968).

It has also been recognized for some time that the

crystallins of cataractous lenses are more extensively

oxidized (specifically Cys and Met residues) than in

healthy lenses (Dische and Zil, 1951; Garner and

Spector, 1980; Truscott and Augusteyn, 1977a). It

has been postulated that H#O#

generated in the lens

* Address correspondence to: Brett Garner, The OxfordGlycobiology Institute, Department of Biochemistry, University ofOxford, South Parks Rd, Oxford OX1 3QU, U.K. Fax: 44 (0) 1865275 342, E-mail : brett!glycob.ox.au.uk

andor the aqueous humour could be directly re-

sponsible for this oxidation (Spector and Garner,

1981; Truscott and Augusteyn, 1977b). Experimental

models which increase lenticular H#O#concentrations

do mimic some of the changes observed in cataractous

lenses (Cui and Lou, 1993; Spector et al., 1993). In

addition, H#O#

concentrations in some cataractous

lenses were reported to be higher than in normals

(Spector and Garner, 1981). Subsequent studies

suggested that metal ion-catalysed formation of HO[via the Fenton reaction (eqn 1) could contribute to the

protein modifications observed in cataract (reviewed

in Garland, 1990; Spector, 1995).

H#O#­M(n−")+UHO[­OH−­Mn+ (1)

In vitro model systems, utilizing a ‘photo-Fenton’

reagent (Guptasarma et al., 1992) or, more com-

monly, exogenously added transition metal ions and

H#O#(or ascorbate), have shown that HO[ formation

could also mimic some of the modifications detected in

cataractous lens proteins (Fu et al., 1998; Garland,

Zigler and Kinoshita, 1986; Smith, Jiang and

Abraham, 1997; Zigler, Huang and Du, 1989). While

the idea that lenticular production of HO[ contributes

to crystallin modification in cataract has received

attention, it was not until 1998 that the first specific

0014–483500010081­08 $35.000 # 2000 Academic Press

82 B. GARNER ET AL.

chemical evidence for the presence of HO[-damaged

proteins in cataractous lenses was provided (Fu et al.,

1998). These studies showed that even in the earliest

stages of cataract, lens proteins contained o-tyrosine,

m-tyrosine, di-tyrosine, dihydroxyphenylalanine,

valine-hydroxide and leucine-hydroxide. Furthermore,

the levels of these oxidized amino acids increased with

severity of cataract or when calf lens crystallins were

exposed to a Fenton system (Fu et al., 1998). As HO[-

damaged proteins become crosslinked, insoluble and

can become colored (Davies and Dean, 1997), it is

plausible that the HO[ plays a key role in cataracto-

genesis.

While there is evidence that HO[ contributes to age-

related nuclear cataract formation, the origin of

lenticular H#O#is uncertain. Diffusion of H

#O#from the

aqueous humour to the interior of the lens is possible,

however, the epithelial cells which cover the anterior

surface of the lens are richly endowed with antioxidant

defences, including catalase, GSH and GSH peroxidase,

which remove aqueous-derived H#O#(Berman, 1991;

Reddan et al., 1999). Recently it has been suggested

that the aqueous humour H#O#

levels previously

reported using the dichlorophenol-indophenol method

(reviewed in Spector, Ma and Wang, 1998) may be

artificially high due to the presence of ascorbate

(Bleau, Giasson and Brunette, 1998; Spector, Ma and

Wang, 1998). Using either a modified xylenol orange

(FOX1) or the titanium porphyrin complex-based

methods, aqueous humour H#O#

levels appear to be

only ca. 1–5 µ (Bleau et al., 1998; Spector, Ma and

Wang, 1998). However, this should not alter the

conclusions ‘qualitatively ’ where healthy and

cataractous lenses have been compared under parallel

experimental conditions (Spector and Garner, 1981).

Another means of H#O#

production could be via

intralenticular autoxidation of ascorbate, GSH and 3-

hydroxykynurenine (Vazquez et al., 1999), a possible

limitation being the low concentration of oxygen

thought to be in the lens nucleus (Eaton, 1997).

Another key issue concerning the HO[ hypothesis is

that it is not certain where, and in what redox state,

transition metal ions are found in the lens. Some (Cook

and McGahan, 1986; Lakomaa and Eklund, 1978;

Nath, Srivastava and Singh, 1969), but not all

(Swanson and Truesdale, 1971) studies have provided

evidence that Fe and Cu are present at increased

concentrations in cataractous lenses (reviewed in

Cook and McGahan, 1986; Garland, 1990). We are

not aware of any study that has attempted to ascertain

possible differences in either metal ion levels or redox

availability at different stages of cataract development.

Here we used EPR spectroscopy to examine the ability

of type II and type IV cataractous lenses to generate

HO[ when provided with a source of H#O#. We also

measured total Fe and Cu levels in lenses at these two

stages of nuclear cataract. The data indicate that the

more diseased lenses are better catalysts of HO[production, particularly in the nuclear region.

2. Materials and Methods

Materials

Ascorbic acid, diethylenetriaminepentaacetic acid

(DETAPAC), ethylenediaminetetraacetic acid (EDTA),

FeCl$and DMPO (purified before use by treatment with

activated charcoal) was from Sigma (St. Louis, MO,

U.S.A.). HPLC grade ethanol (Unichrom) and H#O#

was from Ajax (Auburn NSW, Australia), Chelex-100

resin from Bio-Rad (Regents Park NSW, Australia) and

HNO$

("99% pure) from Aldrich (Milwaukee, MI,

U.S.A.). Other reagents were of the highest quality

available and obtained through commercial suppliers.

Milli-Q water (purified to 18 MΩ cm−#) was used in the

preparation of all solutions.

Lens Collection

Severely diseased cataractous lenses are not readily

obtainable in developed countries including Australia.

Cataractous lenses were therefore imported from India

after intracapsular extraction to alleviate blindness.

The lenses were stored in 70% (vv) medical-grade

ethanol in H#O and kept on ice during transit. Lenses

were subsequently stored at ®20°C until use.

Cataractous lenses were divided into two groups based

on color intensity according to the method of Pirie

(1968). Since type II and IV lenses could be sorted

unequivocally, samples of these were used in the

comparative analyses below. Data on the age and sex

of the cataract patients were not available. Where

indicated, ‘ fresh’ non-cataractous lenses were

obtained locally from donor eyes used for corneal

grafting. Ethical approval was from the Eastern Sydney

Area Health Service –Research Ethics Committee (Ref.

90057) and the University of Wollongong Human

Ethics Committee (Ref. HE96145).

Lens Sectioning and EPR Spectroscopy

Frozen lenses were cored with a 5 mm I.D.

chromium-plated brass trephine and E1 mm of both

the anterior and posterior surface of the nuclear core

was removed. The resulting core sample was defined

as the nuclear region of the lens. All remaining

material was defined as the cortex. Both the nucleus

and cortex were weighed and made up to 15% (wv)

with sterile phosphate-buffered saline (PBS), pH 7±4,

which had been previously treated with Chelex resin

to remove adventitious transition metals (van Reyk et

al., 1995). All EPR spectroscopic comparisons were

made on the basis of identical concentrations of wet

lens material. It is noteworthy that the water content

(mass%) of human lens nucleus and cortex are very

similar (64 and 69% respectively) according to

previous studies (Huizinga et al., 1989). The wet mass

of nuclear and cortical material was typically E44

and 86 mg respectively (dry mass of E20 and 31 mg

respectively). The samples were homogenized in plastic

HYDROXYL RADICALS AND CATARACT 83

Eppendorf tubes using a custom-made teflon pestle.

Various additions were made (described in the ‘results ’

section and figure legends) after the samples were

homogenized.

EPR spectra were recorded at room temperature

using a Bruker EMX X-band spectrometer equipped

with 100 kHz modulation and a cylindrical ER

4103TM cavity. Samples were contained in a

flattened, aqueous-sample cell (WG-813-SQ, Wilmad,

Buena, NJ, U.S.A.) and recording of the spectra was

initiated within 2 min of sample preparation, except

where specified otherwise. Hyperfine couplings were

measured directly from the field scan. Typical EPR

spectrometer settings were: gain 2¬10', modulation

amplitude 0±1 mT, time constant 0±16 sec, scan time

84 sec, resolution 1024 points, centre field 348 mT,

field scan 10 mT, power 25 mW, frequency 9±76 GHz,

with eight scans averaged. The radical species were

identified by comparison of their line shape, hyperfine

structure and g values with published reference data

(Buettner, 1987). The EPR signal intensity (line

height) is directly proportional to the radical con-

centration (providing the line shapes are constant)

and was therefore used to compare radical concen-

trations in the various samples (Weil, Bolton and

Wertz, 1994).

Atomic Absorption Spectrometry

For the determination of total Fe and Cu in lenses,

the tissues were first dried under vacuum and then

made up to 10% (wv) in concentrated HNO$. The

samples were subsequently sealed in plastic tubes

(Eppendorf), which had been previously rinsed with

10% (vv) HNO$in H

#O, wrapped with plastic film and

incubated at 37°C for 48 hr to completely digest the

tissue (to form a clear yellow-colored solution). Blank

samples which contained no lens material were also

prepared. The acid-digested samples were then

analysed by atomic absorption spectrometry (Varian,

SpectrAA 1020), with a graphite furnace employed

for the Cu determinations. Standard solutions of Fe

and Cu were prepared in 10% (vv) HNO$

in acid-

washed glassware.

Statistical Analysis

Statistical significance was determined by using the

two-tailed Student’s t-test for unpaired data for

interlenticular comparisons (eg. type II vs type IV

lenses) and paired data for intralenticular comparisons

(e.g. nuclei vs cortices).

3. Results

Generation of HO[ by Cataractous Lens Homogenates

The ability of cataractous lens homogenates to form

HO[ when exposed to exogenous H#O#was assessed by

EPR spin trapping using DMPO. The DMPO-HO[

F. 1. Cataractous lens stimulates hydroxyl radicalformation. EPR spectra of homogenized type IV cataractouslens in Chelex-treated PBS. The samples contained 0±1 DMPO and 1 m H

#O#. (A) Lens homogenate in PBS; (B) as

(A) except with 50% (vv) ethanol ; (C) as (A) except with10 m DETAPAC. The signals marked (X) and (E) in (A)and (B) are assigned to DMPO-HO[ and DMPO-C[H(CH

$)OH

respectively. The species arrowed in (A) has not beenidentified. Residual signal due to DMPO-HO[ is also presentin (B) and to a lesser extent in (C). The minor unmarkedsignals present [particularly evident in (C)] are due to DMPOdegradation products.

adduct [Fig. 1(A)] was identified by its line shape,

hyperfine structure and g value of 2±0057 (Buettner,

1987). Since it has been shown that the DMPO-HO[adduct can also arise via reaction of superoxide with

this trap, experiments were also conducted in the

presence of 50% (vv) ethanol. HO[ is scavenged by

ethanol to produce a DMPO-C[H(CH$)OH adduct

with a characteristic EPR signal which is not formed

on reaction of ethanol with superoxide (Halliwell and

Gutteridge, 1989). Fig. 1(B) shows that the DMPO-

HO[ signal was inhibited in the presence of ethanol,

and the DMPO-C[H(CH$)OH adduct signal was also

detected. This confirms that cataractous lens homo-

genates can catalyse HO[ production. In the presence

of the metal ion chelator, DETAPAC, the DMPO-HO[adduct signal was suppressed by 60% [Fig. 1(C)]. This

suggests that transition metal ions (such as Fe or Cu)

present in the lens homogenates can catalyse HO[production. A radical signal (broad singlet) was also

observed at a g ca. 2±006 [arrowed, Fig. 1(A)]. The

identity of this species is not known at present but may

be due to protein-derived radicals (Davies and Dean,

1997).

Comparison of Type II and Type IV Lenses

Given that cataractous lens homogenates were able

to catalyse HO[ formation, we then investigated the

relative capacity of type II vs type IV lenses to generate

this species. Type IV lens nuclei generated significantly

higher (P¯0±018) concentrations of DMPO-HO[

84 B. GARNER ET AL.

F. 2. Comparison of type IV and type II lenses for theircapacity to stimulate hydroxyl radical formation. Individuallens nuclei and cortices were homogenized at the sameconcentration (wet weight) and analysed by EPR spec-troscopy for HO[ production. With the exception that 8 m(n¯1) or 10 m (n¯2) H

#O#

was present in the controlconditions, experimental conditions are as in Fig. 1(A). Thedata refer to EPR signal intensity where the averageconcentration of DMPO-HO[ detected in the type IVhomogenates is defined as 1. Data are means³.. (n¯3for Con, n¯7 for Type II and n¯8 for Type IV). Thecontrol (Con) lenses were from one male and two femalesaged 37, 61 and 69 years old, respectively.

compared to type II nuclei (Fig. 2). The same trend

was also observed for the lens cortices (P¯0±078, Fig.

2). The radical of unknown identity [arrowed, Fig.

1(A)] was also found to be increased by 40% in type

IV compared with type II lenses.

In agreement with earlier data derived from EPR

analysis of frozen lens sections (Borkman and Lerman,

1977), we found that non-cataractous lens homo-

genate without additions did not produce detectable

levels of radicals. Even on addition of very high

concentrations of H#O#

(up to 800 m), only a

weak, broad, radical signal (g ca. 2±006) was detected,

possibly due to the formation of crystallin-bound

radicals. Bubbles were generated in these mixtures,

presumably due to catalase activity and the con-

comitant formation of O#, as has been previously

observed (Zigman, Schultz and Schultz, 1998). When

DMPO was also present in the non-cataractous lens

homogenates, a DMPO-HO[ signal was observed,

though approximately ten-fold higher concentrations

of H#O#

(10 m) were required in order to achieve

similar radical adduct concentrations to those detected

in the cataractous lenses (Fig. 2). The levels of DMPO-

HO[ detected in the non-cataractous lens nuclei were

lower than for both the type II (P¯0±012) and type IV

(P¯0±0002) nuclei. Comparing the cortical samples,

there was a significant difference between the non-

cataractous and type IV samples (P¯0±024). This

indicates that, compared to the cataractous tissue, the

non-cataractous lenses were not as effective in

catalysing HO[ production.

When the generation of HO[ by the nucleus and

cortex in each of the individual cataractous lenses was

F. 3. Correlation of the capacity of individual cataractouslens nuclei and cortices to stimulate hydroxyl radicalformation. Individual data from each lens is plotted as afunction of the concentration of DMPO-HO[ generated ineither the cortex or nucleus (E), type IV; (D), type II. Forexperimental conditions see Fig. 1(A).

F. 4. Total lenticular Fe and Cu levels. Cataractous andhealthy ‘control ’ lenses were analysed for total Fe and Culevels using atomic absorption spectrometry as described inthe ‘materials and methods ’ section. Data aremeans³.. (n¯3 for Con, n¯6 for Type II, n¯6 forType IV).

compared, a significant positive correlation was

observed (Fig. 3), i.e. for each lens, the capacity for the

nucleus to generate radicals was related to that of the

cortex. The increased HO[ production observed with

the type IV lenses could be ascribed to several factors.

If the HO[ arises via a Fenton-type system, one

possible reason for this enhancement is that the more

advanced cataractous lenses contained increased

concentrations of transition metal ions. The two

physiologically relevant lenticular metals in this

context are Fe and Cu (Lakomaa and Eklund, 1978).

We therefore measured total lenticular Fe and Cu

contents using atomic absorption spectrometry. On

average, the Fe content of the type IV lenses was E40% higher than that of type II lenses (Fig. 4).

HYDROXYL RADICALS AND CATARACT 85

F. 5. Stimulation of ascorbyl radical formation bycataractous lens. A type IV cataractous lens was divided intonuclear and cortical regions, homogenized in Chelexed PBS,and made up to 5 m with respect to ascorbate con-centration. The nuclear (A) and cortical (B) homogenatesboth generated ascorbyl radicals. (C) buffer ‘control ’ whichcontained Chelex-treated PBS and ascorbate only.

However, due to intersample variation, these

differences were not statistically significant. The Cu

concentration in both classes of cataractous lens was

very similar and E16–29% of the Fe concentration

(Fig. 4). The absolute levels of Fe and Cu detected are

within the ranges reported previously (Lakomaa and

Eklund, 1978) and the higher levels of Fe compared to

Cu also confirm previous data (Eckhert, 1983;

Lakomaa and Eklund, 1978). We did not detect a

significant difference between lenticular Fe or Cu levels

present in the nucleus vs the cortex of individual

lenses (data not shown).

While healthy ‘control ’ (non-cataractous) lenses

were not available from the same (Indian) source, a

comparison was made with lenses from a local Eye

Bank. The total amounts of Fe and Cu detected in these

non-cataractous lenses were lower than in both types

of cataractous lens (Fig. 4), consistent with previous

observations (Cekic, 1998; Cekic et al., 1999; Cook

and McGahan, 1986; Lakomaa and Eklund, 1978;

Nath et al., 1969).

Cataractous Lenses Generate Ascorbyl Radical

The data above indicate that lenticular transition

metals may be present in a redox active form.

However, it is known that if Fe and Cu are tightly

bound to physiological chelators such as ferritin and

ceruloplasmin, respectively, their ability to stimulate

HO[ via Fenton-type reactions is markedly reduced. It

is also possible that the addition of a large (1 m)

bolus dose of H#O#

to lens homogenates might have

altered the structure of metal binding sites in a way

that increased their redox availability. We therefore

also used ascorbate as a probe for metal ion avail-

ability. In the presence of Fe or Cu, ascorbate is

oxidized to give the ascorbyl radical which is readily

detected by EPR spectroscopy. Due to this, the

formation of the ascorbyl radical has been used

previously to indicate the presence of catalytically-

active transition metal ions in biological samples

(McCormick, Buettner and Britigan, 1998).

F. 6. Lack of ascorbyl radical formation in homogenizednon-cataractous lens. A non-cataractous lens (37 year oldmale) was homogenized in Chelexed PBS, and analysed byEPR spectroscopy for the presence of the ascorbyl radical (A).The contralateral lens was also homogenized and made upto 0±1 m with FeCl

$ethylenediaminetetraacetic acid (1:1

molar ratio) (B). Exogenous ascorbate was not added to thesesamples (in contrast to Fig. 5). The ascorbyl radical wasdetected only where the Fe-chelate was added (B).

Fig. 5 shows that the ascorbyl radical was generated

in both the nuclear and cortical fractions of

cataractous lens homogenates. This provides further

evidence that redox-active metal ions were present in

the lens samples. Only a very weak radical signal was

present in the Chelexed phosphate buffer (control),

indicating that it is unlikely that a significant

proportion of the radicals detected by EPR were due to

contaminating metals. These data are in agreement

with a recent report that cataractous lens proteins

oxidize exogenously added ascorbate (Saxena, Saxena

and Monnier, 1998). In contrast to the cataractous

lens, the ascorbyl radical was not detected in fresh

non-cataractous lens homogenates unless metal was

added (Fig. 6). The non-cataractous lens homogenates

contained a sufficient quantity of endogenous

ascorbate to generate a radical signal when metal was

added. For this reason, and in contrast to the

cataractous lens homogenates, it was not necessary to

add ascorbate to the non-cataractous lens homo-

genates.

4. Discussion

In the present work we studied directly the

generation of HO[ in human lenses at different stages

of nuclear cataract severity. It should be emphasized

that the cataractous lenses studied were probably not

in the same physiological state as when they were

removed from patients (e.g. antioxidant defence

mechanisms may have been compromized). However,

the fact that we were able to study both moderately

and severely diseased lenses (with identical post-

surgical histories) in parallel, has allowed us to make

comparisons between the ability of the two classes of

cataractous lens to catalyse HO[ production. The data

showed that the severity of nuclear cataract is

positively correlated with the capacity of the lens to

stimulate HO[ production. Several factors need to be

considered in order to understand the biochemical

mechanisms that may underlie this observation.

86 B. GARNER ET AL.

F. 7. Mechanism of HO[ production in the lens. H#O#

generated in the aqueous humour via autoxidation ofascorbic acid (for example) may diffuse into the lens.Alternatively, given the great capacity for the anteriorepithelial cells to degrade H

#O#

via catalase and GPx,intralenticular autoxidation of ascorbate or 3-hydroxy-kynurenine (for example) could generate a small, butrenewable and potentially toxic, nuclear pool of H

#O#. This

may react with transition metals such as Fe and Cu via theFenton reaction to produce HO[, which has been shown tobe involved in damage of lens proteins in nuclear cataract.Redox cycling of the oxidized metal back to its reduced statecould be achieved via interaction with intralenticularreductants such as ascorbate and GSH. The exact location ofmetal ions within the lens is at present unknown. AA,ascorbic acid ; DHA, dehydroascorbic acid ; GPx, glutathioneperoxidase ; 3OHKyn, 3-hydroxykynurenine.

The lenticular concentration of metal ions could

play a key role. Our data concerning the apparently

higher levels of both Fe and Cu in cataractous vs

healthy ‘control ’ lenses are in general agreement with

earlier studies (Cekic, 1998; Cekic et al., 1999; Cook

and McGahan, 1986; Lakomaa and Eklund, 1978;

Nath et al., 1969). However, while the storage and

processing of these control lenses was conducted in the

same manner as the cataractous lenses (see ‘materials

and methods’) we cannot draw firm conclusions

regarding the levels of these metals in healthy vs

cataractous lenses, as the sample sources and histories

are dissimilar.

Our data indicated that, on average, Fe levels tended

to be higher in the type IV lenses compared to type II.

This could contribute to the differences in HO[generation which we observed. In addition, the site at

which Fe is bound, and both the redox state (i.e. Fe(II)

or Fe(III) and redox availability (i.e. ‘ loosely ’ bound vs

tightly chelated in a region which restricts access to

H#O#

or DMPO) is likely to be important. Little is

known about possible metal ion binding sites within

the nucleus of the human lens. Of possible significance,

patients which have a marked over production of L-

ferritin [which lacks ferroxidase activity (Levi et al.,

1994)] have a high incidence of premature cataract

(Cazzola et al., 1997). Since Cu ions can be more

effective catalysts of HO[ production than Fe (Yugay et

al., 1996), Cu-binding sites may also be crucial in

cataract development.

It is also possible that as a cataract progresses, Fe or

Cu become bound at other sites which can catalyse

radical production. Since there is little or no protein

turnover in the lens nucleus, it is possible that metal-

binding proteins become modified during cataracto-

genesis (e.g. by radical damage) and this alters the

redox availability of previously chelated metal ions.

Similar mechanisms of Fe-induced oxidative stress

have been identified in other cell types, specifically

where Fe-binding proteins are degraded in the

lysosome (Garner et al., 1997). Another possibility is

that the formation of colored, melanin-like complexes

in the cataractous lens could bind metal ions and

thereby contribute to radical formation (Pilas et al.,

1988). Furthermore, recent studies have suggested

that crystallins from cataractous lenses have a greater

capacity to stimulate ascorbate oxidation when com-

pared to crystallins derived from non-diseased lenses

(Saxena et al., 1998). The latter work demonstrated

that DETAPAC was less effective as an inhibitor of

ascorbate oxidation in the cataractous lens material,

and proposed that progressive crystallin modification

(glycation to form advanced glycation end-products)

increased the amount of metal ion binding to the

crystallins. Such a mechanism of increased metal ion-

catalysed radical production is consistent with pre-

vious work, demonstrating increased ascorbyl radical

levels in human nuclear cataracts compared to

controls (Lohmann, Schmehl and Strobel, 1986), and

with our data showing that the nuclei of type IV lenses

are better catalysts of HO[ production when compared

to type II or non-cataractous lenses.

The events which result in the extensive protein

oxidation associated with age-related nuclear cataract

(Fu et al., 1998) are likely to involve decreased

antioxidant defence. Under normal circumstances,

most lenticular H#O#would be removed enzymatically.

However, it appears that there is very little catalase

activity in the nucleus of aged human lenses (Zigman

et al., 1998). It is also recognized that while nuclear

GSH peroxidase is present at the onset of cataract, its

activity is diminished by E70% (Fecondo and

Augusteyn, 1983). Recent data from our laboratory

has shown that as the lens ages, transport of GSH to

the lens nucleus is markedly reduced (Sweeney and

Truscott, 1998). As GSH is required for GSH

peroxidase-mediated H#O#

removal, the age-related

‘barrier ’ to GSH diffusion would be predicted to lead to

an increase in the H#O#

concentration of the nucleus

and thereby increase the metal-catalysed radical flux,

protein oxidation and thus possibly nuclear cataract

(Sweeney and Truscott, 1998).

In summary, the present studies demonstrate that

the capacity for the ocular lens to catalyse HO[

HYDROXYL RADICALS AND CATARACT 87

formation in the presence of H#O#

is correlated with

the stage of cataract. This may be partly explained by

the lenticular metal ion content and its role as a

catalyst in Fenton-type reactions (Fig. 7). Future

studies should be directed towards gaining an under-

standing of the precise metal ion binding sites within

the lens and their possible role in modulating metal

ion-catalysed radical production.

Acknowledgements

We thank Sue Butler (Chemistry Dept., University ofWollongong) for performing the atomic absorption spec-trometry. The surgeons of the Madurai eye clinic (India) andJane Taylor of the Lions Eye Bank (Sydney, Australia) arethanked for the provision of lenses. The research wassupported in part by grants from the Australian NationalHealth and Medical Research Council (RJWT, grantu980495) and the Australian Research Council (MJD, QueenElizabeth II Fellowship and grantu A29906636).

References

Augusteyn, R. C. (1974). Human lens albuminoid. Jap. J.Ophthalmol. 18, 127–34.

Berman, E. R. (1991). Biochemistry of the eye. Plenum Press,New York, U.S.A.

Buettner, G. R. (1987). Spin trapping: ESR parameters ofspin adducts. Free Radic. Biol. Med. 3, 259–303.

Bleau, G., Giasson, C. and Brunette, I. (1998). Measurementof hydrogen peroxide in biological samples containinghigh levels of ascorbic acid. Anal. Biochem. 263, 13–7.

Borkman, R. F. and Lerman, S. (1977). Evidence for a freeradical mechanism in aging and u.v.-irradiated oculardisease. Exp. Eye Res. 25, 303–9.

Cazzola, M., Bergamaschi, G., Tonon, L., Arbustini, E.,Grasso, M., Vercesi, E., Barosi, G., Bianchi, P. E., Cairo,G. and Arosio, P. (1997). Hereditary hyperferritinemia-cataract syndrome: Relationship between phenotypesand specific mutations in the iron-responsive element offerritin light-chain mRNA. Blood 90, 814–21.

Cekic, O. (1998). Copper, lead, cadmium and calcium incataractous lenses. Ophthal. Res. 30, 49–53.

Cekic, O., Bardak, Y., Totan, Y., Kavakli, S., Akyol, O.,Ozdemir, O. and Karel, F. (1999). Nickel, chromium,manganese, iron and aluminium levels in humancataractous and normal lenses. Ophthal. Res. 31, 332–6.

Cook, C. S. and McGahan, M. C. (1986). Copper concen-tration in cornea, iris, normal, and cataractous lensesand intraocular fluid of vertebrates. Curr. Eye Res. 5,69–77.

Cui, X. L. and Lou, M. F. (1993). The effect and recovery oflong term H

#O#

exposure on lens morphology andbiochemistry. Exp. Eye Res. 57, 156–67.

Davies, M. J. and Dean, R. T. (1997). Radical-Mediated ProteinOxidation. From Chemistry to Medicine. Oxford UniversityPress, Oxford, United Kingdom.

Dische, Z. and Zil, H. (1951). Studies on the oxidation ofcysteine to cystine in lens proteins during cataractformation. Am. J. Ophthalmol. 34, 104–13.

Eaton, J. (1997). Is the lens canned? Free Radic. Biol. Med.11, 201–13.

Eckhert, C. D. (1983). Elemental concentrations in oculartissues of various spacies. Exp. Eye Res. 37, 639–47.

Fecondo, J. V. and Augesteyn, R. C. (1983). Superoxidedismutase, catalase and glutathione peroxidase inhuman cataractous lens. Exp. Eye Res. 36, 15–23.

Fu, S., Dean, R. T., Southan, M. and Truscott, R. J. W.(1998). The hydroxyl radical in lens nuclear cataracto-genesis. J. Biol. Chem. 273, 28603–9.

Garland, D. (1990). Role of site-specific, metal-catalyzedoxidation in lens aging and cataract. Exp. Eye Res. 50,677–82.

Garland, D., Zigler, S. J. and Kinoshita, J. (1986). Structuralchanges in bovine lens crystallins induced by ascorbate,metal, and oxygen. Arch. Biochem. Biophys. 251, 771–6.

Garner, B., Li, W., Roberg, K. and Brunk, U. T. (1997). Onthe cytoprotective role of ferritin in macrophages and itsability to enhance lysosomal stability. Free Rad. Res. 27,487–500.

Garner, M. H. and Spector, A. (1980). Selective oxidation ofcysteine and methionine in normal and senilecataractous lenses. Proc. Natl. Acad. Sci. USA 77,1274–7.

Guptasarma, P., Balasubramanian, D., Matsugo, S. andSaito, I. (1992). Hydroxyl radical mediated damage toproteins, with special reference to the crystallins.Biochemistry 31, 4296–303.

Halliwell, B. and Gutteridge, J. M. C. (1989). Free Radicals inBiology and Medicine. Oxford University Press, Oxford,United Kingdom.

Huizinga, A., Bot, A. C. C., De Mul, F. F. M., Vrensen,G. F. J. M. and Greve, J. (1989). Local variation inabsolute water content of human and rabbit eye lensesmeasured by Raman microspectroscopy. Exp. Eye Res.48, 487–96.

Kupfer, C., Underwood, B. and Gillen, T. (1994). Leadingcauses of visual impairment world wide. In: Principlesand Practice of Ophthalmology Basic Science. (Albert,D. M. & Jakobiec, F. A., eds) W. B. Saunders & Co.,Philadelphia, USA.

Lakomaa, E. L. and Eklund, P. (1978). Trace elementanalysis of human cataractous lenses by neutronactivation analysis and atomic absorption spectrometrywith special reference to pseudo-exfoliation of the lenscapsule. Ophthal. Res. 10, 302–6.

Lerman, S. and Borkman, R. (1976). Spectroscopic evalu-ation and classification of the normal, aging, andcataractous lens. Ophthal. Res. 8, 335–53.

Lerman, S., Kuck, J. F., Borkman, R. and Saker, E. (1976).Acceleration of an aging parameter (fluorogen) in theocular lens. Ann. Ophthalmol. 5, 558–61.

Levi, S., Santambrogio, P., Cozzi, A., Rovida, E., Corsi, B.,Tamborini, E., Spada, S. (1994). The role of the L-chainin ferritin iron incorporation-studies of homo andheteropolymers. J. Mol. Biol. 238, 649–54.

Lohmann, W., Schmehl, W. and Strobel, J. (1986). Nuclearcataract : oxidative damage to the lens. Exp. Eye Res.43, 859–62.

McCormick, M. L., Buettner, G. R. and Britigan, B. E. (1998).Endogenous superoxide dismutase levels regulate iron-dependent hydroxyl radical formation in escherichiacoli exposed to hydrogen peroxide. J. Bacteriol. 180,622–5.

Nath, R., Srivastava, S. K. and Singh, K. (1969). Ac-cumulation of copper and inhibition of lactate de-hydrogenase activity in human senile cataractous lens.Exp. Biology 7, 25–6.

Pilas, B., Sarna, T., Kalyanaraman, B. and Swartz, H. M.(1988). The effect of melanin on iron associateddecomposition of hydrogen peroxide. Free Radic. Biol.Med. 4, 285–93.

Pirie, A. (1968). Color and solubility of the proteins ofhuman cataract. Invest. Ophthalmol. 7, 634–42.

Reddan, J. R., Giblin, F. J., Kadry, R., Leverenz, V. R., Pena,J. T. and Dziedzic, D. C. (1999). Protection from oxi-dative insult in glutathione depleted lens epithelial cells.Exp. Eye Res. 68, 117–27.

88 B. GARNER ET AL.

Saxena, P., Saxena, A. K. and Monnier, V. M. (1998).Evidence of transition metal-catalyzed oxidation ofascorbate by advanced glycation end products in thecataractous human lens (abstract). Invest. Ophthalmol.Vis. Sci. 39, S-891.

Smith, J. B., Jiang, X. G. and Abraham, E. C. (1997).Identification of hydrogen peroxide oxidation sites ofalpha-A- and alpha-B-crystallins. Free Rad. Res. 26,103–11.

Spector, A. (1995). Oxidative stress-induced cataract :mechanism of action. FASEB J. 9, 1173–82.

Spector, A. and Garner, W. H. (1981). Hydrogen peroxideand human cataract. Exp. Eye Res. 33, 673–81.

Spector, A., Ma, W. and Wang, R.-R. (1998). The aqueoushumour is capable of generating and degrading H

#O#.

Invest. Ophthalmol. Vis. Sci. 39, 1188–97.Spector, A., Wang, G.-M., Wang, R.-R., Garner, W. H. and

Moll, H. (1993). The prevention of cataract caused byoxidative stress I. H

#O#

and photochemically inducedcataract. Curr. Eye Res. 12, 163–79.

Swanson, A. A. and Truesdale, A. W. (1971). Elementalanalysis in normal and cataractous human lens tissue.Biochem. Biophys. Res. Comm. 45, 1488–97.

Sweeney, M. H. J. and Truscott, R. J. W. (1998). An im-pediment to glutathione diffusion in older normalhuman lenses : A possible precondition for nuclearcataract. Exp. Eye Res. 67, 587–95.

Truscott, R. J. W. and Augusteyn, R. C. (1977a). Changes in

human lens proteins during nuclear cataract formation.Exp. Eye Res. 24, 159–70.

Truscott, R. J. W. and Augusteyn, R. C. (1977b). Oxidativechanges in human lens proteins during senile nuclearcataract formation. Biochim. Biophys. Acta 492, 43–52.

van Reyk, D. M., Brown, A. J., Jessup, W. and Dean, R. T.(1995). Batch-to-batch variation of chelex-100 con-founds metal-catalysed oxidation. Leaching of inhibi-tory compounds from a batch of chelex-100 and theirremoval by a pre-washing procedure. Free Rad. Res. 23,533–5.

Vazquez, S., Garner, B., Sheil, M. M. and Truscott, R. J. W.(1999). Characterisation of the major autoxidationproducts of 3-hydroxykynurenine under physiologicalconditions. Free Rad. Res. in the press.

Weil, J. A., Bolton, J. R. and Wertz, J. E. (1994). Electronparamagnetic resonance : Elementary theory and practicalapplications. John Wiley and Sons, New York, U.S.A.

Yugay, M. T., Pereira, P. C., Leiria, F. and Mota, M. C.(1996). Oxidative damage to lens membranes inducedby metal-catalyzed systems. Ophthal. Res. 28, 92–6.

Zigler, J. S., Huang, Q.-L. and Du, X.-Y. (1989). Oxidativemodification of lens crystallins by H

#O#

and chelatediron. Free Radic. Biol. Med. 7, 499–505.

Zigman, S., Schultz, J. B. and Schultz, M. (1998). Measure-ment of oxygen production by in vitro human andanimal lenses with an oxygen electrode. Curr. Eye Res.17, 115–9.