Graefe's Archive for Clinical and Experimental Ophthalmology

, Volume 242, Issue 12, pp 1000–1007

High-resolution magic angle spinning 1H NMR spectroscopy of metabolic changes in rabbit lens after treatment with dexamethasone combined with UVB exposure


    • Faculty of MedicineNorwegian University of Science and Technology (NTNU)
    • Faculty of Natural Sciences and TechnologyNorwegian University of Science and Technology (NTNU)
    • MR-SenteretSt Olavs Hospital
  • Øystein Risa
    • Faculty of MedicineNorwegian University of Science and Technology (NTNU)
    • Faculty of Natural Sciences and TechnologyNorwegian University of Science and Technology (NTNU)
  • Jitka Čejková
    • Department of Eye Histochemistry and Pharmacology, Institute of Experimental MedicineAcademy of Sciences of the Czech Republic
  • Jostein Krane
    • Faculty of Natural Sciences and TechnologyNorwegian University of Science and Technology (NTNU)
  • Anna Midelfart
    • Faculty of MedicineNorwegian University of Science and Technology (NTNU)
Laboratory Investigation

DOI: 10.1007/s00417-004-1030-8

Cite this article as:
Sæther, O., Risa, Ø., Čejková, J. et al. Graefe's Arch Clin Exp Ophthalmol (2004) 242: 1000. doi:10.1007/s00417-004-1030-8



Long-term steroid treatment and UVB exposure are well-known cataractogenic factors. The purpose of this study was to investigate metabolic changes in the rabbit lens after long-term dexamethasone treatment in combination with UVB exposure, using high-resolution magic angle spinning proton nuclear magnetic resonance (HR-MAS 1H NMR) spectroscopy to analyse intact lens tissues.


Rabbits received topical doses of 0.1% dexamethasone or 0.9% saline (50 μl) four times daily for 36 days. On day 37, the eyes were exposed to UVB radiation (2.05 J/cm2). Twenty-four hours later the animals were killed, and HR-MAS 1H NMR spectra of lens tissues were obtained.


More than 15 major metabolites were assigned in NMR spectra of rabbit lenses. The combined treatment with dexamethasone and UVB induced large reductions in the concentration of reduced glutathione, inositols, taurine and lactate compared with normal lenses. Concurrently, the levels of glucose, sorbitol and sorbitol-3-phosphate were increased. After exposure to UVB radiation only, the most significant finding was a decrease in the concentration of lactate. No lens opacities were detected.


HR-MAS 1H NMR spectroscopy was found to be an efficient tool for analysis of intact lens tissues. High-resolution NMR spectra of intact lens tissue enabled metabolic changes to be quantified. Long-term treatment with dexamethasone combined with UVB exposure induced substantial metabolic changes, dominated by osmolytic regulation processes and loss of glutathione.


Cataract formation is a well-established side effect of long-term treatment with glucocorticoids [41]. However, the exact mechanism of steroid-induced cataract is not known [15]. Biochemical changes such as the occurrence of glucocorticoid-protein adducts [2] and loss of glutathione [42] have been observed within the lens in animal experiments. From human lens protein culture studies, formation of disulfide-linked protein aggregates has been reported following steroid treatment [3].

Cataract is a multifactorial process, so it may not be sufficient to test the lenticular metabolic effect of a drug alone. Instead, different cataractogenic factors should be combined [44] due to their additive or even synergistic effects [41]. Among others, ultraviolet B (UVB) radiation is a risk factor for the development of cataract [45]. In the rabbit lens, cataract develops following relatively low-level UVB irradiation (1–2 mW/cm2) [18]. Thus, exposure to UVB radiation might enhance development of cataract induced by steroids.

By the application of nuclear magnetic resonance (NMR) spectroscopy on biological samples, numerous metabolites can be monitored simultaneously. Metabolic changes in the rabbit lens after treatment with steroids like dexamethasone have previously been investigated by NMR spectroscopy, i.e. organophosphate profiling of incubated lenses and their extracts [12] and analysis of lens extracts after long-term topical treatment [33]. In our laboratory, 19F and 1H NMR spectroscopy have been applied to monitor penetration of dexamethasone into the anterior segment of the eye after short-term topical treatment, analysing tissue extracts from cornea and lens [29]. However, tissue extraction procedures involve extensive treatment of the samples, which might change their biochemical composition. Also, a relatively large amount of tissue is required. Thus, analysis of intact tissue samples is preferable. Such opportunity is provided by the application of high-resolution magic angle spinning (HR-MAS) NMR spectroscopy. Using this technique, peak broadening effects caused by anisotropic interactions are averaged to zero by fast spinning of the sample at a certain angle (54.7°) between the spinning axis and the static magnetic field. As shown previously with other tissues [1, 4, 5, 30], spectral resolution comparable with that obtained from liquid extracts can be achieved by means of HR-MAS 1H NMR spectroscopy.

The purpose of this study was to investigate metabolic changes in the rabbit lens after combined treatment with dexamethasone and UVB exposure, using HR-MAS 1H NMR spectroscopy to analyse intact lens tissues. The feasibility of HR-MAS 1H NMR spectroscopy of intact lens tissue will be discussed.

Materials and methods

Animal experiments

Albino rabbits (New Zealand white, body weight 2.5–3.0 kg) were divided into three groups. In the first group both eyes of four rabbits were topically treated with 0.1% dexamethasone (Decadron; Merck, Sharp & Dohme, Whitehouse Station, NJ, USA) before exposure to UVB radiation (dxm/UVB group). In the second group three rabbits were topically treated with 0.9% saline instilled in both eyes before UVB exposure (UVB group). The third group (three rabbits) served as controls (untreated animals).

The eye drops were administered by application of 50 μl of either drug or saline solution into the lower conjunctival fornix four times daily for 36 days. On day 37, all animals in the dxm/UVB and the UVB group were anaesthetised by i.m. injection of Rometar (xylazine hydrochloride 2%; Spofa, Czech Republic) [(0.2 ml/kg body weight)] and Narkamon (ketamine hydrochloride 5%; Spofa) [(1 ml/kg body weight)] before exposure to UVB radiation. The eyes were open and the cornea was irradiated by UVB rays using an UVB lamp (Bioblock Scientific, Illkirch, France; wavelength 312 nm, 6 W) from a distance of 0.03 m for 10 min (the time of exposure), while the rest of the eye was protected. An intensity of approximately 3.2 mW/cm2 was achieved at this distance during the experiment. The total dose of irradiation was 2.05 J/cm2. The intensity and dose were measured with a VLX-3 W radiometer with microprocessor from Cole–Parmer (Vernon Hills, IL, USA) and a Cole–Parmer UVB sensor (312 nm). Both eyes were treated in the same way. The lenses were examined with a hand-held slit lamp during the treatment with steroids and immediately after the irradiation procedure. Twenty-four hours after the UVB exposure, the animals were killed using thiopental anaesthesia and lens samples were prepared from each eye after enucleation (the lens sample from one eye in the dxm/UVB group was lost during preparation). The samples were immediately frozen and stored at −80°C prior to analysis by NMR spectroscopy. Animal experiments were performed in compliance with the “Principles of laboratory animal care” (NIH publication no. 85-23, revised 1985) and the OPRR Public Health Service Policy on the Humane Care and Use of Laboratory Animals.

NMR spectroscopy

The lenses were prepared for HR-MAS 1H NMR spectroscopy. A cut was made along the lens axis (anterior–posterior), dividing the lens into two parts. A small meridional sector of one of the halves was obtained and weighed (20–40 mg); this sector contained the whole radial profile with both cortical and nuclear parts of the lens. While still frozen, the lens sample was immersed in D2O in a zirconium 4-mm 50-μl HR-MAS rotor. TSP (sodium-3′-trimethylsilylpropionate-2,2,3,3-d4) was used as an internal shift reference substance (0 ppm). HR-MAS 1H NMR spectroscopy was performed on a Bruker Avance DRX 600 spectrometer (14.1 T; Bruker BioSpin, Germany) equipped with a 4-mm HR-MAS 1H/13C probe. Samples were spun at 5,000 Hz, temperature 4°C.

Proton spectra were obtained using a 1D T2-filtered sequence [90° − (τ − 180° − τ)n − acquisition] (spin echo, CPMG—Carr-Purcell-Meiboom-Gill) [27, 31] to suppress signals from lipids and macromolecules. Five hundred and twelve transients were collected, using a 7.2-kHz spectral region with 32K data points. The T2 filter contained delays of τ=1 ms and n=72 loops, giving an effective echo time of 144 ms. Acquisition time was 2.28 s, and zero-filling to 64K and an exponential line broadening of 1 Hz were applied to the raw data before Fourier transformation. A selective pre-saturation pulse was applied to enhance the water suppression. Also, spectra were acquired using the NOESYPR1D pulse sequence [relaxation delay − 90° − t1 − 90° − tm − 90° − acquisition] (Bruker BioSpin), with t1 fixed at 3 μs and a mixing time tm of 100 ms. The water resonance was irradiated during the relaxation delay and tm. The same spectral parameters and number of scans as in the T2-filtered acquisitions were used.

For peak identification purposes, two-dimensional (2D) spectra such as magnitude-mode chemical shift-correlated spectroscopy (COSY) and J-resolved spectra were acquired, all under MAS conditions. Gradient-selected homonuclear COSY experiments were acquired with 64 transients, 256 time-domain points in t1, acquisition time 0.38 s and spectral width 5.4 kHz in both dimensions. The spectra were processed with unshifted sine window functions in both directions. HR-MAS 1H NMR spectra were assigned with the aid of these 2D experiments and by comparison with spectra of authentic compounds, together with reference to previous reports [79, 11, 14, 23, 28, 36, 39]. To analyse metabolic stability of the lens tissues during the HR-MAS NMR spectroscopy, some samples were spun continuously overnight at 4°C prior to a repeated acquisition. No obvious changes in lens metabolic profile were observed.

Statistical analysis

Both CPMG and NOESYPR1D HR-MAS NMR spectra were initially analysed by principal component analysis (The Unscrambler, CAMO, Norway). However, discrimination between treated and non-treated rabbit lenses was not different for the two types of spectra. Thus, further processing of the 1D spectra for the purpose of quantification was based on CPMG spectra, using software for analysis of complex mixtures (AMIX; Bruker BioSpin). The data were reduced to a resolution of 0.6 Hz/point, omitting the spectral region downfield from 5 ppm due to low signal-to-noise ratio. Metabolite concentrations in treated groups were calculated relative to the levels in the untreated group after normalisation (absolute NMR peak integrated/sample weight) [4]. Student’s t test (two-tailed) was used for statistical analysis of the data. Statistical significance was set at P<0.05 (confidence level at 95%).


Representative T2-filtered HR-MAS 1H NMR spectra of the normal rabbit lens and lenses exposed to either only UVB radiation or combination of UVB and steroid pre-treatment are shown in Fig. 1. Signals from more than 15 major metabolites were assigned in the lens spectra, including glucose, reduced glutathione (GSH), sorbitol and sorbitol-3-phosphate, lactate, myo-inositol and scyllo-inositol, glycine, taurine, glycerophosphocholine, phosphocholine, choline, glutamate, acetate, alanine and valine. Treatment with steroids in combination with UVB exposure induced more evident changes in the spectra than did UVB irradiation alone. Compared with normal rabbit lens, substantial changes were observed in the spectral region 3.2–4.7 ppm of the lens treated with dexamethasone in combination with UVB. The most obvious was the decline of the β-amino acid taurine after this treatment (Fig. 1c). As shown in Fig. 1, intense signals from taurine were present in normal (a) and UVB-treated lenses (b), represented by two triplets centred at chemical shifts (δ) 3.26 ppm and 3.42 ppm, respectively. The results did not reveal significant difference in taurine concentration between normal and UVB-treated lenses. However, after the combined treatment with steroids and UVB radiation, the signals from taurine were almost invisible (Fig. 1c).
Fig. 1

HR-MAS 1H NMR spectra of rabbit lenses. a Normal lens; b lens after UVB exposure; c lens treated with dexamethasone in combination with UVB exposure. Assignments: Ace acetate, Ala alanine, Cho choline, Glc glucose, Glu glutamate, Gly glycine, GPC glycerophosphocholine, GSH reduced glutathione, Lac lactate, M-ins myo-inositol, PC phosphocholine, S-ins scyllo-inositol, Sorb sorbitol, Sorb-3-P sorbitol-3-phosphate, Tau taurine, Val valine

As evident in Fig. 1, both the normal spectrum and the spectrum of UVB-exposed lens were to a large extent dominated by signals from inositols in the shift range 3.2–4.1 ppm. In addition, this sugar/polyol region had signals from sorbitol and sorbitol-3-phosphate. Myo-inositol had a triplet at 3.27 ppm, adjacent to a triplet from taurine and singlets from choline-containing compounds (choline, phosphocholine and glycerophosphocholine) in the crowded region 3.20–3.30 ppm. A double doublet at 3.54 and triplets at 3.62 ppm and 4.06 ppm were assigned as myo-inositol. Scyllo-inositol was represented with a singlet at 3.35 ppm. Similar to taurine, the signals from myo-inositol and scyllo-inositol were markedly depressed in the dxm/UVB group. Concomitantly, numerous peaks from glucose indicated an evident increase in its concentration after steroid/UVB treatment, while glucose signals were undetectable in the spectra of normal lenses and lenses exposed to UVB alone (Fig. 1). The accumulation of glucose and depletion of taurine and inositols is clearly demonstrated in the 2D COSY spectra in Fig. 2, presented as contour plots for the relevant spectral region. In the spectrum shown in Fig. 1c, the remaining small peaks of taurine in the dxm/UVB group were hidden under peaks arising from glucose. However, this residual amount of taurine could still be detected by 2D spectroscopy (Fig. 2). Moreover, the dxm/UVB group showed an increase in signals from sorbitol and sorbitol-3-phosphate. In this group, the peaks in the spectral region 3.60–3.90 were mainly attributed to these two metabolites (Fig. 1c). In addition, the increased signal centred at 4.29 ppm was a multiplet assigned as sorbitol-3-phosphate. Figure 3 shows concentration levels of different metabolites in lenses exposed either to UVB radiation or a combination of UVB and steroid pre-treatment, expressed relative to the concentration in normal lenses. The extent of quantification was, however, limited in complex parts of the spectra with overlapping signals. Regarding observed differences between normal and UVB exposed lenses, the only significant changes (P<0.05) were found for the concentration of glutamate and lactate. Thus, average concentration of glutamate in the lens increased by 40%, and lactate decreased by 35% after UVB exposure. Among the other metabolites showing a tendency to concentration change after UVB exposure, but without reaching a significant level, were increasing levels of alanine and the inositols, and a decrease of GSH level (Fig. 3).
Fig. 2

2D 1H COSY contour plots (3.14–3.52 ppm) of rabbit lens treated with dexamethasone in combination with UVB exposure (left) and untreated lens (right). Corresponding 1D T2-filtered spectra are shown at the top (all spectra obtained under MAS conditions). Abbreviations as in Fig. 1

Fig. 3

Changes in metabolite concentrations in rabbit lens after UVB exposure (n=6) and after treatment with dexamethasone in combination with UVB exposure (n=7), measured as percentage change (±95% CI) compared with concentrations in the normal lens (n=6). For explanation of abbreviations, see Fig. 1

Treatment with dexamethasone in combination with UVB exposure induced significant changes for several metabolites compared with normal lens (Fig. 3). The concentration level of GSH was reduced by 54%, together with a substantial drop in inositol levels. The mean decrease in myo-inositol was 72%, equal to the mean reduction of scyllo-inositol. The concentration of lactate was reduced by 34%. Among the quantifiable amino acids, the level of alanine was more than doubled (107% increase) and the level of valine decreased by 28%. In addition, the level of choline decreased by 49% compared with the concentration in normal lens.

Despite all these metabolic changes measured with HR-MAS 1H NMR spectroscopy in the lenses, all lenses remained clear as confirmed by regular slit-lamp examination of the rabbit eyes. No lens opacities could be detected by careful biomicroscopy 1 day after the UVB exposure.


The results of this study show that UVB exposure alone or combined with steroid pre-treatment induced significant changes in the metabolic profile of rabbit lens. One of the striking effects of combined steroid and UVB treatment was the depletion of taurine. Taurine is the most abundant free amino acid in rabbit lens [21]. The rate of taurine uptake in the lens is very low, which implies the presence of an endogenous taurine-synthesising mechanism [32]. Taurine functions as an osmoregulator, ion flux regulator and membrane stabiliser [19]. In diabetic conditions taurine is depleted, which is considered the result of osmotic compensation in response to the accumulation of sorbitol within the lens [32]. Moreover, taurine is assumed to be active as an antioxidant. Reduction of taurine level in the lens might increase the risk of lens protein oxidation with subsequent cataract formation [26].

The combined treatment was also shown to cause a large reduction in the concentrations of inositols. Together with taurine and sorbitol, myo-inositol is one of the major osmolytes in the lens [46]. Using cultured bovine lens epithelial cells, Reeves and Cammarata [35] recognised the movement of myo-inositol from cell to medium, as induced by intracellular polyol accumulation. This polyol-activated release of myo-inositol from the lens is a mechanism supporting our findings. Lowered myo-inositol and taurine levels following the combined treatment in our study suggest a change in osmotic regulation processes in the lens.

Lenses incubated with dexamethasone have shown a decrease in ATP concentration and an increase in sugar phosphates, suggesting that dexamethasone antagonises the cellular uptake or utilisation of glucose [12]. In the present study, the concentration of lenticular glucose was shown to increase considerably following combined long-term steroid treatment/UVB radiation. Concomitantly, the concentration of lactate decreased reaching a level similar to that measured after UVB exposure only. In contrast, Pescosolido et al. [33] did not find any change in lenticular lactate concentration after long-term dexamethasone treatment of rabbit eyes. This implies that the decreased lactate concentration in our study may have been caused by the UVB irradiation. Reduced lactate level might be due to inactivation of glycolytic enzymes after UVB exposure [24, 25, 37, 40].

A possible metabolic pathway of lenticular glucose is the conversion into sorbitol. In diabetic and galactosemic conditions, glucose is funnelled into the polyol pathway for conversion into sorbitol by activation of aldose reductase [16, 20]. Sugar alcohols poorly permeate cell membranes, so polyols accumulate intracellularly [26]. In diabetic lens, sorbitol-3-phosphate is produced from sorbitol by a 3-phosphokinase [22]. In the present study, sorbitol and sorbitol-3-phosphate were, on the basis of spectrum interpretation, found to increase after the combined treatment with steroids/UVB exposure. However, this was not exactly quantifiable due to severe spectral overlap. The decreased inositol and taurine levels, with a concomitant increase in sorbitol/sorbitol-3-phosphate, support previous reports [26, 33].

GSH is a vital lens antioxidant [10]. It is synthesised within the lens, and a common feature of most types of cataracts is a decrease in the GSH level [34]. It has been suggested that glucocorticoid activity mediated through a glucocorticoid receptor [15] is responsible for loss of lenticular GSH, and thereby leading to cataractogenesis [6]. In the present study, the concentration of GSH decreased considerably after the combined treatment. An accompanying increase in oxidised glutathione (GSSG) was not observed, which might indicate that GSH synthesis was impaired. In fact, GSSG was not detected in any lens from treated or untreated eyes. Under normal conditions, essentially all detectable cellular glutathione is in a reduced state [38]. Generally, HR-MAS 1H NMR spectroscopy is an excellent method to discriminate between GSH and GSSG in unprocessed biological tissue.

The rise in alanine in both treated groups might be induced by the UVB radiation, as alanine is an oxidation product of tryptophan in the human lens [17].

As shown previously in studies with other biological tissues, HR-MAS 1H NMR spectroscopy offers the ability to obtain high-resolution spectra from inhomogeneous samples with acceptable signal-to-noise ratio (S/N) [1, 4, 5, 30]. Due to its avascularity and the highly ordered state of its cells, the crystalline lens is a rewarding tissue for application of the MAS technique. To our knowledge, the present study has applied HR-MAS 1H NMR spectroscopy in the analysis of lens tissue sections for the first time. The major advantage of the HR-MAS method is that high-quality spectra can be obtained from intact tissue samples. In contrast to previous reports on NMR spectroscopy of lens extracts from our laboratory and others [13, 28, 36], labour-intensive and tissue-destructive extraction methods are avoided. Tissue extractions involve possible oxidation or hydrolysis of substances, and volatile compounds might be lost.

Compared with enzymatic and chromatographic methods, an inherent advantage of NMR spectroscopy is the simultaneous detection of different groups of compounds, e.g. amino acids and glycolytic intermediates and products. In addition, by means of the HR-MAS technique, water-soluble metabolites and lipids can be measured in the same sample.

Due to the non-destructive analysis in HR-MAS spectroscopy, also compartmentalisation of metabolites within different cellular environments can be assessed. Thus, HR-MAS NMR spectroscopy is a link to, and might be a step towards, in vivo biochemical analysis of the lens by means of NMR spectroscopy.

In this study, frozen lens samples were sectioned and then put into the HR-MAS rotor. Efforts were made to keep the samples cooled until insertion into the spectrometer. During acquisition of spectra the temperature of the samples was kept at 4°C after careful calibration for heating effects due to spinning of the samples. As reported by Waters et al. [43], there may be some variation in the metabolic state of some tissues during MAS 1H NMR experiments. Our study revealed a surprisingly high metabolic stability of lens samples during the analysis, a favourable finding for accomplishment of prolonged experiments with 2D NMR spectroscopy. Spectral quantification of the metabolites was based on CPMG spectra, i.e. filtered spectra in terms of T2 relaxation. Hence, without accurate T2 measurements, absolute quantification could not be performed. However, it was assumed that each metabolite has the same T2 in different lenses, so a relative quantification was performed using absolute peak integrals normalised by sample weight.

The HR-MAS 1H NMR spectra were acquired from a section of each lens. The standardised excised samples contained both cortical and nuclear parts, so spatially averaged metabolic profiles of the lenses were obtained. Minor inconsistencies during this procedure may, however, occur providing lens samples with a larger inhomogeneity than desirable. Using this method, the metabolic asymmetry between nucleus, cortex and epithelium was not taken into account. For further work, an improved sectioning technique has been initiated to provide representative samples from either lens cortex or nucleus. Analysis of these regions separately would increase the benefit from HR-MAS NMR spectroscopy. Generally, spectra of good quality can be obtained from very small lenticular sections, making HR-MAS NMR spectroscopy a potentially powerful tool for the study of disease or lenticular metabolic change due to drugs or endogenous pathophysiological stimuli.

In conclusion, HR-MAS 1H NMR spectroscopy was found to be a valuable method for investigating the metabolic state of intact lens samples. The results show high-quality spectra with high signal-to-noise ratio from small lenticular sections. Significant metabolic changes were detected in the rabbit lens after long-term treatment with dexamethasone combined with a subsequent UVB exposure. The main findings include depletion of taurine and myo-inositol, an increase in glucose and sorbitols and a decrease in the GSH level. The metabolic state of the lens was changed more profoundly by long-term steroid treatment than by short-term UVB exposure. For the level of GSH, the effect of these two factors seems to be additive. The observed metabolic changes are in agreement with those reported by other studies and involved in the development of cataract.


Čestmír Čejka is gratefully acknowledged for radiometric measurements. This study was supported by grants from The Research Council of Norway, grant no. 304/03/0419 from the Grant Agency of the Czech Republic and grant AVOZ5008914 from the Academy of Sciences of the Czech Republic. J.K. thanks Bruker BioSpin GmbH, Rheinstetten, Germany, for continuous support in development of probes for making HR-MAS a viable method for metabolic profiling in tissue.

Copyright information

© Springer-Verlag 2004