Impact of Aging and Hyperbaric Oxygen In Vivo on Guinea Pig Lens Lipids and Nuclear Light Scatter

Abstract

purpose. To measure lipid compositional and structural changes in lenses as a result of hyperbaric oxygen (HBO) treatment in vivo. HBO treatment in vivo has been shown to produce increased lens nuclear light scattering.

methods. Guinea pigs, approximately 650 days old at death, were given 30 and 50 HBO treatments over 10- and 17-week periods, respectively, and the lenses were sectioned into equatorial, cortical, and nuclear regions. Lipid oxidation, composition, and structure were measured using infrared spectroscopy. Phospholipid composition was measured using ³¹P-NMR spectroscopy. Data were compared with those obtained from lenses of 29- and 644-day-old untreated guinea pigs.

results. The percentage of sphingolipid approximately doubled with increasing age (29–544 days old). Concomitant with an increase in sphingolipid was an increase in hydrocarbon chain saturation. The extent of normal lens lipid hydrocarbon chain order increased with age from the equatorial and cortical regions to the nucleus. These order data support the hypothesis that the degree of lipid hydrocarbon order is determined by the amount of lipid saturation, as regulated by the content of saturated sphingolipid. Products of lipid oxidation (including lipid hydroxyl, hydroperoxyl, and aldehydes) and lipid disorder increased only in the nuclear region of lenses after 30 HBO treatments, compared with control lenses. Enhanced oxidation correlated with the observed loss of transparency in the central region. HBO treatment in vivo appeared to accelerate age-related changes in lens lipid oxidation, particularly in the nucleus, which possesses less antioxidant capability.

conclusions. Oxidation could account for the lipid compositional changes that are observed to occur in the lens with age and cataract. Increased lipid oxidation and hydrocarbon chain disorder correlate with increased lens nuclear opacity in the in vivo HBO model.

Morphologic ^{1 2 3 4 5 6} and biophysical ^{7 8 9} studies have shown that membrane derangement occurs in human cataractous lenses. Alterations to lens membranes, possibly as a result of oxidative processes, are believed to contribute to the development of cataract. Protein oxidation in the lens has been shown to initiate at the membrane, ¹⁰ and products of lipid oxidation in the human lens increase with both age ^{11 12} and cataract. ^{11 13 14 15 16 17 18 19} In experimental animal models such as the Royal College of Surgeons (RCS) rat, lipid hydroperoxides have been reported to be a cause of cataract. ^{6 20 21 22 23} We have found that lipid oxidation is an early event in UVB-induced damage in lens epithelial membranes. ²⁴  

Green and blue fluorescence in the lens is characteristic of oxidation. It was found that the nuclear region of the lens contained more specific fluorescence than the cortex. ²⁵ Relative to the lipid content of the membrane, human nuclear cataractous membranes contained 1.6 times more green and blue fluorescence than did clear lenses (calculated from references 25 and 26, as well as Borchman D, unpublished data, 1989). A blue fluorophore with a fluorescence spectrum identical with that found in human lenses could be the result of the oxidation of sphingomyelin. ²⁷ These studies indicate that lipid oxidation and/or compositional changes in the lipid membrane may be a cause of lens opacification. 

Oxidation of membrane lipids could directly or indirectly alter the molecular structure of lens membranes. The structural features of membrane lipids are directly determined by lipid composition. ²⁸ Elevation of sphingolipid levels with age ^{29 30 31} and cataract ^{17 31 32 33} results in greater saturation of the membrane hydrocarbon chain region. ²⁸ This enhanced saturation causes the hydrocarbon chain region to become more ordered (stiff, due to C⋕C trans rotomers) with age ^{12 34} and cataract. ^{7 8 9} Oxidation directly fluidized ordered lens lipids ^{35 36} and fluidized ordered sphingomyelin ^{27 37} by inserting hydrophilic groups in the hydrophobic hydrocarbon chain region. Glycerolipids would be expected to be preferentially oxidized, because glycerolipid hydrocarbon chains have more double bonds. ²⁸ In vivo, oxidation could lead indirectly to a more ordered membrane, as has been observed with lens age ^{12 34} and cataract. ^{7 8 9} We hypothesize that lipases would eliminate oxidized glycerolipids, leaving a membrane composed of more saturated sphingolipids. Fiber membranes, in general, do not contain the machinery necessary for the synthesis of sphingolipids, and it is therefore unlikely that sphingolipids are synthesized at a greater rate than glycerolipids. Sphingolipid content, and therefore lipid order, could increase with age and cataract indirectly due to increased lipid oxidation with age ^{11 12} and cataract. ^{11 13 14 15 16 17 18 19}  

Lipid structural changes could influence a number of factors in the lens. For instance, based on our binding studies, ^{38 39 40 41 42} elevated sphingolipid ^{17 29 30 31 32 33} and cholesterol levels and lipid hydrocarbon chain order, ^{7 8 9 12 34} as observed with age and cataract, would be expected to decrease theα -crystallin–lipid binding constant by 30% and increase the binding capacity of the membrane for α-crystallin by 200%. α-Crystallin could serve as a condensation point to which other crystallins bind and become oxidized. 

Elevated sphingolipid ^{17 29 30 31 32 33} and lipid hydrocarbon chain order, ^{7 8 9 12 34} as observed with age and cataract, have been shown to decrease Ca²⁺-adenosine triphosphatase (ATPase) activity in the lens ^{43 44} and in other systems. ^{45 46} Calcium homeostasis is essential to the clarity of the lens. ⁴⁷ In model liposome systems, elevated levels of sphingolipids and cholesterol and lipid hydrocarbon chain order have also been shown to elevate the levels of light scattering. ⁴⁸ Thus, both protein and lipid structural changes contribute to lens opacity. 

A nuclear cataract model using hyperbaric oxygen (HBO)-treated guinea pigs, developed by Giblin et al. ⁴⁹ and Padgaonkar et al., ⁵⁰ shows morphologic and biochemical changes in the lens nucleus similar to those found in the aging human lens and in immature human nuclear cataracts. Among many effects, HBO treatment causes nuclear opacity associated with increased distension of intracellular spaces, plasma membrane disruption, projection of processes from adjacent fibers, convoluted plasma membranes, increased levels of protein-thiol mixed disulfides, decreased levels of soluble proteins and the disulfide cross-linking of MIP26 and cytoskeletal proteins. ^{49 50} In this study, in the same HBO treatment model, products of lens membrane lipid oxidation and lipid structure were measured by infrared spectroscopy. 

Methods

Male retired breeder Hartley guinea pigs, initially 17 to 18 months old, were obtained from Kuiper Rabbit Ranch (Indianapolis, IN) or Kingstar (Kingston, NH). The animals were held for 1 to 2 weeks before HBO treatment to allow recovery from the stress of shipment and to identify the healthiest animals for the study. During this time the lenses of the guinea pigs were examined carefully by slit lamp biomicroscopy, and animals with cortical or nuclear opacities were excluded. The diet used throughout the study was special guinea pig chow (Guinea Pig Chow 5025; Purina Mills, Richmond, IN), which contained 0.1% ascorbic acid. Guinea pigs treated with HBO have been shown previously to grow at a normal rate. ⁴⁹ These studies adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

HBO Treatment

The HBO treatment protocol developed by Giblin et al. ⁴⁹ was used. Guinea pigs were treated in a pressure vessel 45 in. long and 18 in. in diameter (Amron, Escondido, CA). The vessel had a fully opening hinged door at one end and, at the opposite end, a 6-in. viewpoint for observing the animals during the experiments. Light from a 50-W tungsten halogen projector lamp, located outside the chamber, was led inside through an acrylic light pipe and was kept on during each treatment. Fourteen animals were treated at a time, seven in each of two Lucite boxes with screened tops. Each guinea pig was identified by a marking on the ear. Plastic trays containing wet paper towels were placed inside the chamber to add humidity. Soda lime (Sodasorb; WR Grace, Lexington, MA) was added to absorb CO₂, and ice was added to maintain the temperature below 23°C. After the chamber was sealed, it was flushed for 5 to 10 minutes with approximately 1 volume of 100% O₂ (USP Grade Medical Gas; Liquid Carbonic, Chicago, IL) which was vented outside the building. The pressure was then raised during 15 minutes to 2.5 atmospheres absolute (ATA; 22.3 psig [pounds per square inch gauge] or 50 ft of sea water) of O₂. At the end of a 2.5-hour holding period the pressure was released over a 15-minute period, to 1 ATA (0 psig), and the animals were removed. During the treatments, the guinea pigs were free to move around in the Lucite boxes. The guinea pigs were treated three times per week, on alternate days, at approximately the same time each day. The animals were treated either 30 times over a 10-week period or 50 to 51 times over a 17-week period. Age-matched control animals were included with each group of O₂-treated animals. The mean guinea pig ages ± SD at death (in days) were: 29.3 ± 0.5, 644 ± 55, 655 ± 14, and 657 ± 10 for the young untreated, older untreated, 30 HBO treatment and 50 HBO treatment groups, respectively. 

The transparency of lenses of control and HBO-treated guinea pigs was assessed by a single observer (L-RL) using a slit lamp photograph microscope (Carl Zeiss, Thornwood, NY) after induction of full mydriasis with tropicamide (1%) and phenylephrine (10%). The results were documented by photography. After animals were killed by CO₂ asphyxiation, the eyes were enucleated and the lenses were removed by posterior approach. Lenses, both control and experimental, averaged 115 mg wet weight. To make the best use of the animals killed, especially when they had been treated with O₂ over long periods, other tissues of the eye and other parts of the body were isolated, frozen, and used in additional studies. 

Lens Sectioning

For spectroscopic analysis, isolated lenses were frozen immediately in crushed dry ice and quickly placed in liquid nitrogen for storage until use. A 3-mm outside diameter thin-walled trephine was used to remove a core sample from the center of a frozen lens. Argon gas was directed over the lens which was placed in the center of a small petri dish. To prevent shattering of the lenses, they were warmed to approximately −20°C in a conventional freezer after they were removed from liquid nitrogen, and before taking core samples. The cylindrical core sample was divided into three sections using a surgical blade. The outer two thirds of the core is referred to as the cortical fraction. The central third of the core sample is referred to as the nucleus. The surrounding annulus of equatorial tissue was also saved for analysis. In most instances, tissue from four lenses was pooled. The cortical, nuclear, and equatorial fractions composed approximately 13%, 17%, and 70% of the total lens wet weight, respectively. 

Lipid Preparation for Spectroscopic Studies

The pooled samples described were placed in glass test tubes filled with argon gas. All reagents were bubbled with argon for 10 minutes before use. The samples were sonicated in methanol (2 ml) for 20 minutes in a bath-type sonicator, vortexed, and centrifuged at 7000 rpm. The supernatant was decanted, and the methanol was evaporated under a gentle stream of argon gas. The film was suspended in 2 ml of hexane-isopropanol (2:1) sonicated, vortexed, and centrifuged as before. The clear supernatants were decanted, and the hexane-isopropanol was evaporated under a stream of argon gas. The thin lipid film on the bottom of the tube was solubilized in 300 μl methanol to be used for spectroscopic analysis. 

FTIR Measurement of Lipid Oxidation

For Fourier-transform infrared (FTIR) analysis, each lipid sample prepared as described was layered onto an AgCl window and lyophilized for 12 hours to remove MeOH and trace amounts of water. Infrared spectra of the dried lipid films were measured to quantify lipid oxidation as was done for human lens membranes. ¹² Infrared spectra were acquired with a spectrophotometer (model 500 Magna IR; Nicolet, Freemont, CA). Exactly 300 interferograms were recorded, coadded, and apodized with a Happ–Genzel function before Fourier transformation, yielding an effective spectral resolution of 1.0 cm⁻¹. 

Analysis of Marker Bands

The CH stretching bands between 3100 and 2800 cm⁻¹, which do not change with oxidation, were used as an internal standard. The areas of the bands that change with oxidation (shown later) were divided by the area of the CH stretching bands to determine the relative increase in oxidation. 

Analysis of Hydroxyl and Hydroperoxyl Bands

The CH and OH infrared stretching region for a dried film of control guinea pig lens nuclear lipid is shown in Figure 1A . The intensity of the OH stretching bands (3600–3100 cm⁻¹) reflects the degree of lipid oxidation ⁵¹ and the amount of hydroxyl-containing lipids such as sphingolipids and cholesterol. To quantify the changes in the OH band intensity, the total areas of the OH and CH stretching bands were measured using an integration computer program (Grams 386 software, ver. 2.04; Galactic, Salem, NH). The baselines for the OH stretching band was taken near 3600, 3050 cm⁻¹ and the CH stretching band near 3050, 2750 cm⁻¹. 

The band intensity at 3444 cm⁻¹ is sensitive to changes in the number of lipid hydroperoxyl groups formed by oxidation. ⁵² To quantify the amount of lipid hydroperoxyl groups, the number and position of the major bands that compose the OH stretching region must first be determined using Fourier self-deconvolution and second-derivative analysis. When using Fourier self-deconvolution to determine the number of bands, artifacts due to overdeconvolution can be avoided by setting the smoothing and gamma parameters so that artifactual bands are not visible in the flat baseline region on either side of the OH and CH stretching bands. Second-derivative analysis should be used to confirm the number of peaks identified by Fourier self-deconvolution. Note that two major and two minor bands were detected in our samples (Fig. 1A) . The areas of each of the minor bands were determined using the curve fit algorithm. We curve fit the original spectrum. 

Analysis of cis Double Bonds

The cis double-bond band is located at 3010 cm⁻¹ (Fig. 1A) . When lipids are oxidized, cis double bonds of the hydrophobic chains rearrange to form trans double bonds (see Fig. 1 of reference 12). 

Analysis of Carbonyl Bands

The carbonyl stretching band near 1734 cm⁻¹ (Fig. 1B) arises from the acyl-linked hydrocarbon chains of lipids with a glycerol backbone, such as phosphatidylcholine or phosphatidylethanolamine, and from products of lipid oxidation. ¹²  

Analysis of Aldehyde Bands

In our guinea pig lens sample, at least four major bands were detected in the amide band region near 1670, 1620, 1600, and 1550 cm⁻¹ (Fig. 1B) . The assignments for these bands are: sphingolipid amide I mode and trans double bonds for the 1670-cm⁻¹ band, lipid aldehyde for the 1620- and 1600-cm⁻¹ bands, and the sphingolipid amide II stretching mode for the 1550-cm⁻¹ band. ⁵³ The area of the entire amide band region was measured and with the use of Fourier self-deconvolution and second-derivative analysis the total number of underlying bands was determined. A curve-fitting algorithm was used to measure the area of each underlying band such as the aldehyde band. 

Ordered-to-Disordered Phase Transition Measurement

To measure lipid phase transitions, aqueous HEPES buffer (5 mM; pH 7.4) containing 100 mM KCl was added to the window containing a film of dried lipid. A second window was placed on top of the first so that the sample was sandwiched within a 0.01-mm space. Infrared spectra were acquired with a spectrophotometer (model 500 Magna II; Nicolet). Exactly 3000 interferograms were recorded, coadded, and apodized with a Happ–Genzel function before Fourier transformation, yielding an effective spectral resolution of 1.0 cm⁻¹. Temperature was monitored to within ± 0.4°C by a remote sensor (Neslab, Portsmouth, NH) and maintained in a variable temperature cell (model 21500; Specac, Fairfield, CT). The sample was warmed to 60°C for 1 hour to ensure complete hydration and was then brought to 0°C over a period of 3 hours and allowed to equilibrate for 30 minutes. Temperature was raised at a rate of 0.1°C per minute and spectra measured using the following protocol to measure infrared spectra from 0°C to 100°C. The sample was allowed to equilibrate for 15 minutes every 5°C. After temperature equilibration, an infrared spectrum was measured for approximately 10 minutes. Signal averaging, data smoothing using the Savitsky–Golay procedure, baseline correction, Fourier self-deconvolution, and curve fit analysis were performed by computer (Grams 386 software, ver. 2.04; Galactic). 

Curve Fitting of Temperature-Dependent Infrared Data

Nonlinear regression analysis using a scientific graphing system (Sigma Plot version 4.02; Jandel Scientific, San Rafael, CA) was used to fit the lipid phase transition curves to the following equation as described by Borchman et al. ⁸ :  

\[\mathrm{Frequency}{=}P_{1}{+}{\{}P_{2}/{[}(1{+}P_{3}/T)^{P_{4}}{]}{\}}\]

where P ₁ is the minimum frequency of the CH₂ symmetric stretching vibration in wave numbers for the phase transition and represents the most ordered state of the transition. P ₂, the magnitude of the phase transition, is the net change in wave numbers and is related to the change in the number of trans-to-gauche rotomers (see the Discussion section for details). P ₃ is the transition temperature and indicates the temperature at which half the lipid molecules have undergone a phase change. T is the absolute temperature. P ₄ is the relative cooperativity, which reflects the ability of one lipid to influence the structure of the adjacent lipids. This term is similar mathematically to the Hill coefficient, which was first used to demonstrate the cooperative nature of hemoglobin–oxygen binding. The broader the phase transition the smaller the value for P ₄. 

The percentage of disorder at 36°C was calculated by ⁸  

\[\mathrm{Percentage\ of\ disorder}{=}100{\times}({\nu}-2849\ \mathrm{cm}^{-1})/5.5\ \mathrm{cm}^{-1}\]

where ν represents the frequency (in wave numbers) interpolated from equation 1 . The wave number of the CH₂ symmetric stretching band and the number of gauche and trans rotomers in the hydrocarbon chain(s) used to measure lipid order are related, ²⁸ the apparent frequency being dependent on the summation of the underlying trans and gauche bands. 

Phospholipid Composition by ³¹P-NMR

Phospholipid quantification of guinea pig lens lipid extracts was performed by a ³¹P-nuclear magnetic resonance (NMR) method. ^{29 30} The pooled phospholipid extract was dissolved in 400 μl of deuterochloroform. An aliquot of 250 μl of an EDTA-based reagent (prepared as described previously ^{29 30} ) and using KOH as the counterion source, was added at least 15 minutes before spectral acquisition. The mixture was then shaken, and the aqueous phase was allowed to separate before data acquisition. A spectrometer (500AMX NMR Bruker; Billerica, MA), operating at 202.4 MHz, was used to acquire ³¹P-NMR spectral data. Other acquisition parameters were spectral width, 2032.5 Hz; resolution, 0.50 Hz; acquisition time, 1.0 seconds; pulse length, 10 μsec; dwell time, 246μ sec. The data treatment was performed by computer (WINNMR software; Bruker). The spectra were phase corrected, zero filled, base line corrected, and deconvolved. The percentage of each phospholipid was evaluated by integrating the peak area corresponding to each phospholipid and then calculating the ratio of each area to the sum of all the areas. Nine components were quantified (Table 1) . 

Statistics

Unless indicated, all values are expressed as means ± SEM. Probability values were calculated using the paired Student’s t-test. 

Results

Slit Lamp Photography

The effect of HBO treatment of guinea pigs on lens nuclear light scattering was evaluated using slit-lamp biomicroscopy. The level of nuclear light scattering increased after 30 HBO treatments of the animals (Fig. 2B ) compared with that present in the lenses of the age-matched control animals (Fig. 2A) . The scattering became more intense after 50 treatments (Fig. 2D) . In addition to the increase in light scattering, the overall size of the nucleus appeared to increase as a result of HBO treatment of the animals, in agreement with previously reported trends. ⁴⁹  

Infrared Spectroscopy of Guinea Pig Lens Lipid Extract Dispersion: Temperature-Induced CH₂ Stretch Frequency Shifts

The infrared spectrum corresponding to the CH and OH stretching regions for lipid extracted from the nuclear region of a normal 654-day-old guinea pig lens, is shown in Figure 1A . For all the samples analyzed, the noise was less than 1% of the signal. The frequency of the infrared CH₂ symmetric stretching band was used to measure changes in lipid hydrocarbon chain order (see the Methods section). The CH₂ symmetric stretching-band frequency for control guinea pig lens lipids from the nuclear, cortical, and equatorial regions, increased in general from approximately 2852 cm⁻¹ at 0°C to approximately 2854 cm⁻¹ at 100°C (Fig. 3A ). We interpreted this change to be a result of the conversion of lipid hydrocarbon chain trans rotomers to gauche rotomers with increasing temperature. ⁵⁴ The cooperativity, a measure of the broadness of the lipid phase transition and the ability of one lipid to influence the structure of adjacent lipids, was similar for lipids from guinea pig lenses (all regions) (17.7 ± 0.95; n = 10), compared with that previously measured for bovine lens lipids (12 ± 6). ⁵⁵  

The guinea pig lipid phase transition was approximately twice as broad as that from human lenses, which measured 24 and 39 for cortical and nuclear lipids, respectively. ⁸ We estimate that the phase transition temperature for guinea pig lens lipids is 47.6°C ± 1.0°C. Note that at all temperatures studied, the lipids from lens nuclear control animals exhibited a lower CH₂ symmetric stretching-band frequency than the cortical and equatorial lipids, which were similar (Fig. 3A) . Lower frequencies indicated fewer gauche rotomers and higher hydrocarbon chain order. The CH₂ symmetric stretching-band frequency measured at a given temperature for a given lens region was similar from lens to lens, as evidenced by the overlap of the data for two separate lens pairs shown in Figure 3C (open symbols) and Figure 3D (solid symbols). From the CH₂ symmetric stretching-band frequencies at 36°C, we calculate that the equatorial, cortical, and nuclear guinea pig lens lipids are 31%, 32%, and 43% ordered, respectively. 

The CH₂ symmetric stretching-band frequencies of the lens cortical and nuclear lipids were higher than control animals at all temperatures studied for guinea pigs exposed to 30 HBO treatments (Figs. 3C 3D , respectively). The difference in frequency represents a 33% decrease in lipid hydrocarbon chain order at 36°C as a result of HBO treatment. In contrast, the CH₂ symmetric stretching-band frequencies of the lens equatorial lipids were approximately the same for animals with 30 HBO treatments compared with those not treated (Fig. 3B) . 

Phospholipid Composition by ³¹P-NMR

The ³¹P-NMR spectrum of guinea pig lens lipid extract from the equatorial region of clear, 644-day-old guinea pigs is shown in Figure 4 . Nine phospholipid species were quantified for lipids extracted from the lens equatorial region (Table 1) of control and HBO-treated lenses (30 treatments). Only two to five species were detected for lens lipids extracted from the nuclear region of control and HBO-treated lenses. The sphingolipid to glycerolipid increased by 180% and 114% in the nuclear and equatorial fractions from clear lenses, respectively, and did not change significantly in any region after 30 HBO treatments. The sphingolipids included dihydrosphingomyelin (DHSM) and sphingomyelin (SM). Glycerolipids included all the phospholipids, excluding the sphingolipids. 

Regional Changes in Lipid Composition and Lipid Oxidation with Age and HBO Treatment

Infrared OH Stretching-Band Region.

The intensity of the OH stretching region (3600–3000 cm⁻¹), shown for guinea pig lens nuclear lipids in Figure 1A , reflects the degree of lipid oxidation ^{51 52} and the amount of hydroxyl-containing lipids such as sphingolipids ⁵³ and cholesterol. Anhydrous films were used to eliminate interference from water OH stretching bands. The OH stretching band could be resolved into two major and two minor bands using a curve fit algorithm (Fig. 1A) . The band intensity at 3440 cm⁻¹ is sensitive to changes in the number of lipid hydroperoxyl groups formed by oxidation. ⁵² The other band intensities are sensitive to lipid hydroxyl groups induced by oxidation. ⁵¹ A study of control lenses from guinea pigs averaging 644 days old showed that the intensity ratios of the OH and OOH to CH stretching bands were similar for the equatorial, cortical, and nuclear region lipids, averaging 0.79 ± 0.05 (n = 38) for the OH band and 0.22 ± 0.02 (n = 41) for the OOH band. The significant and greater than twofold increase in the OH intensity ratio between the 29- and 644-day-old control guinea pig lenses, nuclear and equatorial regions (Figs. 5 A and 6 A, respectively), may reflect an increase in cholesterol, sphingolipid and/or oxidation. A significant difference between the lipid hydroperoxyl band intensity between the 29-day- and 644-day-old control guinea pig lenses, nuclear and equatorial regions (Figs. 5B 6B , respectively), confirms that increased lipid oxidation occurs with age, concomitant with an increase in the percentage of sphingolipid (see last paragraph of the Results section). 

The relative intensity of the lens nuclear lipid OH stretching band increased significantly by 57% and 88%, after 30 and 50 HBO treatments, respectively, compared with the 644-day-old control (Fig. 5A) . Similarly, the OOH stretching band increased by 45% and 65% after 30 and 50 HBO treatments, respectively, compared with the 644-day-old control (Fig. 5B) but the increase after 30 treatments was not statistically significant (P > 0.10). In contrast to results for the lens nucleus after 30 HBO treatments, there was no significant change in the relative intensity of the equatorial lipid OH or OOH bands (Figs. 6A 6B , respectively). Only after 50 treatments were there significant increases for the OH and the OOH equatorial region band relative intensities, compared with the 0 and 30 treatment pools (Figs. 6A 6B) . The relative intensities of the OH and OOH bands for the lens cortical region gave relative standard deviations that were approximately twice those of the lens equatorial and nuclear region data, and the data from this region therefore will not be discussed. 

Infrared Amide Band Region.

The amide band region is composed of at least four major bands near 1670, 1620, 1600, and 1550 cm⁻¹ (Fig. 1B) . The assignments for these bands are: sphingolipid amide I mode and trans double bonds for the 1670 cm⁻¹ band, lipid aldehyde for the 1620- and 1600-cm⁻¹ bands, and the sphingolipid amide II for the 1550 cm⁻¹ band. ⁵³ A curve-fitting algorithm showed that the aldehyde bands at 1620 and 1600 cm⁻¹ make up more than 90% of the total band intensity of the region. There was no difference between the relative intensities of the amide band from lipids of the equatorial, cortical, and nuclear regions of a control guinea pig lens (644 days old); the average value was 0.44 ± 0.03 (n = 35). 

For the nuclear and equatorial regions of control lenses from 644-day-old control guinea pigs, the integrated area of the amide band region was significantly larger—by 140% and 80%, respectively—than in the 29-day-old control guinea pigs (Figs. 5C 6C , respectively). We interpret this difference as an increase in sphingolipid content with age. For the nuclear and equatorial regions of lenses from 644-day-old control guinea pigs, the total 1600 cm⁻¹ lipid aldehyde band intensity was significantly larger by 200% and 100%, respectively, compared with lenses from 29-day-old control guinea pigs (Figs. 5D and 6D , respectively). 

For the lens nuclear region of animals treated 30 and 50 times with HBO, the total amide bands increased significantly by 40% and 70%, respectively (Fig. 5C) . The 1600 cm⁻¹ lipid aldehyde band remained statistically the same (P = 0.3) for lens nuclei of animals treated 30 times with HBO, but increased significantly by 40% for animals treated 50 times with HBO (Fig. 5D) . In contrast to results for the nucleus, there was no significant change (P > 0.1) in the relative intensities of the lens equatorial lipid total amide or aldehyde bands after 30 HBO treatments (Figs. 6C 5D , respectively). Only after 50 treatments was there a significant increase in the total relative intensities of the two bands compared with the control animals with no treatment. 

Infrared cis C⋕C Stretching Band.

The cis double-bond band is located at 3010 cm⁻¹ (Fig. 1A) . The relative intensity of the cis C⋕C bond band for guinea pig lens lipid control samples (644-day-old mean) decreased regionally in the order: equatorial, cortex, nucleus (Table 2) . A similar trend was observed for the pool of lenses from 29-day-old guinea pigs (Table 2) . The relative intensity of the lens nuclear lipid infrared cis double-bond band from guinea pigs receiving 30 HBO treatments was 35% lower, 0.00036 ± 0.00008, n = 6, compared with control animals (control shown in Table 2 , 644-day average), but the difference was only marginally significant (P = 0.1). The relative intensity of the cis C⋕C band for lipids extracted from any of the three regions, was not significantly different (P > 0.1) between results for control animals and 50 HBO-treated guinea pigs (data not shown). 

Changes in Membrane Glycero- and Sphingolipid Content with Age and Region

The acyl linkage of the hydrocarbon chains for glycerophospholipids contributes predominantly to the intensity of the carbonyl band at approximately 1740 cm⁻¹ (Fig. 1B) . The relative intensity of the carbonyl band for guinea pig lens lipid control samples (644-day-old mean) decreased regionally in the order: equatorial, cortex, and nucleus (Table 2) . A similar trend was observed for the pool of lenses from 29-day-old guinea pigs (Table 2) . The greater percentage of glycerolipids in the 29-day-old guinea pig lenses, compared with the 644-day-old lenses, is evident by the larger relative intensity of the carbonyl band in the lipids of the 29-day-old lenses (Table 2) . A concomitant 80% and 156% increase in the total infrared amide band for control lipids (29 days versus 644 days old) from the equatorial (Fig. 6C) and nuclear regions (Fig. 5C) , respectively, indicates that the age-related distribution of the glycerolipids changes concomitantly with the sphingolipids. From this percentage of change data, the glycerolipid in the lens equatorial region from the 29-day-old guinea pig was calculated to be 74% and to decrease to 52% at 644 days old. The glycerolipid in the lens nuclear region from a 29-day-old guinea pig was calculated to be 70% and to decrease to 37% in the 644-day-old guinea pig lens. These estimates represent lower limits for the percentage of glycerolipid and upper limits for sphingolipid content, because oxidation, which contributes to the carbonyl and amide bands, increases with age and the contribution of the plasmalogen lipid, which does not contain an acyl or amide linkage, was not accounted for. The age-related changes were substantiated by the regional differences in the intensity of the carbonyl band intensities for a given age by comparing the carbonyl intensity in the nucleus with that of the equatorial region (Table 2) . The trend observed qualitatively from our infrared data are comparable with data measured quantitatively using ³¹P-NMR, in which, with increasing guinea pig age, lens sphingomyelin was found to increase by 165% and 224% in the equatorial and nuclear regions, respectively, and glycerolipids such as phosphatidylcholine were found to decrease (Table 1) . Free fatty acids released from oxidized phospholipids may cause the intensity of the carbonyl band to increase; however, only a marginally statistical (P = 0.04) 27% increase in this band was evident in the lens nucleus of the guinea pig after 30 HBO treatments 0.043 ± 0.003 (n = 7) compared with the control (0.034 ± 0.003; n = 16).  

Discussion

Impact of Oxidation on Light Scattering

HBO treatment of guinea pigs resulted in light scattering in the lens nuclei after as few as 30 treatments and became more intense after 50 treatments (Fig. 2) . This increased light scatter was previously reported to be coupled with lens membrane structural invaginations, loss of glutathione, and increased protein disulfide bonds in the nuclear region. ⁵⁰ In contrast, the cortical region of the lens remained clear after in vivo HBO treatment, with none of these alterations detected. The major finding of this study was that after 30 treatments of the animals with HBO, lipid disorder and products of lipid oxidation increased in the lens nucleus (Figs. 3D 5A 5B 5C 5D , respectively), but not in the equatorial region (Figs. 3B 6A 6B 6C 6D) , compared with age-matched control animals. Thus, the appearance of lipid alterations in the lens nucleus after 30 HBO treatments corresponded with increased light scattering. A correlation between lens opacity and lipid structural alterations has been made for human age-related cataractous lenses. ^{7 8} Oxidation of lens nuclear lipids may also have been associated with the development of nuclear cataracts in humans who were treated therapeutically with HBO. ⁵⁶  

The reason only lens nuclear and not equatorial lipids were affected by HBO after 30 treatments of the animals could not be attributed to differences in the composition of the nuclear lipids that might predispose them to be more susceptible to oxidation. The results of this study indicate that indeed the opposite was true. We found that the lens nuclear lipids contained a lower level of glycerolipids (all but DHSM and SM; Table 1 ) and were 58% more saturated (less cis C⋕C bonds) than lipids present in the equatorial membranes (Table 2) . Both of these factors would act to decrease the inherent susceptibility of the lens nuclear lipids to oxidation, compared with lipids present in the equatorial region. It is likely, as has been suggested previously, ⁴⁹ that because the nucleus contains considerably less antioxidant activity compared with the cortex, this region is significantly more susceptible to oxidation. For example, the five times lower level of GSH present in the nuclear region of the guinea pig lens compared with that in the cortex, ⁴⁹ would make the lipids in the central region more prone to oxidation. 

After 50 treatments of the animals with HBO, increased levels of lipid oxidation over control values were observed in both the equatorial region of the lens (Figs. 6A 6B 6C 6D) , as well as in the nucleus (Figs. 5A 5B 5C 5D) . A possible explanation for why lipid oxidation was detected in the equatorial region without a coincident increase in light scattering (Fig. 2) or protein oxidation in that region ^{49 50} may be linked to the observed increase in the size of the lens nucleus after 50 HBO treatments. ⁴⁹ Because the diameter of the core of the nuclear region sample was kept the same throughout the study, it is possible that the equatorial region sample of the 50 HBO treatment lens contained a portion of the enlarged nucleus. This possibility is made more likely by the fact that whereas in our previous studies, the nucleus comprised 25% of the total lens weight, ^{49 50} the nuclear component in the present investigation made up only 17% of the total. Thus, for the analysis of lipids in this study, some of the enlarged nuclear component may have been sectioned along with the equatorial region. 

Impact of Sphingolipid Content on Lipid Order

Using techniques similar to those developed from human lens studies ⁸ we used infrared CH symmetric stretching-band frequencies at 36°C (Fig. 3A) to determine the lens lipid hydrocarbon chain order. Equatorial, cortical, and nuclear lipids from 644-day-old guinea pigs were found to be 31%, 32%, and 42% ordered, respectively. This degree of lipid hydrocarbon chain order fits well with previous studies for a variety of tissues ²⁸ showing that lipid order decreases with phosphatidylcholine content and increases with sphingolipid content (Fig. 7) . The reason for this correlation is that sphingolipids are highly saturated, and saturated lipids enhance van der Waal’s interactions, resulting in a higher phase transition temperature and thus a higher order at 36°C. ²⁸ In this study, guinea pig lens lipid saturation was found to correlate with lipid order. In the control 644-day-old guinea pig lens, lipid saturation increasedfrom the equatorial region of the lens, to the cortex, to the nucleus, as is evident from the decrease in relative intensity of the cis double-bond bands (Table 2) . A similar trend was found in the lenses of 29-day-old guinea pigs (Table 2) . 

With increasing age, the sphingomyelin content in guinea pig lens was found to increase by 165% and 224% in the equatorial and nuclear regions, respectively, and glycerolipids such as phosphatidylcholine were found to decrease (Table 1) . These results are similar to those observed in aging human lens. ^{29 30} The increase in sphingolipid content with age undoubtedly contributes ²⁸ to the increase in lipid order observed with increasing age in the human lens ³⁴ which could influence lens membrane functional properties and the clarity of the lens (see introduction). Because there is no evidence for lipid synthesis in the nuclear region of the lens, the elevation in the percentage of sphingolipid in the human and guinea pig lens with age must result from degradation of other lipids such as the glycerolipids; however, there have been no studies to support this hypothesis. Glycerolipids would be expected to be preferentially oxidized because glycerolipid hydrocarbon chains have more double bonds. ²⁸ We hypothesize that lipases eliminate oxidized glycerolipids, leaving a membrane composed of more saturated sphingolipids. 

Impact of Oxidation on Lipid Order

After 30 HBO treatments of the animals, lipid order was found to decrease in the lens nucleus (Fig. 3D) , to decrease to a lesser extent in the cortex (Fig. 3C) , and to remain essentially unchanged in the equatorial region (Fig. 3B) . The observed decreased lipid order in the nucleus was similar to our previously reported data on oxidation-induced decrease in lipid order in studies of lipid from rabbit ³⁵ and bovine ³⁶ lenses and from purified sphingomyelin membranes. ^{27 37} However, this finding is contrary to other results showing that oxidation actually increases the order of certain disordered membranes that contain a low content of sphingolipid. ^{57 58 59 60 61 62} We may conclude from these structural studies that although lipid order increases in the human lens with age ³⁴ and cataract, ^{7 8 9} this increase is the result of an elevation in sphingolipid content (Fig. 7) and is not due to oxidation, which in fact would result in the disorder of the membrane. In the guinea pig lens of the HBO model, but not in the aging human lens, the ordering effect of sphingolipids and oxidation-induced trans double bonds is overcome by the disordering effect of oxidation-induced lipid hydroxyl groups. Hydrophylic OH groups added to the hydrophobic region of lipid hydrocarbon chains with oxidation would be expected to disrupt van der Waal’s interactions between adjacent hydrocarbon chains, thus disordering the membrane. In the present study, a decrease in lipid order in the lens nucleus (Fig. 3D) was associated with an increase in oxidation (Figs. 5A 5B 5C 5D) in agreement with observations made for regions containing small focal opacities in fixed human lenses. ⁶³ The equatorial region of the guinea pig lens, which showed no change in lipid order after 30 HBO treatments (Fig. 3B) , also exhibited no effects of oxidation during this period (Figs. 6A 6B 6C 6D) . 

Correlations with Human Lens Regional Age and Cataract Studies

In the present study, we used the same spectroscopic technique to study guinea pig lens lipid oxidation that we had used previously to investigate this parameter in the human lens as a function of age. ¹² Products of lipid oxidation were found to increase with age in clear lenses from both the nuclear and cortical regions of the guinea pigs and humans (Table 3) . We found that in the lens nucleus of the HBO-treated animal, lipid hydroxyls (Fig. 5A) , hydroperoxyls (Fig. 5B) , and aldehydes (Fig. 5C) increased after both 30 and 50 treatments, compared with control animals. The measured level of oxidation was substantial and comparable to that observed as a human lens ages from 20 to 80 years old. ¹² The introduction of hydrophilic hydroxyl and hydroperoxyl groups into lipids is likely to cause membranes to become more permeable to cations, ⁶⁴ and the introduction of the same groups into the hydrophobic hydrocarbon regions of lipids would be expected to decrease the order of this region, as was observed. The appearance of an increase in light scattering in the lens of the HBO-treated guinea pig (Fig. 2) coincided with the appearance of products of lipid oxidation (Fig. 5) and structural changes (Fig. 3D) in the nucleus. A similar correlation between products of lipid oxidation and lens opacity has been made in human cataractous lenses (Table 3) . ^{6 20} No products of lipid oxidation (Figs. 6A 6B 6C 6D) , changes in structure (Fig. 3B) , or opacity (Fig. 2) were observed in the lens equatorial region after 30 HBO treatments. The overall ordering seen in membranes of human cataractous tissues (Table 3) ^{7 8 9} suggests that oxidation may target unsaturated glycerolipids; these types of lipids are predominant in guinea pig, but not in human lenses. We postulate that these lipid compositional differences are responsible for the disordering that is associated with lens opacity in guinea pigs and the ordering that is observed in human cataractous lenses. 

Morphologically, after in vivo HBO treatment, the nuclear region of the guinea pig lens showed distension of intercellular spaces at the Y junctions between neighboring fiber cells, as well as cell membrane breakage and convoluted membranes. ⁴⁹ It can be speculated that lipid oxidation would contribute to the breakage and convolution of the lens nuclear membranes, as was observed. Similar membrane fragmentation, vesicle formation, and disruption of cell-to-cell contacts have been attributed to products of lipid oxidation in muscle sarcoplasmic reticulum and retinal rod and pigment epithelial cells. ⁶¹ In the HBO guinea pig cataract model, lens nuclear MIP26 and cytoskeletal proteins were found to be disulfide cross linked, and a portion of nuclear MIP26 was degraded to MIP22. ⁵⁰ Because oxygen is five times more soluble in the lipid membrane ⁶⁵ than in the cytoplasm, it is reasonable that oxidation may begin at the site of the membrane, as has been observed in human lenses. ¹⁰ However, the species of oxygen that may be responsible for the lipid oxidation remains speculative. Products of lipid oxidation, similar to those that were observed in this study have been identified in lens membranes after generation of hydroxyl radical through an ascorbate-iron system. ³⁶ Singlet oxygen produced from molecular oxygen photosensitizers ^{15 66} may also contribute to lipid oxidation. 

Previous studies have shown that HBO in vivo accelerates aging in the nuclear region of the guinea pig lens with regard to loss of water-soluble and cytoskeletal proteins, damage to plasma membranes, formation of protein disulfide, and degradation of MIP26. ^{49 50} Such modifications are similar to those that occur in the nuclei of aging and cataractous human lenses. ⁶⁶ Although the causes of the changes in lipid composition with age may or may not be different from those brought about by HBO treatment, a similar acceleration of aging in the lens nucleus was observed in this study, as evidenced by an increase in the relative intensities of lipid hydroxyl (Fig. 5A) , hydroperoxyl (Fig. 5B) , amide (Fig. 5C) , and aldehyde (Fig. 5D) bands, and a decrease in the number of lipid cis double bonds (Table 2) , with both age and HBO treatment. In the human lens, the levels of oxidized lipids ¹² have been reported to increase with age and cataract. It is possible that a threshold level of lens membrane oxidation exists, above which membrane disruption and lens opacity results. The guinea pig HBO model for nuclear cataract formation appears to be useful for studying this hypothesis, because subtle oxidatively induced changes in lens nuclear membrane lipids and proteins can be examined as a function of loss of transparency in the nucleus. 

 Supported by Public Health Service Grants EYO7975 (DB), EYO11657 (MCY), EYO0484 (VNR), EYO2027 (FJG), and EYO5230 (Core Center Grant); the Kentucky Lions Eye Foundation, Louisville; and an unrestricted grant from Research to Prevent Blindness, New York. The study was part of the Cooperative Cataract Research Group (CCRG) program.

 Submitted for publication January 31, 2000; revised April 18, 2000; accepted April 26, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

(A) Infrared OH and CH stretching region for a thin anhydrous film of lipids from the nuclear region of a pair of 654- and 661-day-old guinea pig lenses. Lower bands in the OH stretching region indicate the results of a curve-fitting algorithm showing the two major and two minor peaks (arrows) that compose the OH stretching region. The CH₂ symmetric stretching band (CH₂ sym.) was used in hydrated samples to measure lipid hydrocarbon chain structural order (see Fig. 3 ). The C⋕C cis band is indicative of lipid unsaturation. (B) Infrared fingerprint region for a thin anhydrous film from the nuclear region of a pair of 654- and 661-day-old guinea pig lenses. The carbonyl, amide, and trans C⋕C bands may be used as markers for lipid oxidation. The carbonyl band is also a marker of glycerolipids and the amide band a marker for sphingolipids. The phosphate stretching bands (PO⁻ ₂ asym. and sym.) are markers for phospholipids. Other bands representative of lipid hydrocarbon chain stretching modes are the terminal CH₃ stretching band and the CH₂ rocking band.

Table 1.

 

Phospholipid Composition by ³¹P-NMR

Phospholipid	Phospholipid Composition (%)
		29-Day-Old	644-Day-Old	655-Day-Old 30 HBO Treatments
Equatorial region
PG	ND	0.3 ± 0.2	5 ± 2
DHSM	2 ± 1	2 ± 0.5	ND
PE plas	25 ± 2	14 ± 1	ND
PE	9 ± 1	8 ± 0.5	4 ± 2
PS	5 ± 2	4 ± 0.5	ND
SM	11 ± 1	30 ± 1	30 ± 3
LPC	0.5 ± 0.4	0.3 ± 0.1	5 ± 2
AAPC	1 ± 0.5	2 ± 0.5	3 ± 2
PC	47 ± 2	40 ± 1	52 ± 3
SM/glycerolipid	0.18	0.38	0.31
Nuclear region
PE plas	25 ± 4	15 ± 5	<5
PE	7 ± 2	<5	<5
PS	1 ± 0.5	<5	<5
SM	13 ± 2	42 ± 5	55 ± 5
PC	53 ± 5	43 ± 5	49 ± 5
SM/glycerolipid	0.18	0.49	0.51

 Glycerolipids were defined as all phospholipids excluding SM and DHSM. Data are mean percentages. ND, not detectable; PG, phosphatidylglycerol; DHSM, dihydrosphingomyelin; SM, sphingomyelin; PE plas, phosphatidylethanolamine plasmologen; PE, phosphatidylethanolamine; PS, phosphatidylserine; LPC, lysophosphatidylcholine; AAPC, alkylacylphosphatidylcholine; PC, phosphatidylcholine.

Figure 2.

 

Slit lamp photographs of guinea pig eyes. (A) A 20.5-month-old control lens; (B) 20.5-month-old lens after 30 treatments with HBO over a 10-week period; (C) 22.25-month-old control lens; and (D) 22.25-month-old lens after 51 treatments with HBO over a 17-week period.

Figure 3.

 

Effect of 30 HBO treatments on the change in the CH₂ symmetric stretching-band frequency versus temperature for hydrated samples of guinea pig lens lipids. (A) Control animals (644 days old) all regions: (•), nuclear region; (○), cortical region; and (▵), equatorial region. (B) Equatorial region: (•), control; and (○), HBO-treated animals. (C) Cortical region: (•), control animals; (○) and (▵), HBO-treated animals, two separate lens pairs. (D) Nuclear region (•) and (▴), control animals, two separate lens pairs; and (○), HBO-treated. The increase in the CH₂ symmetric stretching-band frequency was interpreted as an increase in the structural disorder of the lipid hydrocarbon chains as a result of a change from trans to gauche rotomers. A frequency of 2849 cm⁻¹ indicates a completely ordered hydrocarbon chain. A frequency of 2854.5 cm⁻¹ indicates a completely disordered hydrocarbon chain. Each symbol is for lipid extracted from a pool of paired lenses.

Figure 4.

 

31P-NMR spectrum of a pool of equatorial lens lipids extracted from guinea pigs averaging 544 days old. PG, phosphatidylglycerol; DHSM, dihydrosphingomyelin; SM, sphingomyelin; PE plas, phosphatidylethanolamine plasmalogen; PE, phosphatidylethanolamine; PS, phosphatidylserine; LPC, lysophosphatidylcholine; AAPC, alkylacylphosphatidylcholine; PC, phosphatidylcholine.

Figure 5.

 

Effect of age and in vivo HBO on guinea pig lens nucleus infrared spectral parameters used to assess lipid oxidation. The y-axes of (A) and (B) give lens nuclear region products associated with lipid oxidation measured from thin anhydrous films. The products are ratioed to the CH symmetric stretching-band intensity to quantitate the amount of oxidation relative to the amount of lipid. (C) Entire amide region. (D) Intensity of the 1600-cm⁻¹ band due to aldehydes that were calculated from a curve-fitting algorithm. Thirty and 50 HBO treatments correspond to 10 and 17 weeks of treatment, respectively. Values are ± SEM with the number of samples in parentheses.

Figure 6.

 

Description of data is as in Figure 5 but for the equatorial region of the guinea pig lens.

Table 2.

 

Relative Intensities for cis C⋕C Stretching and Carbonyl Bands for Lipids of Three Regions of Young and Old Guinea Pig Lenses

	29-Day-Old	644-Day-Old
Stretching bands (I cisC⋕C/ICH)
Equatorial	0.00517 ± 0.00006 (4)	0.00304 ± 0.00022 (16)
Cortical	NA	0.00086 ± 0.00022 (7)
Nuclear	0.0028 ± 0.0008 (4)	0.000554 ± 0.000008 (15)
Carbonyl bands (IC⋕O/ICH)
Equatorial	0.149 ± 0.002 (4)	0.103 ± 0.007 (15)
Cortical	NA	0.073 ± 0.006 (7)
Nuclear	0.106 ± 0.022 (4)	0.034 ± 0.003 (16)

 Data are means ± SEM. The number of samples analyzed is in parentheses. NA, not analyzed.

Figure 7.

 

Correlation between sphingolipid percentage of total phospholipid and hydrocarbon chain order at 36°C. Lipid order as used in this figure was calculated directly from the CH₂ symmetric band frequency. A 100% order on the y-axis corresponds to a CH₂ symmetric band frequency of 2849 cm⁻¹, whereas 0% order corresponds to a CH₂ symmetric band frequency of 2854.5 cm⁻¹. It is likely that this order parameter relates to the number of C⋕C trans rotomers. Solid line is the linear regression line through the data. The following data are from our prior studies: sarcoplasmic reticulum (SR) data, ²⁸ human lens lipid order, ³⁴ human lens lipid composition, ²⁹ and bovine lens lipid order ⁵⁵ . Human lens lipid order ranged from 44% to 66%, with the average plotted. Bovine lens lipid composition is from Broekhuyse, ⁶⁸ rod outer segment (ROS) disc and plasma membrane composition from Boesze–Battaglia and Albert, ⁶⁹ and order from Lamba et al. ⁷⁰ Guinea pig lens lipid composition and order are from the current study, and rabbit lens lipid composition is from Iwata et al. ⁷¹ Rabbit lens lipid order is from the heating curve of Sato et al. ³⁸

Table 3.

 

Correlations between Guinea Pig Lens Lipid Data Measured in This Study and Human Lens Membranes Measured in Other Studies

	Age and Regional Correlations	Correlations Regarding Nuclear Cataracts
Changes in sphingolipid membrane content	Elevated with age in both human 29 30 31 and guinea pig lenses in all regions	Elevated in human cataracts. 17 31 32 33 No change in the guinea pig HBO-induced nuclear cataract model.
Changes in glycerolipid membrane content	Diminished with age 29 30 31 in both human and guinea pig lenses in all regions	Diminished in human cataracts. 17 31 32 33 No change in the guinea pig HBO-induced nuclear cataract model.
Sphingolipid (SL) content	Higher in human lenses 29 30 31 ; nuclear and cortical regions have the same SL content in contrast to the guinea pig lens nuclear region, which has a higher SL content than the cortical region	Much higher in human cataractous lenses compared with the guinea pig HBO-induced nuclear cataract model.
Glycerolipid (GL) content	Lower in human lenses 29 30 31 ; nuclear and cortical regions have the same GL content in contrast to the guinea pig lens nuclear region, which has a lower GL content than the cortical region	Much lower in human cataractous lenses compared with the guinea pig HBO-induced nuclear cataract model.
Lipid hydrocarbon chain structural order (fluidity) at 36°C	Higher in human lenses 12 34 ; similar in human lens nuclear and cortical regions. 12 The guinea pig lens nuclear region has a higher hydrocarbon chain order than that of the cortical region	Compared with clear regions of the lens, lipid order is higher in opaque regions of human cataractous lenses 7 8 9 and lower in opaque regions of guinea pig lenses treated with HBO.
Change in lipid oxidation	Elevated with age 11 12 in all regions in both human and guinea pig lenses. No regional differences were detected in both systems	Correlates with lens opacity in both human cataracts 11 13 14 15 16 17 18 19 and in the guinea pig HBO-induced nuclear cataract model.

The authors thank Ann Dunlop for the long-term care of the guinea pigs and the following students for assistance in treating the animals with HBO: Alex Hung, Cristina Kapustij, Whitney Lakin, Olivia Luther, and Vaidehee Padgaonkar.