Resonance Raman detection of carotenoid antioxidants in living human tissue

Abstract

Increasing evidence points to the beneficial effects of carotenoid antioxidants in the human body. Several studies, for example, support the protective role of lutein and zeaxanthin in the prevention of age-related eye diseases. If present in high concentrations in the macular region of the retina, lutein and zeaxanthin provide pigmentation in this most light sensitive retinal spot, and as a result of light filtering and/or antioxidant action, delay the onset of macular degeneration with increasing age. Other carotenoids, such as lycopene and beta-carotene, play an important role as well in the protection of skin from UV and short-wavelength visible radiation. Lutein and lycopene may also have protective function for cardiovascular health, and lycopene may play a role in the prevention of prostate cancer. Motivated by the growing importance of carotenoids in health and disease, and recognizing the lack of any accepted noninvasive technology for the detection of carotenoids in living human tissue, we explore resonance Raman spectroscopy as a novel approach for noninvasive, laser optical carotenoid detection. We review the main results achieved recently with the Raman detection approach. Initially we applied the method to the detection of macular carotenoid pigments, and more recently to the detection of carotenoids in human skin and mucosal tissues. Using skin carotenoid Raman instruments, we measure the carotenoid response from the stratum corneum layer of the palm of the hand for a population of 1375 subjects and develop a portable skin Raman scanner for field studies. These experiments reveal that carotenoids are a good indicator of antioxidant status. They show that people with high oxidative stress, like smokers, and subjects with high sunlight exposure, in general, have reduced skin carotenoid levels, independent of their dietary carotenoid consumption. We find the Raman technique to be precise, specific, sensitive, and well suitable for clinical as well as field studies. The noninvasive laser technique may become a useful method for the correlation between tissue carotenoid levels and risk for malignancies or other degenerative diseases associated with oxidative stress.

Fig. 1

Absorption spectra, molecular structures, and energy level scheme of major carotenoid species found in human tissue, including β-carotene, zeaxanthin, lycopene, lutein, and phytofluene. Carotenoid molecules, which feature an unusual even parity excited state (see inset). As a consequence, absorption transitions are electricdipole allowed in these molecules but spontaneous emission is forbidden. The resulting absence of any strong fluorescence in carotenoids is the main reason for the possibility to use RRS, shown as a solid, downward-pointing arrow (optical transition) in the inset, as a noninvasive means of carotenoid detection in human tissue. All carotenoid molecules feature a linear, chainlike conjugated carbon backbone consisting of alternating carbon single (C—C) and double bonds (C==C) with varying numbers of conjugated C==C double bonds, and a varying number of attached methyl side groups. Beta-carotene, lutein, and zeaxanthin feature additional ionone rings as end groups. In beta-carotene and zeaxanthin, the double bonds of these ionone rings add to the effective C==C double bond length of the molecules. Lutein and zeaxanthin have an OH group attached to the ring. Lycopene has 11 conjugated C==C bonds, beta carotene has 11, zeaxanthin has 11, lutein has 10, and phytofluene has 5. The absorptions of all species occur in broad bands in the blue/green spectral range, with the exception of phytofluene, which as a consequence of the shorter C==C conjugation length absorbs in the near UV. Note also, that a small (~10 nm) spectral shift exists between lycopene and lutein absorptions. The spectral shifts can be explored with RRS to selectively detect some of the carotenoids existing as a mixture in human tissues.

Fig. 2

Resonance Raman spectra of β-carotene, zeaxanthin, lycopene, lutein, and phytofluene solutions, showing the three major “spectral fingerprint” Raman peaks of carotenoids originating from rocking motions of the methyl components (C–CH₃) and stretch vibrations of the carbon–carbon single bonds (C—C) and double bonds (C==C). In all carotenoids except phytofluene, these peaks appear at 1008, 1159, and 1525 cm⁻¹, respectively. In phytofluene, the C==C stretch frequency is shifted by ~40 cm⁻¹ to higher frequencies. Note the large contrast between Raman response and broad background signal, which is due to the inherently weak fluorescence of carotenoids.

Fig. 3

Absorption spectra and resonance Raman responses for solutions of β-carotenes, lycopenes, and a mixture of both. The Raman response for the mixture corresponds to the sum of responses for individual concentrations, measured with 488 nm excitation. Results demonstrate capability of resonance Raman spectroscopy to detect composite response of several carotenoids if excited at proper spectral wavelength within their absorption bands.

Fig. 4

RRS responses in the spectral vicinity of the C==C stretching frequency, measured for various antioxidants as a function of concentration in ethanol under identical excitation and detection conditions—●, β-carotene; □, ascorbic acid; ▵, α-tocopherol—from Ref. . Note the absence of a Raman response for noncarotenoid compounds.

Fig. 5

RRS response of various carotenoid solutions under 488-nm excitation, shown as a function of increasing concentrations—filled triangle, β-carotene; open triangle, zeaxanthin; open square, lutein; and open circle, lycopene. The response is linear up to ~5 μg/mL, which is a carotenoid concentration exceeding that found in human skin. Slight variations in the slopes for each carotenoid solution are in very good agreement with their respective excitation efficiencies. Note the near coincidence of responses for lutein and zeaxanthin.

Fig. 6

(a) Fundus photograph of healthy human retina, showing optic nerve head (bright spot at left) and macula [dark shaded area outlining macular pigment (MP) distributions], and (b) schematic representation of retinal layers participating in light absorption, transmission, and scattering of excitation and emission light: ILM, inner limiting membrane; NFL, nerve fiber layer; HPN, Henle fiber, plexiform, and nuclear layers; PhR, photoreceptor layer; RPE, retinal pigment epithelium. In Raman scattering, the scattering response originates from the MP, which is located anteriorly to the photoreceptor layer. The influence of deeper fundus layers such as the RPE is avoided. In autofluorescence spectroscopy, light emission of deeper fundus layers, such as lipofuscin emission from the RPE, can be stimulated on purpose to generate an instrinsic “light source” for single-path absorption measurements of anteriorly located MP layers (see text).

Fig. 7

(a) Schematic diagram of macular pigment resonance Raman detector designed for human clinical studies. Light from an argon laser is routed via optical fiber into an optical probe head (outlined by dashed line) and from there into the eye of a subject where it is projected as ~1 mm diameter spot onto the retina. L1 to L4, lenses; F, laser line filter; BS: dichroic beamsplitter; NF, notch filter; VHF, volume holographic grating; and LED, red-light-emitting diode for visualization of fiber bundle. Raman scattered light is collected in backscattering geometry with lens L3, split off by BS, and sent into a spectrograph via fiber bundle for light dispersion. A CCD array is used to detect the spectrally dispersed light. The instrument is interfaced to a personal computer for light exposure control, data acquisition, and processing. Typical settings are 1 mW of 488-nm laser light for 0.2 s with a 1-mm spot size on the macula. (b) Subject looking into the optical probe head of the instrument. (c) Typical Raman spectra from the retina of a healthy volunteer, measured with dilated pupil (8 mm), and displayed on the computer monitor of the instrument. Left panel, spectrum obtained after a single measurement, clearly showing the carotenoid Raman signals superimposed upon a broad fluorescent background; right panel, same enlarged spectrum obtained after fitting background with a fourth-order polynomial and subtracting it from original spectrum. For quantitation of MP concentration the software displays the Raman response of the strongest peak, corresponding to the C==C stretch, as an intensity score.

Fig. 8

Correlation of RRS signals obtained for the C==C double bond vibration at 1525 cm⁻¹ with the carotenoid content of six monkey retinae as determined by HPLC. A linear fit to the data results in a correlation coefficient of 0.68.

Fig. 9

RRS MP measurements of 33 normal eyes for a young group of subjects ranging in age from 21 to 29 years. Note the large (~10-fold) variation of RRS levels between individuals. Since the ocular transmission properties in this age group can be assumed to very similar, the variations are can be assigned to differing MP levels. Subjects with extremely low carotenoid levels may be at higher risk of developing macular degeneration later in life.

Fig. 10

(a) Schematic diagram used for resonance Raman imaging of the MP distribution of volunteer subjects. Light from blue and green excitation laser wavelengths is sequentially projected onto the retina of a subject as a ~5-mm-diam spot, while the fellow eye is looking at a fixation target. The Raman-scattered light is collimated by the eye lens and imaged simultaneously onto the 2-D arrays of two separate digital imaging CCD cameras, using a beamsplitter (QBS) to generate two separate light paths. One of the cameras images the spatial distribution of the excitation light (reference), while the other camera images light levels in the spectral range of the C==C stretch Raman peak. A rotatable, narrowband filter (F2) is used to selectively realize C==C on-peak and off-peak transmission for the scattered light. Following the registration of an on-peak and an off-peak image, the two images are digitally subtracted (following alignment using vessel landmarks) and displayed as topographic pseudocolor images or surface plots showing the spatial distribution of MP concentrations. L1 to L3: lenses; F1, laser line filter; BS, dichroic beam splitters; F2, tunable filter; F3, bandpass filter. (b) Pseudocolor surface plot of MP distribution in a living human eye measured with out Raman imaging instrument. Raman intensities are coded according to the intensity scale shown on the upper right-hand side of the image. Note the roughly circularly symmetric MP distribution in the central area of the fovea but the appearance of side bumps in peripheral areas. (c) Line scan of the image, in (b) obtained by plotting intensities along a horizontal meridional line running through the center of the macular pigment distribution. The width at half maximum is about 150 μm in this individual.

Fig. 11

Layer structure of human skin as seen in a microscope after staining, showing the morphology of dermis, basal layer, stratum spinosum, stratum granulosum, and stratum corneum. Cells of the stratum corneum have no nucleus (missing dark stain spots) and form a relatively homogeneous optical medium well suited for Raman measurements.

Fig. 12

(a) Skin carotenoid resonance Raman detector, showing argon laser, spectrograph, light delivery/collection module, and excitation laser spot on the palm of the hand of a subject. A typical measurement involves the placement of the palm of the hand against the window of the module and exposing the palm for about 1 min at laser intensities of ~10 mW in a 2-mm-diam spot. Carotenoid Raman signals are detected with a 2-D CCD camera integrated with the spectrograph (left side of image), and processed similar to the ocular Raman instrument. (b) Typical skin carotenoid Raman spectra measured in vivo. Spectrum shown on the left is spectrum obtained directly after exposure, and reveals broad, featureless, and strong autofluorescence background of skin with superimposed sharp Raman peaks characteristic for carotenoid molecules. Spectrum at the right is difference spectrum obtained after fitting fluorescence background with a fourth-order polynomial and subtracting it from the spectrum on the left. The main characteristic carotenoid peaks are clearly resolved with good SNRs at 1159 and 1524 cm⁻¹.

Fig. 13

Spectral dependencies of the C==C double bond RRS skin response (open triangles), skin AF background (open circles), and the resulting ratio between both responses (filled circles), measured with three blue/green argon laser excitation wavelengths and one frequency-doubled Nd:YAG laser wavelength for the inner palm stratum corneum layer of a volunteer subject. The spectral dependence of the ratio identifies an optimum excitation wavelength around 500 nm.

Fig. 14

(a) Correlation of skin resonance Raman intensity measured in the inner palm of the hand with serum carotenoids determined by HPLC, obtained for a group of 104 healthy male and female adults. Note the high correlation coefficient of r=0.78 (p<0.001). (b) Histogram of skin carotenoid resonance Raman response measured in the palm of hands for 1375 subjects, showing wide distribution of skin carotenoid levels in a large population.

Fig. 15

(a) Resonance Raman intensity (counts) versus reported number of daily consumed servings of fruits and vegetables, demonstrating increase of skin carotenoid concentrations with increased fruit and vegetable uptake. Means ± standard deviation (SD). (b) Resonance Raman intensity (counts) versus self-reported exposure of skin to sunlight, showing decrease of skin carotenoid levels with increased sun exposure. (c) Resonance Raman intensity (counts) in nonsmokers and cigarette smokers, showing ~30% decrease of skin carotenoid levels in smokers.

Fig. 16

(a) Resonance Raman spectrum of an acetone solution of lycopene (solid lines) and β-carotene (dotted lines), measured under argon laser excitation at 488 (left panel) and 514.5 nm (right panel). Both solutions had identical carotenoid concentrations. Raman spectra were recorded using identical excitation power and sensitivity-matched instruments. Strongest Raman peaks correspond to the stretch vibrations of the carbon single and double bonds of the molecule (at ~1159 and ~1525 cm⁻¹, respectively). Note strongly reduced Raman response of C==C stretch for lycopene compared to β-carotene under 514.5-nm excitation. (b) Beta-carotene and lycopene skin Raman levels measured with selective resonant Raman spectroscopy for seven subjects. Note strong intersubject variability of β-carotene to lycopene ratios (indicated above bar graphs).

Fig. 17

C==C double-bond Raman spectra of lutein (thick line) and zeaxanthin (thin line) dissolved in methanol and measured at room temperature under excitation with 488-nm light. A small (~2.5 cm⁻¹) but readily detectable chemical shift of the double-bond stretching frequency in these two very similar compounds, which differ in effective conjugation length by one double bond, could be potentially used for selective detection of each carotenoid in retinal tissue.

Fig. 18

Handheld optical probes used for RRS carotenoid detection of human and other biological tissues: LLF, laser line filter; HNF, holographic notch filter; M, mirror; W, optical window; BS, beamsplitter. (a) and (b) Probes use colinear excitation-detection geometry and can be used to sample near the tissue surface or inside the tissue if interfaced with a 1-mm-diam fiber needle. (c) and (d) Raman probes use oblique excitation-detection geometry, having the advantage of lowering confounding tissue fluorescence and eliminating the need for a dichroic beam splitter otherwise placed inside the probe. (d) Two-excitation channel probe is used for selective detection of lycopene and carotenes, respectively, in human skin and ensures identical probe volume for each wavelength during measurements.