629400
research-article2016
JDRXXX10.1177/0022034516629400Journal of Dental ResearchNonlinear Optical Signals for Caries Detection
Research Reports: Biological
Laboratory Studies of Nonlinear Optical Signals for Caries Detection
Journal of Dental Research 1–6 © International & American Associations for Dental Research 2016 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0022034516629400 jdr.sagepub.com
E. Terrer1,2, I.V. Panayotov1, A. Slimani1, D. Tardivo3, D. Gillet1, B. Levallois1, O. Fejerskov4, C. Gergely5, F.J.G. Cuisinier1, H. Tassery1,2, and T. Cloitre5
Abstract Multiphoton confocal microscopy and nonlinear spectroscopy are used to investigate the caries process in dentin. Although dentin is a major calcified tissue of the teeth, its organic phase comprises type I collagen fibers. Caries drive dentin demineralization and collagen denaturation. Multiphoton microscopy is a powerful imaging technique: the biological materials are transparent to infrared frequencies and can be excited to penetration depths inaccessible to 1-photon confocal microscopy. The laser excitation greatly reduces photodamage to the sole focal region, and the signal-to-noise ratio is improved significantly. The method has been used to follow pathologic processes involving collagen fibrosis or collagen destruction based on their 2-photon excited fluorescence (2PEF) emission and second harmonic generation (SHG). Combining multiphoton imaging with nonlinear spectroscopy, we demonstrate that both 2PEF and SHG intensity of human dentin are strongly modified during the tooth caries process, and we show that the ratio between SHG and 2PEF signals is a reliable parameter to follow dental caries. The ratio of the SHG/2PEF signals measured by nonlinear optical spectroscopy provides valuable information on the caries process, specifically on the degradation of the organic matrix of dentin. The goal is to bring these nonlinear optical signals to clinical application for caries diagnosis. Keywords: biophotonics, microscopy, demineralization, diagnostic systems, restorative dentistry, risk factor(s)
Introduction Multiphoton microscopy (MPM) is a widely used imaging technique in fundamental or clinical biomedical research (Zipfel et al. 2003), not only for various cellular (Brockbank et al. 2008; Zoumi et al. 2002), tissue (Teng et al. 2006), or organ (Campagnola et al. 2002) studies but also for investigation of collagen fibers, tumors, and brain (Mohler et al. 2003; König et al. 2005). The method has been introduced for dental research as, for example, in caries diagnosis (Hall and Girkin 2004; Kao 2004; Lin et al. 2010) or in monitoring structural modifications of enamel (Chen et al. 2007; Chen et al. 2008; J Chen et al. 2009; WL Chen et al. 2009). Among the various methods used to study collagen denaturation—such as atomic force microscopy, x-ray diffraction, or Fourier transform infrared spectroscopy—MPM is a nonlinear high-resolution imaging technique that can be applied to collagen studies under in vivo conditions (Abraham et al. 2010). The main advantage of MPM lies in its simultaneous image creation based on the fluorescence signals produced by the process of multiphoton excitation (2-photon excited fluorescence [2PEF]) and a coherent nonlinear phenomenon, second harmonic generation (SHG), providing useful information on the structure and optical properties of a specimen. The SHG signal highlights the structure of noncentrosymmetric biological systems, such as collagen fibrils, muscle fibers, and microtubules (Abraham et al. 2010; Pan et al. 2014). Noncentrosymmetric
materials lack inversion symmetry. If the molecule is not centrosymmetric, the movement of electrons will be asymmetrical, and the scattered radiation will contain the usual frequency component and nonlinear components. The nonlinear component is the harmonic light scattering. Only noncentrosymmetric molecules produce the SHG signal. This is an interaction in which 2 photons at the same frequency are converted into 1 photon at twice the incident frequency. It has been demonstrated that type I collagen is efficient in generating second harmonic signals because it lacks inversion symmetry (Williams et al. 2005). Monitoring SHG has already been shown to have potential applicability for cancer diagnosis by revealing changes in the extracellular matrix in tumors (epithelial cancers such as breast cancer) as compared with normal tissues. 1
Laboratoire Bio ingénierie et Nanosciences, Université de Montpellier, Montpellier, France 2 Université d’Aix-Marseille, Marseille, France 3 Laboratoire Anthropologie bio-culturelle, droit, éthique et santé, UMR 7268 CNRS-Université d’Aix-Marseille, Marseille, France 4 Department of Biomedicine-Anatomy Health, Aarhus University, Aarhus, Denmark 5 Laboratoire Charles Coulomb, UMR 5221 CNRS-Université de Montpellier, Montpellier, France. Corresponding Author: E. Terrer, 9 montée de Bel Air, le vieux Peypin, 13124 Peypin, France. Email: [email protected]
Downloaded from jdr.sagepub.com at Middle East Technical Univ on February 12, 2016 For personal use only. No other uses without permission. © International & American Associations for Dental Research 2016
2
Journal of Dental Research diamond saw (Isomet 1000; Buehler, Lake Bluff, IL, USA) with a thickness of up to 0.5 mm. Samples were grinded to 0.25 mm and polished on carbure disks with diamond pastes (6, 1, and 0.25 µm) on an Escil polishing machine (Escil, Lyon, France). Finally, specimens were thoroughly cleaned in water with an ultrasonic bath for 5 min to avoid significant residual surface contamination.
Soprolife Camera
Figure 1. Longitudinal section of dentin observed with Soprolife camera. Green: sound dentin (1); red: carious dentin (2).
The potential of MPM to investigate the structure of hard tooth tissues has already been demonstrated (Chen et al. 2007). In dentin tubule, 2PEF intensity varies due to protein content variation. Intertubular dentin produces both SHG and 2PEF signals. Tubules are surrounded by a thin circular zone with a lower SHG signal than the bulk dentin, and the presence of collagen fibers perpendicular to the tubule longitudinal axis is indicated by strong SHG responses. The dentin-enamel junction appears as a low-intensity line on the 2PEF. The SHG signal is completely absent for prismatic enamel, and contrarily, aprismatic enamel shows a homogeneous low 2PEF signal. In a previous investigation, we presented a detailed description of the tooth enamel and dentin at the dentin-enamel junction (Cloitre et al. 2013). In the present study, we investigated the structural modifications during the dentin caries process by monitoring the changes of the SHG signal (images) and the SHG/2PEF ratio recorded by nonlinear optical spectroscopy (NLOS). The SHG/2PEF ratio is a new parameter that could provide precise information on the degradation of the dentin organic matrix and modifications in the collagen structure. Clinically, the modification of collagen allows us to distinguish infected and affected dentin. In fact, in infected dentin, the collagen is completely destroyed. Moreover, we used a Soprolife camera (Acteon, La Ciotat, France) to distinguish infected or affected dentin; indeed, the fluorescence of the infected dentin is different from the fluorescence of the affected dentin (Terrer et al. 2010).
Materials and Methods Specimen Preparation Ten adult teeth, freshly extracted, for orthodontic or periodontal reasons were collected (written consent was obtained before extraction and with the agreement of the ethical committee of Montpellier hospital): 5 sound teeth and 5 carious teeth (checked with Soprolife camera). They were cleaned with an air polisher (AIR-N-GO; Acteon, Bordeaux, France) and then stored at 4 °C in sterile water, renewed daily. Freshly extracted teeth were sectioned in the longitudinal axis by means of a
The intraoral Soprolife camera utilizes 2 series of LED that can illuminate tooth surfaces within the visible domain. The camera can detect and locate differences in the density, structure, and chemical composition of biological tissue subjected to continuous lighting in 1 frequency band, while making it generate a fluorescence phenomenon in a second frequency band. This provides an anatomic image superposed with an image created by the autofluorescence of the teeth. The camera is equipped with an image sensor (a 0.25-in. charge-coupled device [CCD] sensor) consisting of a mosaic of pixels covered with filters of complementary colors. The collected data, relating to the energy received by each pixel, enable an image of the tooth to be retrieved. The Soprolife camera operates in 3 modes: a daylight mode (4 white LEDs generate daylight), a fluorescence diagnostic mode, and a fluorescence treatment mode (for the latter 2 modes, 4 blue LEDs emitting at 450 nm are used). The difference between the diagnostic and treatment modes is due to different CCD settings. Using the Soprolife camera (Fig. 1) as a supplementary aid for visual diagnosis of caries, we identified the presence of caries first on the whole tooth and then after preparation on dentin cutting to check. We have determined the regions of interest to precisely allocate the origin of the recorded SHG intensities. In the fluorescence diagnostic mode of the camera, the sound dentin appears in an acid green color, whereas the carious dentin is shown in red.
Multiphoton Microscopy SHG and 2PEF images were recorded with a custom-built multiphoton microscope based on a Zeiss Imager Z1 upright microscope. For sample excitation, we used a Spectra-Physics Tsunami Ti-Sapphire laser operated in pulsed mode (wavelength range, 820 to 890 [typically 870 nm]; repetition rate, 80 MHz; pulse duration, ~100 fs; Zoumi et al. 2002). The laser beam was fed into the back port of the microscope, deflected by a short-pass 670-nm dichroic beam splitter (Chroma Technology Corp, Bellows Falls, VT, USA) and focused onto the sample with a Fluar oil-immersion microscope objective (40×, NA = 1.3; Zeiss, Iena, Germany; Stoller et al. 2002). Images were created by scanning the sample with an x/y scanning piezo stage (P-542-2CD; Physik Instrumente, Karlsruhe, Germany). Fluorescence signal from the samples was epicollected through the objective; discriminated from the backscattered laser light and SHG signal by the combination of the
Downloaded from jdr.sagepub.com at Middle East Technical Univ on February 12, 2016 For personal use only. No other uses without permission. © International & American Associations for Dental Research 2016
3
Nonlinear Optical Signals for Caries Detection dichroic mirror, a 700-nm short-pass emission filter, and a 475-nm long-pass filter; and detected with a photomultiplier tube (R928P; Hamamatsu, Tokyo, Japan). The SHG signal was collected through a 1.4-NA oil-immersion condenser with a wide-pass CG-BG39 filter and a 447-nm high-performance band-pass filter and then detected by a H7422P photomultiplier operated in photon-counting mode (Hamamatsu). The method enables acquisition of in-depth Z-stacks of 2-dimensional images of the sample based on their nonlinear optical response (SHG and 2PEF emission) when excited in the infrared domain (Fig. 2). The image is created by scanning the sample point by point (similar to confocal scanning) and collecting the nonlinearly generated photons from each pixel individually. MPM images were obtained by integrating the spectral intensity emitted by the samples from 410 to 470 nm and from 475 to 700 nm for the SHG and 2PEF images, respectively, as illustrated by the blue and green boxes in Figure 2. Indeed, there is a thin disturbed surface layer with a mean thickness of about 2 µm. However, we realized different measurements in Z-scanning to check the possible influence of the smear layer: We scanned until a depth of 50 µm by step of 2 µm and noticed that 2 µm from the surface no disturbance in the images and the spectra could be evidenced. Accordingly, all the spectra and the images presented in this study were obtained within 2 µm from the surface. Images were compiled with the Image J software (http://rsb.info.nih.gov/ij/; National Institutes of Health, Bethesda, MD, USA) through 16-color lookup tables. Intensity profile plots of SHG images were calculated with the same software.
Nonlinear Optical Spectroscopy NLOS was realized in epicollection mode with the same experimental setup, with the adjunction of a spectrometer (Acton SP215i-CCD PIXIS 400B Excelon; Princeton Instruments, Trenton, NJ, USA) connected to the camera port of the microscope via an optical fiber. The backscattered laser light was filtered out with the same dichroic/700-nm short-pass filter combination (Raub et al. 2008). SHG and 2PEF are 2 nonlinear optical signals. Spectra were recorded with WinSpec software from Princeton Instruments. They were collected from 5 different samples of sound and carious dentin. All spectra were normalized following the laser power and the acquisition time. Data were plotted: the mean and the standard deviation were calculated and the statistical analyses were performed with the nonparametric Wilcoxon test. When the data were compared 2 × 2, a Wilcoxon test was used. A P value
research-article2016
JDRXXX10.1177/0022034516629400Journal of Dental ResearchNonlinear Optical Signals for Caries Detection
Research Reports: Biological
Laboratory Studies of Nonlinear Optical Signals for Caries Detection
Journal of Dental Research 1–6 © International & American Associations for Dental Research 2016 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0022034516629400 jdr.sagepub.com
E. Terrer1,2, I.V. Panayotov1, A. Slimani1, D. Tardivo3, D. Gillet1, B. Levallois1, O. Fejerskov4, C. Gergely5, F.J.G. Cuisinier1, H. Tassery1,2, and T. Cloitre5
Abstract Multiphoton confocal microscopy and nonlinear spectroscopy are used to investigate the caries process in dentin. Although dentin is a major calcified tissue of the teeth, its organic phase comprises type I collagen fibers. Caries drive dentin demineralization and collagen denaturation. Multiphoton microscopy is a powerful imaging technique: the biological materials are transparent to infrared frequencies and can be excited to penetration depths inaccessible to 1-photon confocal microscopy. The laser excitation greatly reduces photodamage to the sole focal region, and the signal-to-noise ratio is improved significantly. The method has been used to follow pathologic processes involving collagen fibrosis or collagen destruction based on their 2-photon excited fluorescence (2PEF) emission and second harmonic generation (SHG). Combining multiphoton imaging with nonlinear spectroscopy, we demonstrate that both 2PEF and SHG intensity of human dentin are strongly modified during the tooth caries process, and we show that the ratio between SHG and 2PEF signals is a reliable parameter to follow dental caries. The ratio of the SHG/2PEF signals measured by nonlinear optical spectroscopy provides valuable information on the caries process, specifically on the degradation of the organic matrix of dentin. The goal is to bring these nonlinear optical signals to clinical application for caries diagnosis. Keywords: biophotonics, microscopy, demineralization, diagnostic systems, restorative dentistry, risk factor(s)
Introduction Multiphoton microscopy (MPM) is a widely used imaging technique in fundamental or clinical biomedical research (Zipfel et al. 2003), not only for various cellular (Brockbank et al. 2008; Zoumi et al. 2002), tissue (Teng et al. 2006), or organ (Campagnola et al. 2002) studies but also for investigation of collagen fibers, tumors, and brain (Mohler et al. 2003; König et al. 2005). The method has been introduced for dental research as, for example, in caries diagnosis (Hall and Girkin 2004; Kao 2004; Lin et al. 2010) or in monitoring structural modifications of enamel (Chen et al. 2007; Chen et al. 2008; J Chen et al. 2009; WL Chen et al. 2009). Among the various methods used to study collagen denaturation—such as atomic force microscopy, x-ray diffraction, or Fourier transform infrared spectroscopy—MPM is a nonlinear high-resolution imaging technique that can be applied to collagen studies under in vivo conditions (Abraham et al. 2010). The main advantage of MPM lies in its simultaneous image creation based on the fluorescence signals produced by the process of multiphoton excitation (2-photon excited fluorescence [2PEF]) and a coherent nonlinear phenomenon, second harmonic generation (SHG), providing useful information on the structure and optical properties of a specimen. The SHG signal highlights the structure of noncentrosymmetric biological systems, such as collagen fibrils, muscle fibers, and microtubules (Abraham et al. 2010; Pan et al. 2014). Noncentrosymmetric
materials lack inversion symmetry. If the molecule is not centrosymmetric, the movement of electrons will be asymmetrical, and the scattered radiation will contain the usual frequency component and nonlinear components. The nonlinear component is the harmonic light scattering. Only noncentrosymmetric molecules produce the SHG signal. This is an interaction in which 2 photons at the same frequency are converted into 1 photon at twice the incident frequency. It has been demonstrated that type I collagen is efficient in generating second harmonic signals because it lacks inversion symmetry (Williams et al. 2005). Monitoring SHG has already been shown to have potential applicability for cancer diagnosis by revealing changes in the extracellular matrix in tumors (epithelial cancers such as breast cancer) as compared with normal tissues. 1
Laboratoire Bio ingénierie et Nanosciences, Université de Montpellier, Montpellier, France 2 Université d’Aix-Marseille, Marseille, France 3 Laboratoire Anthropologie bio-culturelle, droit, éthique et santé, UMR 7268 CNRS-Université d’Aix-Marseille, Marseille, France 4 Department of Biomedicine-Anatomy Health, Aarhus University, Aarhus, Denmark 5 Laboratoire Charles Coulomb, UMR 5221 CNRS-Université de Montpellier, Montpellier, France. Corresponding Author: E. Terrer, 9 montée de Bel Air, le vieux Peypin, 13124 Peypin, France. Email: [email protected]
Downloaded from jdr.sagepub.com at Middle East Technical Univ on February 12, 2016 For personal use only. No other uses without permission. © International & American Associations for Dental Research 2016
2
Journal of Dental Research diamond saw (Isomet 1000; Buehler, Lake Bluff, IL, USA) with a thickness of up to 0.5 mm. Samples were grinded to 0.25 mm and polished on carbure disks with diamond pastes (6, 1, and 0.25 µm) on an Escil polishing machine (Escil, Lyon, France). Finally, specimens were thoroughly cleaned in water with an ultrasonic bath for 5 min to avoid significant residual surface contamination.
Soprolife Camera
Figure 1. Longitudinal section of dentin observed with Soprolife camera. Green: sound dentin (1); red: carious dentin (2).
The potential of MPM to investigate the structure of hard tooth tissues has already been demonstrated (Chen et al. 2007). In dentin tubule, 2PEF intensity varies due to protein content variation. Intertubular dentin produces both SHG and 2PEF signals. Tubules are surrounded by a thin circular zone with a lower SHG signal than the bulk dentin, and the presence of collagen fibers perpendicular to the tubule longitudinal axis is indicated by strong SHG responses. The dentin-enamel junction appears as a low-intensity line on the 2PEF. The SHG signal is completely absent for prismatic enamel, and contrarily, aprismatic enamel shows a homogeneous low 2PEF signal. In a previous investigation, we presented a detailed description of the tooth enamel and dentin at the dentin-enamel junction (Cloitre et al. 2013). In the present study, we investigated the structural modifications during the dentin caries process by monitoring the changes of the SHG signal (images) and the SHG/2PEF ratio recorded by nonlinear optical spectroscopy (NLOS). The SHG/2PEF ratio is a new parameter that could provide precise information on the degradation of the dentin organic matrix and modifications in the collagen structure. Clinically, the modification of collagen allows us to distinguish infected and affected dentin. In fact, in infected dentin, the collagen is completely destroyed. Moreover, we used a Soprolife camera (Acteon, La Ciotat, France) to distinguish infected or affected dentin; indeed, the fluorescence of the infected dentin is different from the fluorescence of the affected dentin (Terrer et al. 2010).
Materials and Methods Specimen Preparation Ten adult teeth, freshly extracted, for orthodontic or periodontal reasons were collected (written consent was obtained before extraction and with the agreement of the ethical committee of Montpellier hospital): 5 sound teeth and 5 carious teeth (checked with Soprolife camera). They were cleaned with an air polisher (AIR-N-GO; Acteon, Bordeaux, France) and then stored at 4 °C in sterile water, renewed daily. Freshly extracted teeth were sectioned in the longitudinal axis by means of a
The intraoral Soprolife camera utilizes 2 series of LED that can illuminate tooth surfaces within the visible domain. The camera can detect and locate differences in the density, structure, and chemical composition of biological tissue subjected to continuous lighting in 1 frequency band, while making it generate a fluorescence phenomenon in a second frequency band. This provides an anatomic image superposed with an image created by the autofluorescence of the teeth. The camera is equipped with an image sensor (a 0.25-in. charge-coupled device [CCD] sensor) consisting of a mosaic of pixels covered with filters of complementary colors. The collected data, relating to the energy received by each pixel, enable an image of the tooth to be retrieved. The Soprolife camera operates in 3 modes: a daylight mode (4 white LEDs generate daylight), a fluorescence diagnostic mode, and a fluorescence treatment mode (for the latter 2 modes, 4 blue LEDs emitting at 450 nm are used). The difference between the diagnostic and treatment modes is due to different CCD settings. Using the Soprolife camera (Fig. 1) as a supplementary aid for visual diagnosis of caries, we identified the presence of caries first on the whole tooth and then after preparation on dentin cutting to check. We have determined the regions of interest to precisely allocate the origin of the recorded SHG intensities. In the fluorescence diagnostic mode of the camera, the sound dentin appears in an acid green color, whereas the carious dentin is shown in red.
Multiphoton Microscopy SHG and 2PEF images were recorded with a custom-built multiphoton microscope based on a Zeiss Imager Z1 upright microscope. For sample excitation, we used a Spectra-Physics Tsunami Ti-Sapphire laser operated in pulsed mode (wavelength range, 820 to 890 [typically 870 nm]; repetition rate, 80 MHz; pulse duration, ~100 fs; Zoumi et al. 2002). The laser beam was fed into the back port of the microscope, deflected by a short-pass 670-nm dichroic beam splitter (Chroma Technology Corp, Bellows Falls, VT, USA) and focused onto the sample with a Fluar oil-immersion microscope objective (40×, NA = 1.3; Zeiss, Iena, Germany; Stoller et al. 2002). Images were created by scanning the sample with an x/y scanning piezo stage (P-542-2CD; Physik Instrumente, Karlsruhe, Germany). Fluorescence signal from the samples was epicollected through the objective; discriminated from the backscattered laser light and SHG signal by the combination of the
Downloaded from jdr.sagepub.com at Middle East Technical Univ on February 12, 2016 For personal use only. No other uses without permission. © International & American Associations for Dental Research 2016
3
Nonlinear Optical Signals for Caries Detection dichroic mirror, a 700-nm short-pass emission filter, and a 475-nm long-pass filter; and detected with a photomultiplier tube (R928P; Hamamatsu, Tokyo, Japan). The SHG signal was collected through a 1.4-NA oil-immersion condenser with a wide-pass CG-BG39 filter and a 447-nm high-performance band-pass filter and then detected by a H7422P photomultiplier operated in photon-counting mode (Hamamatsu). The method enables acquisition of in-depth Z-stacks of 2-dimensional images of the sample based on their nonlinear optical response (SHG and 2PEF emission) when excited in the infrared domain (Fig. 2). The image is created by scanning the sample point by point (similar to confocal scanning) and collecting the nonlinearly generated photons from each pixel individually. MPM images were obtained by integrating the spectral intensity emitted by the samples from 410 to 470 nm and from 475 to 700 nm for the SHG and 2PEF images, respectively, as illustrated by the blue and green boxes in Figure 2. Indeed, there is a thin disturbed surface layer with a mean thickness of about 2 µm. However, we realized different measurements in Z-scanning to check the possible influence of the smear layer: We scanned until a depth of 50 µm by step of 2 µm and noticed that 2 µm from the surface no disturbance in the images and the spectra could be evidenced. Accordingly, all the spectra and the images presented in this study were obtained within 2 µm from the surface. Images were compiled with the Image J software (http://rsb.info.nih.gov/ij/; National Institutes of Health, Bethesda, MD, USA) through 16-color lookup tables. Intensity profile plots of SHG images were calculated with the same software.
Nonlinear Optical Spectroscopy NLOS was realized in epicollection mode with the same experimental setup, with the adjunction of a spectrometer (Acton SP215i-CCD PIXIS 400B Excelon; Princeton Instruments, Trenton, NJ, USA) connected to the camera port of the microscope via an optical fiber. The backscattered laser light was filtered out with the same dichroic/700-nm short-pass filter combination (Raub et al. 2008). SHG and 2PEF are 2 nonlinear optical signals. Spectra were recorded with WinSpec software from Princeton Instruments. They were collected from 5 different samples of sound and carious dentin. All spectra were normalized following the laser power and the acquisition time. Data were plotted: the mean and the standard deviation were calculated and the statistical analyses were performed with the nonparametric Wilcoxon test. When the data were compared 2 × 2, a Wilcoxon test was used. A P value
