Automated Assessment of the DWI-FLAIR Mismatch in Patients with Acute Ischemic Stroke: Added Value to Routine Clinical Practice

BACKGROUND AND PURPOSE

The DWI-FLAIR mismatch is used to determine thrombolytic eligibility in patients with acute ischemic stroke when the time since stroke onset is unknown. Commercial software packages have been developed for automated DWI-FLAIR classification. We aimed to use e-Stroke software for automated classification of the DWI-FLAIR mismatch in a cohort of patients with acute ischemic stroke and in a comparative analysis with 2 expert neuroradiologists.

MATERIALS AND METHODS

In this retrospective study, patients with acute ischemic stroke who had MR imaging and known time since stroke onset were included. The DWI-FLAIR mismatch was evaluated by 2 neuroradiologists blinded to the time since stroke onset and automatically by the e-Stroke software. After 4 weeks, the neuroradiologists re-evaluated the MR images, this time equipped with automated predicted e-Stroke results as a computer-assisted tool. Diagnostic performances of e-Stroke software and the neuroradiologists were evaluated for prediction of DWI-FLAIR mismatch status.

RESULTS

A total of 157 patients met the inclusion criteria. A total of 82 patients (52%) had a time since stroke onset of ≤4.5 hours. By means of consensus reads, 81 patients (51.5%) had a DWI-FLAIR mismatch. The diagnostic accuracy (area under the curve/sensitivity/specificity) of e-Stroke software for the determination of the DWI-FLAIR mismatch was 0.72/90.0/53.9. The diagnostic accuracy (area under the curve/sensitivity/specificity) for neuroradiologists 1 and 2 was 0.76/69.1/84.2 and 0.82/91.4/73.7, respectively; both significantly (P < .05) improved to 0.83/79.0/86.8 and 0.89/92.6/85.5, respectively, following the use of e-Stroke predictions as a computer-assisted tool. The interrater agreement (κ) for determination of DWI-FLAIR status was improved from 0.49 to 0.57 following the use of the computer-assisted tool.

CONCLUSIONS

This automated quantitative approach for DWI-FLAIR mismatch provides results comparable with those of human experts and can improve the diagnostic accuracies of expert neuroradiologists in the determination of DWI-FLAIR status.

Arterial spin-labeling perfusion MRI stratifies progression-free survival and correlates with epidermal growth factor receptor status in glioblastoma

BACKGROUND AND PURPOSE

Glioblastoma is a common primary brain tumor with a poor but variable prognosis. Our aim was to investigate the feasibility of MR perfusion imaging by using arterial spin-labeling for determining the prognosis of patients with glioblastoma.

MATERIALS AND METHODS

Pseudocontinuous arterial spin-labeling with 3D background-suppressed gradient and spin-echo was acquired before surgery on 53 patients subsequently diagnosed with glioblastoma. The calculated CBF color maps were visually evaluated by 3 independent readers blinded to patient history. Pathologic and survival data were correlated with CBF map findings. Arterial spin-labeling values in tumor tissue were also quantified by using manual fixed-size ROIs.

RESULTS

Two perfusion patterns were characterized by visual evaluation of CBF maps on the basis of either the presence (pattern 1) or absence (pattern 2) of substantial hyperperfused tumor tissue. Evaluation of the perfusion patterns was highly concordant among the 3 readers (κ = 0.898, P < .001). Pattern 1 (versus pattern 2) was associated with significantly shorter progression-free survival by Kaplan-Meier analysis (median progression-free survival of 182 days versus 485 days, P < .01) and trended with shorter overall survival (P = .079). There was a significant association between pattern 1 and epidermal growth factor receptor variant III expression (P < .01).

CONCLUSIONS

Qualitative evaluation of arterial spin-labeling CBF maps can be used to stratify survival and predict epidermal growth factor receptor variant III expression in patients with glioblastoma.

Perfusion deficits detected by arterial spin-labeling in patients with TIA with negative diffusion and vascular imaging

BACKGROUND AND PURPOSE

A substantial portion of clinically diagnosed TIA cases is imaging-negative. The purpose of the current study is to determine if arterial spin-labeling is helpful in detecting perfusion abnormalities in patients presenting clinically with TIA.

MATERIALS AND METHODS

Pseudocontinuous arterial spin-labeling with 3D background-suppressed gradient and spin-echo was acquired on 49 patients suspected of TIA within 24 hours of symptom onset. All patients were free of stroke history and had no lesion-specific findings on general MR, DWI, and MRA sequences. The calculated arterial spin-labeling CBF maps were scored from 1-3 on the basis of presence and severity of perfusion disturbance by 3 independent observers blinded to patient history. An age-matched cohort of 36 patients diagnosed with no cerebrovascular events was evaluated as a control. Interobserver agreement was assessed by use of the Kendall concordance test.

RESULTS

Scoring of perfusion abnormalities on arterial spin-labeling scans of the TIA cohort was highly concordant among the 3 observers (W = 0.812). The sensitivity and specificity of arterial spin-labeling in the diagnosis of perfusion abnormalities in TIA was 55.8% and 90.7%, respectively. In 93.3% (70/75) of the arterial spin-labeling CBF map readings with positive scores (≥2), the brain regions where perfusion abnormalities were identified by 3 observers matched with the neurologic deficits at TIA onset.

CONCLUSIONS

In this preliminary study, arterial spin-labeling showed promise in the detection of perfusion abnormalities that correlated with clinically diagnosed TIA in patients with otherwise normal neuroimaging results.

Intracochlear schwannoma presenting as diffuse cochlear enhancement: diagnostic challenges of a rare cause of deafness

Intracochlear schwannoma is a rare, treatable, cause of unilateral hearing loss. Due to the small size, position, and variable clinical and imaging features, diagnosis presents a significant challenge and is often delayed. We present a case of a patient with an intracochlear schwannoma presenting as a diffuse enhancement of the cochlea, mimicking an infectious or inflammatory process. The absence of focal nodularity in this lesion on multiple high-resolution MRI examinations led to a delay of over 3 years from the patient's initial presentation to surgical diagnosis. Clinical history and examination, imaging features, pathologic findings, and surgical management options are described.

Sugar-mediated crosslinking of alpha-biotinylated-Lys to cysteamine-agarose support: a method to isolate Maillard Lys-Lys-like crosslinks

Advanced glycation end products (AGEs) and, specifically, protein-protein AGE crosslinks have long been studied for their potential role in aging, diabetic complications and Alzheimer disease. With few exceptions, the chemical nature of these structures remains unknown. We report here a simple approach that allows the preparation and isolation of milligram quantities of sugar-mediated AGE Lys-Lys-like crosslinks from glycation mixtures. The method is based on a sugar-dependent incorporation of N(alpha)-biotinyl-L-Lys into cysteaminyldisulfide Sepharose 6B (AE-S-S-Sepharose 6B). Glycation mixtures with six different sugars showed a time- and sugar-dependent decrease in the concentration of the support-bound primary amino groups and accounted for almost 90% loss of cysteaminyl amino groups at the end of the various incubation periods. 4-Hydroxyazobenzene-2-carboxylic acid-avidin assays indicated the incorporation of N(alpha)-biotinyl-L-Lys equal to 8% of the total support amino groups with methylglyoxal after 7 d and 1% with fructose and glucose after 1 mo of incubation. Treatment of the washed, sugar-modified supports with 2-mercaptoethanol released the bulk of the bound AGE modifications and the crosslinks. Subsequent fractionation of these preparations over a monomeric avidin column afforded a complete separation of sugar-mediated AGE modifications and the crosslinks. Depending on the sugar employed, micromolar amounts of biotinylated Lys-Lys-like crosslinks were generated by this two-step procedure from 8 mL of the original AE-S-S-Sepharose 6B.

Spontaneous generation of superoxide anion by human lens proteins and by calf lens proteins ascorbylated in vitro

The proteins isolated from aged human lenses and brunescent cataracts exhibit extensive disulfide bond formation. Diabetic rat lenses similarly contain disulfide-bonded protein aggregates. These observations are consistent with the known link between diabetes, glycation and oxidative damage, and suggest a role for reactive oxygen species (ROS) in this process. To assess whether the glycation-related modifications in human lens proteins spontaneously generate ROS, superoxide anion formation was measured using both cataractous lens proteins and calf lens proteins glycated in vitro with ascorbic acid (ascorbylated). The water-insoluble fraction from aged normal human lenses generated 0.3-0.6 nmol superoxide h(-1)mg protein(-1), whereas the activity increased to 0.5-1.8 nmol h(-1)mg protein(-1)with the WI fraction from brunescent cataracts, and 2.3 nmol h(-1)mg protein(-1)with calf lens proteins ascorbylated for 4 weeks in vitro. The activity in the human lens proteins was observed in both the water-soluble and water-insoluble fractions, and was completely dependent upon the presence of oxygen. The pH optimum curve for superoxide formation increased from pH 6.5 to 10 with both the cataract and ascorbylated proteins. The superoxide-generating activity in human lens was completely bound to a boronate affinity column, but only partially bound with the ascorbylated proteins. The superoxide anion produced by a 5 m m solution of purified N(epsilon)-fructosyl-lysine was barely detectable, and therefore, could not account for the superoxide formed by any of the lens protein preparations. Also, superoxide formation increased 10-fold at pH 8.8 with fructosyl-lysine, but only 1.3-1.8-fold with human lens proteins. The addition of copper-stimulated superoxide formation with glycated bovine serum albumin, but no stimulation was seen with cataractous proteins. Assays of specific compounds showed that catechol, hydroquinone, 3-OH kynurenine and 3-OH anthranylic acid exhibited the greatest activity for superoxide generation, but had a very short halflife. 2,3-Dihydroxypyridine and 4,5 dihydroxynaphthalene were one and two orders of magnitude less reactive. In long-term incubations at 37 degrees, cataractous proteins retained the potential to produce superoxide anion, losing only half of the initial activity after 6-7 days. Therefore, the water-insoluble fraction from aged human lenses and dark brown cataracts are potentially capable of generating >100 nmol mg protein(-1)and >170 nmol mg protein(-1)of superoxide anion respectively, likely due to the presence of advanced glycation endproducts in human lens proteins. This spontaneous generation of superoxide anion in vivo could account for a major portion of the oxidation of sulfur amino acids seen during aging and cataract formation.

Superoxide free radical generation by Amadori compounds: the role of acyclic forms and metal ions

Generation of oxygen free radicals by glycated proteins is widely believed to be one of the causes of oxidative stress in diabetes and aging. Metal ion catalysis is regarded as an essential part of the oxidative mechanism. In this work, we also considered an alternative "metal-free" superoxide radical formation by a number of fructose-amino acids (Amadori compounds) derived from glycine and lysine, which represent the simplest models for early glycated proteins. In the superoxide dismutase-dependent cytochrome c assay, 1 mM Chelex-treated aqueous solutions of monofructose-amino acids 4-6 generated 0.9-3.6 x 10(-10) M s-1 O2*- at pH 7. Surprisingly, the rates of superoxide radical formation in the solutions of difructose-amino acids 7-9 were significantly higher (0.75-5.8 x 10(-9) M s-1 O2*-). The percentage of acyclic sugar anomers (</=0. 8-85%) and their "enolization" rate constants (5 x 10(-6) to 2 x 10(-3) s-1) varied broadly for the compounds studied and positively correlated with the rates of superoxide radical formation. The presence of Cu2+ markedly increased the rate of superoxide radical formation at metal concentrations higher than 1 microM, while Fe3+ did not accelerate the reaction even at 100 microM. Therefore, in addition to the metal ion-catalyzed oxygen free radical formation, metal-free enol oxidation of fructosyl groups on glycated amino acid residues may contribute to the generation of oxygen free radicals and their subsequent oxidative damage to proteins.

The generation of superoxide anions in glycation reactions with sugars, osones, and 3-deoxyosones

Glycoxidation is a process whereby glycated proteins chemically generate oxygen free radicals. Superoxide anion formation was measured by the superoxide dismutase-dependent reduction of ferricytochrome C in glycation reactions at pH 7.0 in the absence of transition metal ions. Assays were linear over 1 h, and most activity was seen after a 2 d incubation of 5 mM L-threose and 10 mM alpha-N-acetyl-lysine (N-Ac-Lys) or 10 mg/mL RNase A. Trioses, tetroses and their corresponding osones and 3-deoxyosones had the highest activity (12-16 nmoles O.-2/hr/ml) with N-Ac-Lys. Osones and 3-deoxyosones alone generated considerable O.-2, whereas aldose sugars largely did not. Xylosone and 3-deoxyxylosone produced 6 and 10 nmoles O.-2/hr/ml respectively with N-Ac-Lys, however, xylose was inactive, as were glucose and fructose. Glycation assays with 3-deoxyglucosone and glyoxal showed no activity, however, methyl glyoxal generated 1.7 and 2.0 nmoles O.-2/hr/ml with N-Ac-Lys and N-Ac-Arg, respectively. Therefore, Amadori compounds composed of lysine and short chain sugars can rapidly generate superoxide anion in the absence of metal ions.

The aggregation in human lens proteins blocks the scavenging of UVA-generated singlet oxygen by ascorbic acid and glutathione

One hour of UVA irradiation of air-saturated solutions of 2 mg/mL solubilized lens protein aggregates from aged human lens is able to produce on accumulated concentration of more than 2mM 1O2, along with oxidation of 120 nmol/mL of both Trp and His amino acid residues. Increasing concentrations of either sodium azide or ascorbic acid (up to 10 mM) during the irradiation decreased th His destruction by no more than 50-60% with the intact aggregates, but completely prevented the His loss with proteolyzed aggregates. Glutathione (up to 10 mM) was able to protect less than 10% of the aggregate His residues from oxidative damage, whereas His loss was almost completely prevented in the proteolyzed aggregates. Similar data were obtained for teh UVA photolysis of the Trp residues. This finding led us to study the role a protein conformation of these aggregates plays in the diminishing of antioxidant ability to prevent UVA-mediated photolysis of 1O2-sensitive amino acid residues. We found that Trp, His, and Cys are buried in the aggregates and cannot be oxidized by a relatively high concentration of 1O2 generated externally to the protein. Increasing urea denaturation of the aggregates caused exposure of the buried Trp residues as determined by the red shift of the fluorescence maximum and by a marked increase in the acrylamide and iodide fluorescence quenching. The ability of glutathione to prevent Trp oxidation by UVA light correlated directly with the extent of Trp exposure. These data suggest that the aggregation of the lens crystallins during aging produces a barrier, which prevents the access of water-soluble antioxidants to the sites of UVA-dependent singlet oxygen generation. In this case UVA proteolysis of the lens proteins can proceed even in the presence of physiological levels of antioxidants.

UVA irradiation of human lens proteins produces residual oxidation of ascorbic acid even in the presence of high levels of glutathione

The oxidation products of ascorbic acid (AscH-) can rapidly glycate and crosslink lens proteins in vitro, producing fluorophores and browning products similar to those present in cataractous lenses. The accumulation of AscH- oxidation products, however, would largely be prevented by the millimolar levels of glutathione (GSH) present in human lens. Here we investigate whether protein aggregation could allow the oxidation of AscH- by UVA-induced reactive oxygen species in the presence of physiological levels of GSH. The metal-catalyzed oxidation of 1.0 mM AscH- by 50 microM Cu(II) was almost complete after 1 h, but no oxidation was seen in the presence of GSH concentrations as low as 0.5 mM. UVA irradiation of protein aggregates from human lens, which accumulated more than 2.0 mM singlet oxygen after 1 h, caused a 50-60% oxidation of 1.0 mM AscH-. The addition of 204 mM GSH, however, decreased AscH- oxidation by less than half, and 30% of the AscH- was oxidized even in the presence of 15 mM GSH. This diminished protection may be due, in part, to the ability of AscH-, but not GSH, to penetrate to the sites of singlet oxygen generation located within the protein. Consistent with this hypothesis, greater GSH protection was seen when a proteolytic digest of the human proteins was subjected to the same irradiation or when singlet oxygen was chemically generated from 3-(4-methyl-1-naphthyl)propionic acid endoperoxide (MNPAE) at 37 degrees C in the medium. The addition of 50 microM Cu(II) had no effect on the rate of degradation of dehydroascorbic acid (DHA). Singlet oxygen, either UVA- or MNPAE-generated, increased the rate of DHA loss. This secondary oxidation of DHA by singlet oxygen would allow the accumulation of AscH- oxidation products was not reducible by GSH. Therefore, the data presented here argue that the protein aggregation seen in older human lenses may permit oxidized AscH--induced crosslinking to occur even at physiological GSH levels.

The relative UV sensitizer activity of purified advanced glycation endproducts

The oxidation products of ascorbic acid react with lens proteins to form advanced glycation endproducts (AGE) that are capable of generating reactive oxygen species when irradiated with UVA light. L-Threose, the most active of these oxidation products, was reacted with N-acetyl lysine and six AGE peaks were isolated by RP-HPLC. Each peak exhibited fluorescence and generated superoxide anion and singlet oxygen in response to UV light. Solutions of these AGE peaks (50 micrograms/mL) generated 5-10 nmol/mL of superoxide anion during a 30 min irradiation. This activity was 100-fold less than the superoxide anion generated by kynurenic acid and 400-fold less than riboflavin. Ultraviolet irradiation generated from 1.2 to 2.7 mumol/mL of singlet oxygen with the purified threose AGE compounds. This activity was similar to that seen with other purified AGE compounds (pentosidine, LM-1 and Ac-FTP) and with kynurenine and 3-OH kynurenine. This considerable singlet oxygen formation, however, was still 40-fold less than that obtained with kynurenic acid and 100-fold less than riboflavin under the same irradiation conditions. In spite of this lower sensitizer efficiency, the purified AGE generated 20-60-fold more singlet oxygen on a weight basis than either crude ascorbic acid glycated proteins or a preparation of water-insoluble proteins from aged normal human lenses. On a molar basis, therefore, AGE could account for the sensitizer activity in these protein preparations if they represented less than 1% of the total amino acids.

Quantitation of the singlet oxygen produced by UVA irradiation of human lens proteins

Ultraviolet irradiation of aged human lens proteins in vitro causes extensive photolytic damage of His and Trp residues. Protection by sodium azide argues for a process mediated by singlet oxygen (1O2). In the work described here, the synthesis of 1O2 was measured by the bleaching of N,N-dimethyl-4-nitrosoaniline (RNO), the oxidation of added histidine and the oxidation of furfuryl alcohol. To obtain a more accurate value for 1O2 generation, a known quantity of 1O2 was generated by the thermal dissociation of 3-(4-methyl-naphthyl)propionic acid endoperoxide, and the efficiency of each assay method to report on the 1O2 generated was determined. The values obtained were 0.003 mol of RNO bleached/mol of 1O2 generated, 0.55 mol of furfuryl alcohol oxidized/mol 1O2 and 0.5 mol of His oxidized/mol 1O2 generated. Irradiation of the human lens proteins with UVA light produced from 2.1 to 2.4 mM of 1O2 by RNO bleaching, 2.6-2.8 mM 1O2 by furfuryl alcohol oxidation and up to 1.9 mM of 1O2 by histidine oxidation during a 1 h irradiation period. The average value (2.2 mM of 1O2) corresponds to the theoretical production of 30 nmol of singlet oxygen at UVA light intensities equivalent to a 1 h exposure to sunlight at noon in the northern hemisphere.

The generation of superoxide anion by the UVA irradiation of human lens proteins

In this study we report the generation of superoxide anion and hydrogen peroxide by a water-insoluble protein fraction from aged human lenses in response to UVA light. Irradiation with 1.5 kJ cm-2 of UVA light ( > 338 nm) over a 1 hr period caused the formation of 20 +/- 0.1 microM superoxide radical and 37 +/- 0.5 microM hydrogen peroxide. A linear photolysis of SH-groups (21 nmol ml-1, 26%). His (117 nmol ml-1, 26%) and Trp (72 nmol ml-1, 27%) residues was seen following 60 min of irradiation. The addition of SOD, however, had no effect on the photolytic destruction of any of these amino acid residues. Incubation of the human WISS proteins and bovine alpha-crystallin in the presence of 43-49 microM of O2- generated in a xanthine oxidase/hypoxanthine system over a 1 hr period, caused no loss of histidine, little or no loss of tryptophan and loss of 7-9 nmol ml-1 of sulfhydryl groups with both proteins. This argues that O2- can only account for the destruction of at most 4-8 nmol SH-groups in human water-insoluble proteins following 1 hr of UVA irradiation.

Quantitation of the reactive oxygen species generated by the UVA irradiation of ascorbic acid-glycated lens proteins

The oxidation products of ascorbic acid rapidly glycate proteins and produce protein-bound, advanced glycation endproducts. These endproducts can absorb UVA light and cause the photolytic oxidation of proteins (Ortwerth, Linetsky and Olesen, Photochem. Photobiol. 62, 454-463, 1995), which is mediated by the formation of reactive oxygen species. A dialyzed preparation of calf lens proteins, which had been incubated for 4 weeks with 20 mM ascorbic acid in air, was irradiated for 1 h with 200 mW/cm2 of absorbed UVA light (gamma > 338 nm), and the concentration of individual oxygen free radicals was measured. Superoxide anion attained a level of 76 microM as determined by the superoxide dismutase (SOD)-dependent increase in hydrogen peroxide formation and of 52 microM by the SOD-inhibitable reduction of cytochrome c. Hydrogen peroxide formation increased linearly to 81 microM after 1 h. Neither superoxide anion nor hydrogen peroxide, however, could account for the UVA photolysis of Trp and His seen in this system. Singlet oxygen levels approached 1.0 mM as measured by the oxidation of histidine, which was consistent with singlet oxygen measurements by the bleaching of N,N-dimethyl-4-nitrosoaniline. High concentrations of sodium azide, a known singlet oxygen quencher, inhibited the photolytic destruction of both His and Trp. Little or no protein damage could be ascribed to hydroxyl radical based upon quenching experiments with added mannitol. Therefore, superoxide anion and H2O2 were generated by the UVA irradiation of ascorbate advanced glycation endproducts, however, the major reactive oxygen species formed was singlet oxygen.

Ascorbic acid glycation of lens proteins produces UVA sensitizers similar to those in human lens

Soluble calf lens proteins were extensively glycated during a 4 week incubation with ascorbic acid in the presence of oxygen. Amino acid analysis of the dialyzed proteins removed at weekly intervals showed an increasing loss of lysine, arginine and histidine, consistent with the extensive protein cross-linking observed. Irradiation of the dialyzed samples with UVA light (1.0 kJ/cm2 total illumination through a 338 nm cutoff filter) caused an increasing loss of tryptophan, an additional loss of histidine and the production of micromolar concentrations of hydrogen peroxide. No alteration in amino acid content and no photolytic effects were seen in proteins incubated without ascorbic acid or in proteins incubated with glucose for 4 weeks. The rate of hydrogen peroxide formation was linear with each glycated sample with a maximum production of 25 nmol/mg protein illuminated. The possibility that the sensitizer activity was due to an ascorbate-induced oxidation of tryptophan was eliminated by the presence of a heavy metal ion chelator during the incubation and by showing equivalent effects with ascorbate-incubated ribonuclease A, which is devoid of tryptophan. The ascorbate-incubated samples displayed increasing absorbance at wavelengths above 300 nm and increasing fluorescence (340/430) as glycation proceeded. The spectra of the 4 week glycated proteins were identical to those obtained with a solubilized water-insoluble fraction from human lens, which is known to have UVA sensitizer activity. The incubation of lens proteins with dehydroascorbic acid or L-threose, but not fructose, produced equivalent glycation, protein crosslinking and sensitizer activity.(ABSTRACT TRUNCATED AT 250 WORDS)

The generation of hydrogen peroxide by the UVA irradiation of human lens proteins

The water-insoluble proteins from aged human lens are known to contain protein-bound chromophores that act as UVA senisitizers. The irradiation of a sonication-solubilized, water-insoluble fraction from human lenses (55-75 years) with UVA light (1.5 kJ/cm2, gamma > 338nm) caused an oxygen-dependent photolysis of tryptophan, not seen when either alpha-crystallin or lysozyme were irradiated. The suggested requirement for active oxygen species was consistent with a linear increase in hydrogen peroxide formation, which was also observed. A final concentration of 55 microM H2O2 was attained, with no H2O2 being detected in either dark-incubated controls or in irradiated samples of native proteins. The UVA-dependent H2O2 formation was increased 50% by superoxide dismutase (SOD) and abolished by catalase, arguing for the initial generation of superoxide anion. A linear photolysis of histidine and tryptophan was also seen; however, the addition of SOD or SOD and catalase had no effect on the photolytic destruction of either amino acid. Superoxide dismutase increased the oxidation of protein SH groups implicating H2O2, but SOD and catalase caused a decrease in SH oxidation only at later time periods. The direct addition of H2O2 to a water-insoluble sonicate supernatant fraction caused only a slight oxidation of SH groups, but this was increased four- to eight-fold when the protein was denatured in 4.0 M guanidine hydrochloride. Overall, the data suggest a UVA-dependent oxidation of protein SH groups via H2O2 generated within the large protein aggregates of the water-insoluble fraction. These data also provide a mechanism for oxidation of the sulfur-containing amino acids in vivo--a process that is known to accompany the formation of age-onset cataracts.