1354
OPTICS LETTERS / Vol. 40, No. 7 / April 1, 2015
Third-harmonic generation microscopy reveals dental anatomy in ancient fossils Yu-Cheng Chen,1 Szu-Yu Lee,2 Yana Wu,3 Kirstin Brink,4 Dar-Bin Shieh,3 Timothy D. Huang,5 Robert R. Reisz,4,6 and Chi-Kuang Sun1,2,* 1 2
3
Molecular Imaging Center, National Taiwan University, Taipei 10617, Taiwan
Department of Electrical Engineering and Graduate Institute of Photonics and Optoelectronics, National Taiwan University, Taipei City 10617, Taiwan
Institute of Oral Medicine, National Cheng Kung University College of Medicine and Hospital, Tainan City 701, Taiwan 4 Department of Biology, University of Toronto Mississauga, Mississauga, Ontario L5L 1C6, Canada 5
National Chung Hsing University, Taichung City 402, Taiwan
6
Department of Optics and Photonics, National Central University, Jhongli, Taoyuan City 320, Taiwan *Corresponding author: [email protected] Received October 24, 2014; revised February 12, 2015; accepted February 16, 2015; posted February 17, 2015 (Doc. ID 225362); published March 23, 2015
Fossil teeth are primary tools in the study of vertebrate evolution, but standard imaging modalities have not been capable of providing high-quality images in dentin, the main component of teeth, owing to small refractive index differences in the fossilized dentin. Our first attempt to use third-harmonic generation (THG) microscopy in fossil teeth has yielded significant submicrometer level anatomy, with an unexpectedly strong signal contrasting fossilized tubules from the surrounding dentin. Comparison between fossilized and extant teeth of crocodilians reveals a consistent evolutionary signature through time, indicating the great significance of THG microscopy in the evolutionary studies of dental anatomy in fossil teeth. © 2015 Optical Society of America OCIS codes: (180.4315) Nonlinear microscopy; (190.4160) Multiharmonic generation; (320.7110) Ultrafast nonlinear optics; (170.1850) Dentistry; (170.3880) Medical and biological imaging; (170.1420) Biology. http://dx.doi.org/10.1364/OL.40.001354
Modern paleontology investigates the evolution of organisms and communities through time, and their interactions in the biosphere. A particularly rich fossil record has allowed paleontologists to study in great detail more than 300 million years (Ma) of terrestrial vertebrate evolution. Of particular significance is their recent focus on reconstructing the feeding behavior of ancient vertebrates in an attempt to understand the evolution of communities of organisms, their trophic structure, and the overall structure of their ecosystem. The most common inferences for the study of feeding behavior are based on preserved hard tissues of the vertebrates, with fossil teeth providing the best indicators of diet. Most studies to date have focused on enamel wear or overall tooth morphology [1], and more recently tooth microstructures [2]. However, few studies have focused on the ultrastructure of dentin, despite their critical function in shaping and maintaining the overall integrity of the tooth [3–5]. A major obstacle in the study of dentin is in image modality, especially the image contrast. Recent advances in image modality [6–10] have been widely applied to both living and fossil teeth. Among all, optical microscopy is considered an ideal tool in imaging extant teeth (extracted teeth from living animals) for its noninvasiveness, convenience, and resolution. Nonetheless, no studies have been able to demonstrate clearly the anatomy of dentinal tubules in fossil teeth, while most modalities rely strongly on refractive index differences, staining, or sampledependent auto-fluorescence [9]. A method not yet used is third-harmonic generation (THG) microscopy. THG is a virtual-transition-based nonlinear process that deposits no energy [11] to the interacted specimens. Due to its third-order nonlinear nature, THG microscopy provides 0146-9592/15/071354-04$15.00/0
submicron resolution under reasonable depth and is extremely sensitive to fine structural variations [12–14]. It has been widely applied for noninvasive visualizations of 3D sub-cellular structures inside human [14–17] and living animals [18–20]. THG contrast was found to be enhanced by real electronic transitions in elastin [20], lipid [11,21,22], melanin [14], and hemoglobin [23]. In this Letter, we first apply THG microscopy to the study of fossil teeth of crocodilians and reveal that fossilization process in teeth provides remarkable difference in third-order nonlinear susceptibility χ 3 between the fossilized tubules and adjacent surrounding tissues. Thus, the enhanced THG contrast enables us to study fossilized tubules regardless of similar refractive indices [24,25] throughout the fossilized dentine. Taking advantage of the unique contrast, we investigate the dentinal tissues of an extant Alligator mississippiensis, and compare its anatomy to the 1.5 million-year-old fossilized tissues of the same species, and with those of a 93 million-year-old crocodilian from the Kem Kem Beds of Morrocco. Despite the difference in linear and nonlinear optical properties, correlation studies between extant and fossil teeth show a consistent morphology. This novel approach opens new research opportunities for paleontologists to resolve major issues in the evolutionary history of vertebrates by providing analyzable patterns and morphometric quantification of tubules, and allows for insights into the complex interactions between teeth and food within an evolutionary context [26]. THG, second-harmonic generation (SHG), and twophoton fluorescence (TPF) microscopies were simultaneously performed with a Cr:forsterite laser centered at 1230 nm with a 140-fs pulsewidth [20]. The laser output was focused 20 μm beneath the tooth surfaces by a © 2015 Optical Society of America
April 1, 2015 / Vol. 40, No. 7 / OPTICS LETTERS
0.9 NA objective with 50 mW power, except when otherwise specified. For comparison of extant and fossil teeth, crocodilians were chosen as our study organism since their fossil record spans over 90 million years, and Alligator mississippiensis is known from both fossil and extant individuals. In Fig. 1, we compare TPF, SHG, and THG images of extant and fossilized dentinal tubules in crocodilians. The PMT voltages were fixed at 900 V and 2 kV for TPF and SHG imaging modalities, respectively, while the voltage for THG had to be reduced to 700 V for both fossils, different from the 1 kV applied for THG imaging of the extant tooth. To our surprise, precise and consistent THG signals were observed in all fossil and extant teeth, allowing us to study fossilized dentinal tubules as in living taxa. Besides consistency, two other features help THG to distinguish dentinal tubules from the surrounding matrix: high image contrast and the high spatial resolution that is smaller than the tubule diameter. In comparison, TPF and SHG modalities lack the consistency to reflect the dentinal tubules in fossil dentin. Strong SHG signals found in extant tooth were in response to existing collagen fibers in dentin, while few SHG signals were observed in fossilized teeth, representing strain distribution [10]. In addition, TPF signals only appeared in the fossil teeth, probably as a result of mineral deposition. Hence, THG microscopy simply stands out as our major modality because of its clarity, excellent contrast, and detailed morphological visualization capability of dentin structures in fossil imaging. All extant speciments were skeltonized in a special “bug” room, using dermestid beetles. As such, all soft
Fig. 1. Multimodal imaging comparisons between extant and fossil teeth of crocodilians. (a)–(d) Images of an extant Alligator tooth acquired under (a) TPF, (b) SHG, (c) THG microscopies, and (d) overlay of three channels. (e)–(h) Images of a fossil Alligator (1.5 Ma) tooth under (e) TPF, (f) SHG, (g) THG microscopies, and (h) overlay of three channels. (i)–(l) Images of a fossil Kem Kem crocodilian (93 Ma) tooth under (i) TPF, (j) SHG), (k) THG microscopies, and (l) overlay of three channels. TPF, SHG, and THG are presented by blue, red, and yellow pseudocolors. All scale bars, 40 μm. Ma, millions of years. All laser power fixed at 50 mW. PMT voltages for (a), (e), (i): 900 V; for (b), (f), (j): 2000 V; for (c): 1000 V; for (g), (k): 700 V.
1355
tissues were consumed by the beetles over a period of about a week. After that, the specimens were frozen and then soaked in ammonia water to retard mold formation. Subsequently, the teeth were air dried thoroughly before being placed in the research collection, where they were kept under very dry conditions. Thus, all soft tissues of the alligator teeth, including cells like the odontoblasts and their extensions into the dentinal tubules, would have died and dried out, leaving only air in the tubules [27,28]. Hence, the refractive index (n) inside the tubules is close to 1 while its third-order nonlinear susceptibility χ 3 for THG generation inside this space is almost 0. The large difference in linear and nonlinear susceptibilities, between the vacant tubules and its surrounding tissues provided the required optical inhomogeneity for efficient THG generation in extant teeth. However, during fossilization, dentin tubules were filled with various minerals derived from groundwater after the decay of the odontoblasts [29], while the areas around the tubules are usually made of diagenetically modified hydroxyapatite (HA). As a result, the refractive indices in fossilized tubules [24] become very similar to HA [30], making it less possible to distinguish under linear mechanism. Similar to linear properties, THG inside the fossilized tubules should also be strongly modified by this fossilization process; it is thus important to investigate the THG similarity and difference between the morphology of extant and fossil teeth. To examine the correlations between the THG-imaged structures in fossil and living teeth and to study the fossilization-created THG property change in dentine tubules, the dentinal tissues of three crocodilian teeth spanning more than 90 million years were compared under THG microscopy. Here the dentin-enamel junction (DEJ) of the extant tooth, the 1.5 Ma fossil tooth, and the 93 Ma fossil tooth are shown in Figs. 2(a), 2(b), and 2(c).
Fig. 2. Comparison of the structure of fossil and extant crocodilian teeth under THG microscopy. (a)–(c) Dentin-enamel junction images of the (a) 93 Ma, (b) 1.5 Ma, and (c) extant crocodilian teeth. (d)–(f) High magnification images of the dentinal tubules of the (a) 93 Ma, (b) 1.5 Ma, and (c) extant crocodilian teeth. (g) and (h) Laser power dependent THG signal of the (g) fossil tooth (1.5 Ma) and (h) extant tooth under 10, 20, 30, 40, 80, and 100 mW applied average power on the sample surface at a similar anatomical region using the same PMT with the same 700-V bias voltage. All scale bars, 10 μm. Ma, millions of years.
1356
OPTICS LETTERS / Vol. 40, No. 7 / April 1, 2015
Table 1. Comparisons of Tubule Characteristics Items Alligator (Extant) Alligator (1.5 Ma) Kem Kem Crocodile (93 Ma)
Diameter (μm)
Branching (#/100 um2 )
Density (%/image)
0.99 0.94 0.93
2.10 2.12 2.20
38.8 37.4 35.2
Despite diagenetic alteration, the morphology and organization of the dentinal tubules in both fossil and extant specimens were surprisingly consistent under THG microscopy. Different from the hollow tubules in extant teeth [Fig. 2(f)], solid tubules can be clearly distinguished in Figs. 2(d) and 2(e) in the fossilized teeth. As summarized in Table 1, the patterns of distribution are alike in terms of tubule diameter, branching rate, and density, presenting similar data and strong correlations. This strong morphological correlation not only confirms that dentinal tubule structures have remained relatively constant in crocodilians across more than 90 Ma, but also gives us confidence to state that the THG-imaged tubule structures in fossil vertebrates represents the morphology of the tubules before fossilization. This result demonstrates that the physical structures in fossil teeth may be more reliable than chemical analyses since chemicals can be altered through diagenesis [29]. Interestingly, fossilized tubules appear to produce much stronger THG signals compared to extant tubules, thus providing a new tool for identifying dentinal tubules in fossil teeth. By measuring carefully with the THG microscope, the power-dependent study of the fossil tooth (1.5 Ma) and the extant tooth are presented in Figs. 2(g) and 2(h), respectively. Our result indicates that 10 mW of excitation laser power is enough to image the fossil teeth, while the extant teeth require at least 40 mW excitation laser power under the same microscopy setting including the PMT bias. Combined with the fact that THG intensity is cubic-dependent to excitation power, our study thus indicates a 64-fold stronger THG signal in fossilized versus non-fossilized dentinal tubules under the same incident power. A widely accepted analytical model [12,22] was used to estimate the relative THG χ 3 value. Now we consider a Gaussian beam propagating in the z direction and tightly focused right at the center of a tubule inside a specimen. The electric field of the fundamental wave E ω can be expressed as [31] Aω x2 y2 ; (1) exp − 2 E ω x; y; z 1 i zb wω0 1 i zb where Aω is related to field amplitude, w0 is the beam waist of the fundamental beam, and the confocal parameter 2b is defined with b kw20 ∕2 πnw20 ∕λ. By adopting the trial solution, the electric field of the propagating THG wave E 3ω can be expressed as A3ω z x2 y2 ; (2) exp − 2 E 3ω x; y; z 1 i zb w3ω0 1 i zb with an assumption that w23ω0 w2ω0 ∕3 and b3ω bω b (i.e., n n3 n). By following the THG derivation with:
P NL 3ω 3ω
1 ε0 χ 3 E 3ω ; 22
(3)
where χ 3 is the third-order nonlinear susceptibility for THG, the THG power I 3ω can be approximated as I 3ω ∝ I 3ω jJ 3 j2 ;
(4)
where the integral J 3 is defined by Z J 3 Δk; z
χ 3 z0 iΔkz0 0 e dz ; 0 2 −∞ 1 i z b z
(5)
in which J 3 is the phase matching integral of THG under a tightly focused condition, and Δk 3kω − k3ω is the phase mismatch. In the case for an extant tooth, we assume that the χ 3 value of the HA surrounding the 3 value of the empty tubules is denoted as χ 3 ha , and the χ (air) tubules is approximately 0. For the case in a fossil tooth, we assume that the χ 3 value of the HA surround3 ing the tubules to be the same as χ 3 ha ha, while the χ value inside the fossilized tubules is not zero and is denoted as χ 3 fos . Therefore, the THG power when focused in the center of the tubules can be compared by following Eq. (5). By substituting the respective refractive indices nha , nair and the presumed nonlinear susceptibilities χ 3 ha , 3 3 χ air , χ fos into both cases, the power of THG signal generated in extant tubules (I 3ω;ext ) and fossilized tubules (I 3ω;fossil ) can be estimated as 2 d d d d I 3ω;ext ∞I 3ω J ha −∞; J air − ; J ha − ; ∞ ; 2 2 2 2 (6)
2 d dd d I 3ω;fossil ∞I 3ω J ha −∞; J fos − ; J ha − ;∞ ; 2 2 2 2 (7) where the annotations ha, air, and fos stand for hydroxyapatite, air, and fossil, and d is the tubule diameter. Here d was set to be ∼1 μm based on our observation as shown in Figs. 1 and 2. Note that Δk of HA was estimated at ω and 3ω based on the previous published data [27,32] and by assuming n 1.651 − 3.46 10−6 λ (λ:nm). Based on our empirical observation in Figs. 2(g) and 2(h), here the theoretical calculation implies that the THG χ 3 value in the mineralized dentinal tubule is 9 times higher than its corresponding value in the regular dentine tissues consisting mostly of HA. This result indicates that it is the outstanding nonlinear susceptibility χ 3 of the mineral fillings, which provides the required THG property difference between tubules and its surroundings, responsible for producing the observed
April 1, 2015 / Vol. 40, No. 7 / OPTICS LETTERS
astonishingly strong THG signals contrasting the tubules inside the fossil teeth. The above analytical model treats the round tubule as a thin film, which might cause estimation error. Our experimental condition indicates a w3ω0 value of 290 nm, which is much narrower than the tubule diameter and justifies the simple model. By shortening the interaction length d, we found that our result is independent of the interaction length. As for the unknown refractive index inside fossil tubules, we assumed that it is the same as surrounding. This would lead to errors in Δk inside the tubules. By setting the Δk values of fossil tubules as 0 up to 4 times the value of the hydroxyapatite, we found that the χ 3 fos value would have at most 1.6% change, indicating the insensitivity of our result to the refractive index value of the fossil tubules. The final estimation error could be coming from the assumption that after fossilization, the χ 3 ha value remains the same. Our observation of the strong THG in fossils indicates the enhancement of THG by diagenetic modification. Due to the fact that our observed THG signal intensity inside dentine tubules is determined by the difference between the THG χ 3 values inside and outside the fossilized dentine tubules, the underestimation of the χ 3 ha value after fossilization make us underestimate the χ 3 fos value in fossilized tubules. Correctly speaking, the χ 3 fos value in fossilized tubule should be even larger than 9 times of the corresponding value χ 3 ha in the extant dentine tissues consisting mostly of hydroxyapatite. In conclusion, our report represents the first successful imaging and clear visualization of dentinal microstructures in fossil teeth. The strong THG signal contrast created inside the fossilized tubules allows us to examine dental anatomy in the same way as in living species. By providing reliable and consistent signals in fossil teeth, THG microscopy serves as an ideal new tool for studying dentition and dental evolution in vertebrates. Furthermore, the close correlations in dental anatomy that were demonstrated here offer new avenues for investigating vertebrate evolution. We thank D. C. Evans and K. Seymour for access to collections of the Royal Ontario Museum. This work is supported by MOST, MOE, and NHRI of Taiwan, under NHRI-EX101-9936EI, MOST 103-2221-E-137-MY3, and NSC 102-3011-P-002-010, and by the NSERC Discovery Grant and University of Toronto to RRR, in Canada. References 1. G. M. Erickson, B. A. Krick, M. Hamilton, G. R. Bourne, M. A. Norell, E. Lilleodden, and W. G. Sawyer, Science 338, 98 (2012). 2. K. S. Brink and R. R. Reisz, Nat. Commun. 5, 3269 (2014).
1357
3. A. R. H. LeBlanc and R. R. Reisz, PLoS ONE 8, e74697 (2013). 4. C. Sullivan, R. R. Reisz, and R. M. H. Smith, Proc. R. Soc. B 270, 173 (2003). 5. V. Imbeni, J. J. Kruzic, G. W. Marshall, S. J. Marshall, and R. O. Ritchie, Nat. Mater. 4, 229 (2005). 6. D. C. Kalthoff, J. Morphol. 272, 641 (2011). 7. P. C. J. Donoghue, S. Bengtson, X. P. Dong, and N. J. Gostling, Nature 442, 680 (2006). 8. R. S. Scott, P. S. Ungar, T. S. Bergstrom, C. A. Brown, F. E. Grine, M. F. Teaford, and A. Walker, Nature 436, 693 (2005). 9. M. H. Chen, W. L. Chen, P. T. Fwu, and C. Y. Dong, J. Biomed. Opt. 12, 064018 (2007). 10. S. Y. Chen, C. Y. S. Hsu, and C.-K. Sun, Opt. Express 16, 11670 (2008). 11. C.-K. Sun, S. W. Chu, S. Y. Chen, T. M. Liu, C. Y. Lin, and H. J. Tsai, J. Struct. Biol. 147, 19 (2004). 12. R. W. Boyd, Nonlinear Optics, 3rd ed. (Academic, 2008). 13. D. Yelin and Y. Silberberg, Opt. Express 5, 169 (1999). 14. S. Y. Chen, S. U. Chen, H. Y. Wu, W. J. Lee, Y. H. Liao, and C.-K. Sun, IEEE J. Sel. Top. Quantum Electron. 16, 478 (2010). 15. M.-R. Tsai, S. Y. Chen, D.-B. Shieh, P. J. Lou, and C.-K. Sun, Biomed. Opt. Express 2, 2317 (2011). 16. Y. H. Liao, S. Y. Chen, S. Y. Chou, M. R. Tsai, and C. K. Sun, Biomed. Opt. Express 4, 77 (2013). 17. M. R. Tsai, Y. H. Cheng, J.-S. Chen, Y.-S. Sheen, Y. H. Liao, and C.-K. Sun, J. Biomed. Opt. 19, 036001 (2014). 18. C.-S. Hsieh, S. U. Chen, Y. W. Lee, Y.-S. Yang, and C. K. Sun, Opt. Express 16, 11574 (2008). 19. S. P. Tai, W. J. Lee, D. B. Shieh, H. Y. Huang, C. H. Yu, and C. K. Sun, Opt. Express 14, 6178 (2006). 20. C.-H. Yu, S.-P. Tai, C.-T. Kung, I.-J. Wang, H.-C. Yu, H.-J. Huang, W.-J. Lee, Y.-F. Chan, and C.-K. Sun, Opt. Express 15, 11167 (2007). 21. D. Débarre, W. Supatto, A.-M. Pena, A. Fabre, T. Tordjmann, L. Combettes, M.-C. Schanne-Klein, and E. Beaurepaire, Nat. Methods 3, 47 (2006). 22. M. Müller, J. Squier, K. R. Wilson, and G. J. Brakenhoff, J. Microsc. 191, 266 (1998). 23. C. F. Chang, C. H. Yu, and C.-K. Sun, J. Biophotonics 3, 678 (2010). 24. H. Shinoda, Chem. Today (in Japanese) 162, 43 (1984). 25. A. Boyde, J. Dental Research 58, 981 (1979). 26. R. R. Reisz, Cell TREE 12, 218 (1997). 27. Y. Dauphin and C. T. Williams, CrystEngComm 9, 1252 (2007). 28. J. Enax, H. O. Fabritius, A. Rack, O. Prymak, D. Raabe, and M. Epple, J. Struct. Biol. 184, 155 (2013). 29. M. J. Kohn, M. J. Schoeninger, and W. W. Barker, Geochim. Cosmochim. Acta 63, 2737 (1999). 30. D. Spitzer and J. T. Bosch, Calcif. Tissue Res. 17, 129 (1975). 31. H. A. Haus, Waves and Fields in Optoelectronics (PrenticeHall, 1984). 32. B. W. Darwell, Material Science for Dentistry, 9th ed. (Woodhead, 2009).
OPTICS LETTERS / Vol. 40, No. 7 / April 1, 2015
Third-harmonic generation microscopy reveals dental anatomy in ancient fossils Yu-Cheng Chen,1 Szu-Yu Lee,2 Yana Wu,3 Kirstin Brink,4 Dar-Bin Shieh,3 Timothy D. Huang,5 Robert R. Reisz,4,6 and Chi-Kuang Sun1,2,* 1 2
3
Molecular Imaging Center, National Taiwan University, Taipei 10617, Taiwan
Department of Electrical Engineering and Graduate Institute of Photonics and Optoelectronics, National Taiwan University, Taipei City 10617, Taiwan
Institute of Oral Medicine, National Cheng Kung University College of Medicine and Hospital, Tainan City 701, Taiwan 4 Department of Biology, University of Toronto Mississauga, Mississauga, Ontario L5L 1C6, Canada 5
National Chung Hsing University, Taichung City 402, Taiwan
6
Department of Optics and Photonics, National Central University, Jhongli, Taoyuan City 320, Taiwan *Corresponding author: [email protected] Received October 24, 2014; revised February 12, 2015; accepted February 16, 2015; posted February 17, 2015 (Doc. ID 225362); published March 23, 2015
Fossil teeth are primary tools in the study of vertebrate evolution, but standard imaging modalities have not been capable of providing high-quality images in dentin, the main component of teeth, owing to small refractive index differences in the fossilized dentin. Our first attempt to use third-harmonic generation (THG) microscopy in fossil teeth has yielded significant submicrometer level anatomy, with an unexpectedly strong signal contrasting fossilized tubules from the surrounding dentin. Comparison between fossilized and extant teeth of crocodilians reveals a consistent evolutionary signature through time, indicating the great significance of THG microscopy in the evolutionary studies of dental anatomy in fossil teeth. © 2015 Optical Society of America OCIS codes: (180.4315) Nonlinear microscopy; (190.4160) Multiharmonic generation; (320.7110) Ultrafast nonlinear optics; (170.1850) Dentistry; (170.3880) Medical and biological imaging; (170.1420) Biology. http://dx.doi.org/10.1364/OL.40.001354
Modern paleontology investigates the evolution of organisms and communities through time, and their interactions in the biosphere. A particularly rich fossil record has allowed paleontologists to study in great detail more than 300 million years (Ma) of terrestrial vertebrate evolution. Of particular significance is their recent focus on reconstructing the feeding behavior of ancient vertebrates in an attempt to understand the evolution of communities of organisms, their trophic structure, and the overall structure of their ecosystem. The most common inferences for the study of feeding behavior are based on preserved hard tissues of the vertebrates, with fossil teeth providing the best indicators of diet. Most studies to date have focused on enamel wear or overall tooth morphology [1], and more recently tooth microstructures [2]. However, few studies have focused on the ultrastructure of dentin, despite their critical function in shaping and maintaining the overall integrity of the tooth [3–5]. A major obstacle in the study of dentin is in image modality, especially the image contrast. Recent advances in image modality [6–10] have been widely applied to both living and fossil teeth. Among all, optical microscopy is considered an ideal tool in imaging extant teeth (extracted teeth from living animals) for its noninvasiveness, convenience, and resolution. Nonetheless, no studies have been able to demonstrate clearly the anatomy of dentinal tubules in fossil teeth, while most modalities rely strongly on refractive index differences, staining, or sampledependent auto-fluorescence [9]. A method not yet used is third-harmonic generation (THG) microscopy. THG is a virtual-transition-based nonlinear process that deposits no energy [11] to the interacted specimens. Due to its third-order nonlinear nature, THG microscopy provides 0146-9592/15/071354-04$15.00/0
submicron resolution under reasonable depth and is extremely sensitive to fine structural variations [12–14]. It has been widely applied for noninvasive visualizations of 3D sub-cellular structures inside human [14–17] and living animals [18–20]. THG contrast was found to be enhanced by real electronic transitions in elastin [20], lipid [11,21,22], melanin [14], and hemoglobin [23]. In this Letter, we first apply THG microscopy to the study of fossil teeth of crocodilians and reveal that fossilization process in teeth provides remarkable difference in third-order nonlinear susceptibility χ 3 between the fossilized tubules and adjacent surrounding tissues. Thus, the enhanced THG contrast enables us to study fossilized tubules regardless of similar refractive indices [24,25] throughout the fossilized dentine. Taking advantage of the unique contrast, we investigate the dentinal tissues of an extant Alligator mississippiensis, and compare its anatomy to the 1.5 million-year-old fossilized tissues of the same species, and with those of a 93 million-year-old crocodilian from the Kem Kem Beds of Morrocco. Despite the difference in linear and nonlinear optical properties, correlation studies between extant and fossil teeth show a consistent morphology. This novel approach opens new research opportunities for paleontologists to resolve major issues in the evolutionary history of vertebrates by providing analyzable patterns and morphometric quantification of tubules, and allows for insights into the complex interactions between teeth and food within an evolutionary context [26]. THG, second-harmonic generation (SHG), and twophoton fluorescence (TPF) microscopies were simultaneously performed with a Cr:forsterite laser centered at 1230 nm with a 140-fs pulsewidth [20]. The laser output was focused 20 μm beneath the tooth surfaces by a © 2015 Optical Society of America
April 1, 2015 / Vol. 40, No. 7 / OPTICS LETTERS
0.9 NA objective with 50 mW power, except when otherwise specified. For comparison of extant and fossil teeth, crocodilians were chosen as our study organism since their fossil record spans over 90 million years, and Alligator mississippiensis is known from both fossil and extant individuals. In Fig. 1, we compare TPF, SHG, and THG images of extant and fossilized dentinal tubules in crocodilians. The PMT voltages were fixed at 900 V and 2 kV for TPF and SHG imaging modalities, respectively, while the voltage for THG had to be reduced to 700 V for both fossils, different from the 1 kV applied for THG imaging of the extant tooth. To our surprise, precise and consistent THG signals were observed in all fossil and extant teeth, allowing us to study fossilized dentinal tubules as in living taxa. Besides consistency, two other features help THG to distinguish dentinal tubules from the surrounding matrix: high image contrast and the high spatial resolution that is smaller than the tubule diameter. In comparison, TPF and SHG modalities lack the consistency to reflect the dentinal tubules in fossil dentin. Strong SHG signals found in extant tooth were in response to existing collagen fibers in dentin, while few SHG signals were observed in fossilized teeth, representing strain distribution [10]. In addition, TPF signals only appeared in the fossil teeth, probably as a result of mineral deposition. Hence, THG microscopy simply stands out as our major modality because of its clarity, excellent contrast, and detailed morphological visualization capability of dentin structures in fossil imaging. All extant speciments were skeltonized in a special “bug” room, using dermestid beetles. As such, all soft
Fig. 1. Multimodal imaging comparisons between extant and fossil teeth of crocodilians. (a)–(d) Images of an extant Alligator tooth acquired under (a) TPF, (b) SHG, (c) THG microscopies, and (d) overlay of three channels. (e)–(h) Images of a fossil Alligator (1.5 Ma) tooth under (e) TPF, (f) SHG, (g) THG microscopies, and (h) overlay of three channels. (i)–(l) Images of a fossil Kem Kem crocodilian (93 Ma) tooth under (i) TPF, (j) SHG), (k) THG microscopies, and (l) overlay of three channels. TPF, SHG, and THG are presented by blue, red, and yellow pseudocolors. All scale bars, 40 μm. Ma, millions of years. All laser power fixed at 50 mW. PMT voltages for (a), (e), (i): 900 V; for (b), (f), (j): 2000 V; for (c): 1000 V; for (g), (k): 700 V.
1355
tissues were consumed by the beetles over a period of about a week. After that, the specimens were frozen and then soaked in ammonia water to retard mold formation. Subsequently, the teeth were air dried thoroughly before being placed in the research collection, where they were kept under very dry conditions. Thus, all soft tissues of the alligator teeth, including cells like the odontoblasts and their extensions into the dentinal tubules, would have died and dried out, leaving only air in the tubules [27,28]. Hence, the refractive index (n) inside the tubules is close to 1 while its third-order nonlinear susceptibility χ 3 for THG generation inside this space is almost 0. The large difference in linear and nonlinear susceptibilities, between the vacant tubules and its surrounding tissues provided the required optical inhomogeneity for efficient THG generation in extant teeth. However, during fossilization, dentin tubules were filled with various minerals derived from groundwater after the decay of the odontoblasts [29], while the areas around the tubules are usually made of diagenetically modified hydroxyapatite (HA). As a result, the refractive indices in fossilized tubules [24] become very similar to HA [30], making it less possible to distinguish under linear mechanism. Similar to linear properties, THG inside the fossilized tubules should also be strongly modified by this fossilization process; it is thus important to investigate the THG similarity and difference between the morphology of extant and fossil teeth. To examine the correlations between the THG-imaged structures in fossil and living teeth and to study the fossilization-created THG property change in dentine tubules, the dentinal tissues of three crocodilian teeth spanning more than 90 million years were compared under THG microscopy. Here the dentin-enamel junction (DEJ) of the extant tooth, the 1.5 Ma fossil tooth, and the 93 Ma fossil tooth are shown in Figs. 2(a), 2(b), and 2(c).
Fig. 2. Comparison of the structure of fossil and extant crocodilian teeth under THG microscopy. (a)–(c) Dentin-enamel junction images of the (a) 93 Ma, (b) 1.5 Ma, and (c) extant crocodilian teeth. (d)–(f) High magnification images of the dentinal tubules of the (a) 93 Ma, (b) 1.5 Ma, and (c) extant crocodilian teeth. (g) and (h) Laser power dependent THG signal of the (g) fossil tooth (1.5 Ma) and (h) extant tooth under 10, 20, 30, 40, 80, and 100 mW applied average power on the sample surface at a similar anatomical region using the same PMT with the same 700-V bias voltage. All scale bars, 10 μm. Ma, millions of years.
1356
OPTICS LETTERS / Vol. 40, No. 7 / April 1, 2015
Table 1. Comparisons of Tubule Characteristics Items Alligator (Extant) Alligator (1.5 Ma) Kem Kem Crocodile (93 Ma)
Diameter (μm)
Branching (#/100 um2 )
Density (%/image)
0.99 0.94 0.93
2.10 2.12 2.20
38.8 37.4 35.2
Despite diagenetic alteration, the morphology and organization of the dentinal tubules in both fossil and extant specimens were surprisingly consistent under THG microscopy. Different from the hollow tubules in extant teeth [Fig. 2(f)], solid tubules can be clearly distinguished in Figs. 2(d) and 2(e) in the fossilized teeth. As summarized in Table 1, the patterns of distribution are alike in terms of tubule diameter, branching rate, and density, presenting similar data and strong correlations. This strong morphological correlation not only confirms that dentinal tubule structures have remained relatively constant in crocodilians across more than 90 Ma, but also gives us confidence to state that the THG-imaged tubule structures in fossil vertebrates represents the morphology of the tubules before fossilization. This result demonstrates that the physical structures in fossil teeth may be more reliable than chemical analyses since chemicals can be altered through diagenesis [29]. Interestingly, fossilized tubules appear to produce much stronger THG signals compared to extant tubules, thus providing a new tool for identifying dentinal tubules in fossil teeth. By measuring carefully with the THG microscope, the power-dependent study of the fossil tooth (1.5 Ma) and the extant tooth are presented in Figs. 2(g) and 2(h), respectively. Our result indicates that 10 mW of excitation laser power is enough to image the fossil teeth, while the extant teeth require at least 40 mW excitation laser power under the same microscopy setting including the PMT bias. Combined with the fact that THG intensity is cubic-dependent to excitation power, our study thus indicates a 64-fold stronger THG signal in fossilized versus non-fossilized dentinal tubules under the same incident power. A widely accepted analytical model [12,22] was used to estimate the relative THG χ 3 value. Now we consider a Gaussian beam propagating in the z direction and tightly focused right at the center of a tubule inside a specimen. The electric field of the fundamental wave E ω can be expressed as [31] Aω x2 y2 ; (1) exp − 2 E ω x; y; z 1 i zb wω0 1 i zb where Aω is related to field amplitude, w0 is the beam waist of the fundamental beam, and the confocal parameter 2b is defined with b kw20 ∕2 πnw20 ∕λ. By adopting the trial solution, the electric field of the propagating THG wave E 3ω can be expressed as A3ω z x2 y2 ; (2) exp − 2 E 3ω x; y; z 1 i zb w3ω0 1 i zb with an assumption that w23ω0 w2ω0 ∕3 and b3ω bω b (i.e., n n3 n). By following the THG derivation with:
P NL 3ω 3ω
1 ε0 χ 3 E 3ω ; 22
(3)
where χ 3 is the third-order nonlinear susceptibility for THG, the THG power I 3ω can be approximated as I 3ω ∝ I 3ω jJ 3 j2 ;
(4)
where the integral J 3 is defined by Z J 3 Δk; z
χ 3 z0 iΔkz0 0 e dz ; 0 2 −∞ 1 i z b z
(5)
in which J 3 is the phase matching integral of THG under a tightly focused condition, and Δk 3kω − k3ω is the phase mismatch. In the case for an extant tooth, we assume that the χ 3 value of the HA surrounding the 3 value of the empty tubules is denoted as χ 3 ha , and the χ (air) tubules is approximately 0. For the case in a fossil tooth, we assume that the χ 3 value of the HA surround3 ing the tubules to be the same as χ 3 ha ha, while the χ value inside the fossilized tubules is not zero and is denoted as χ 3 fos . Therefore, the THG power when focused in the center of the tubules can be compared by following Eq. (5). By substituting the respective refractive indices nha , nair and the presumed nonlinear susceptibilities χ 3 ha , 3 3 χ air , χ fos into both cases, the power of THG signal generated in extant tubules (I 3ω;ext ) and fossilized tubules (I 3ω;fossil ) can be estimated as 2 d d d d I 3ω;ext ∞I 3ω J ha −∞; J air − ; J ha − ; ∞ ; 2 2 2 2 (6)
2 d dd d I 3ω;fossil ∞I 3ω J ha −∞; J fos − ; J ha − ;∞ ; 2 2 2 2 (7) where the annotations ha, air, and fos stand for hydroxyapatite, air, and fossil, and d is the tubule diameter. Here d was set to be ∼1 μm based on our observation as shown in Figs. 1 and 2. Note that Δk of HA was estimated at ω and 3ω based on the previous published data [27,32] and by assuming n 1.651 − 3.46 10−6 λ (λ:nm). Based on our empirical observation in Figs. 2(g) and 2(h), here the theoretical calculation implies that the THG χ 3 value in the mineralized dentinal tubule is 9 times higher than its corresponding value in the regular dentine tissues consisting mostly of HA. This result indicates that it is the outstanding nonlinear susceptibility χ 3 of the mineral fillings, which provides the required THG property difference between tubules and its surroundings, responsible for producing the observed
April 1, 2015 / Vol. 40, No. 7 / OPTICS LETTERS
astonishingly strong THG signals contrasting the tubules inside the fossil teeth. The above analytical model treats the round tubule as a thin film, which might cause estimation error. Our experimental condition indicates a w3ω0 value of 290 nm, which is much narrower than the tubule diameter and justifies the simple model. By shortening the interaction length d, we found that our result is independent of the interaction length. As for the unknown refractive index inside fossil tubules, we assumed that it is the same as surrounding. This would lead to errors in Δk inside the tubules. By setting the Δk values of fossil tubules as 0 up to 4 times the value of the hydroxyapatite, we found that the χ 3 fos value would have at most 1.6% change, indicating the insensitivity of our result to the refractive index value of the fossil tubules. The final estimation error could be coming from the assumption that after fossilization, the χ 3 ha value remains the same. Our observation of the strong THG in fossils indicates the enhancement of THG by diagenetic modification. Due to the fact that our observed THG signal intensity inside dentine tubules is determined by the difference between the THG χ 3 values inside and outside the fossilized dentine tubules, the underestimation of the χ 3 ha value after fossilization make us underestimate the χ 3 fos value in fossilized tubules. Correctly speaking, the χ 3 fos value in fossilized tubule should be even larger than 9 times of the corresponding value χ 3 ha in the extant dentine tissues consisting mostly of hydroxyapatite. In conclusion, our report represents the first successful imaging and clear visualization of dentinal microstructures in fossil teeth. The strong THG signal contrast created inside the fossilized tubules allows us to examine dental anatomy in the same way as in living species. By providing reliable and consistent signals in fossil teeth, THG microscopy serves as an ideal new tool for studying dentition and dental evolution in vertebrates. Furthermore, the close correlations in dental anatomy that were demonstrated here offer new avenues for investigating vertebrate evolution. We thank D. C. Evans and K. Seymour for access to collections of the Royal Ontario Museum. This work is supported by MOST, MOE, and NHRI of Taiwan, under NHRI-EX101-9936EI, MOST 103-2221-E-137-MY3, and NSC 102-3011-P-002-010, and by the NSERC Discovery Grant and University of Toronto to RRR, in Canada. References 1. G. M. Erickson, B. A. Krick, M. Hamilton, G. R. Bourne, M. A. Norell, E. Lilleodden, and W. G. Sawyer, Science 338, 98 (2012). 2. K. S. Brink and R. R. Reisz, Nat. Commun. 5, 3269 (2014).
1357
3. A. R. H. LeBlanc and R. R. Reisz, PLoS ONE 8, e74697 (2013). 4. C. Sullivan, R. R. Reisz, and R. M. H. Smith, Proc. R. Soc. B 270, 173 (2003). 5. V. Imbeni, J. J. Kruzic, G. W. Marshall, S. J. Marshall, and R. O. Ritchie, Nat. Mater. 4, 229 (2005). 6. D. C. Kalthoff, J. Morphol. 272, 641 (2011). 7. P. C. J. Donoghue, S. Bengtson, X. P. Dong, and N. J. Gostling, Nature 442, 680 (2006). 8. R. S. Scott, P. S. Ungar, T. S. Bergstrom, C. A. Brown, F. E. Grine, M. F. Teaford, and A. Walker, Nature 436, 693 (2005). 9. M. H. Chen, W. L. Chen, P. T. Fwu, and C. Y. Dong, J. Biomed. Opt. 12, 064018 (2007). 10. S. Y. Chen, C. Y. S. Hsu, and C.-K. Sun, Opt. Express 16, 11670 (2008). 11. C.-K. Sun, S. W. Chu, S. Y. Chen, T. M. Liu, C. Y. Lin, and H. J. Tsai, J. Struct. Biol. 147, 19 (2004). 12. R. W. Boyd, Nonlinear Optics, 3rd ed. (Academic, 2008). 13. D. Yelin and Y. Silberberg, Opt. Express 5, 169 (1999). 14. S. Y. Chen, S. U. Chen, H. Y. Wu, W. J. Lee, Y. H. Liao, and C.-K. Sun, IEEE J. Sel. Top. Quantum Electron. 16, 478 (2010). 15. M.-R. Tsai, S. Y. Chen, D.-B. Shieh, P. J. Lou, and C.-K. Sun, Biomed. Opt. Express 2, 2317 (2011). 16. Y. H. Liao, S. Y. Chen, S. Y. Chou, M. R. Tsai, and C. K. Sun, Biomed. Opt. Express 4, 77 (2013). 17. M. R. Tsai, Y. H. Cheng, J.-S. Chen, Y.-S. Sheen, Y. H. Liao, and C.-K. Sun, J. Biomed. Opt. 19, 036001 (2014). 18. C.-S. Hsieh, S. U. Chen, Y. W. Lee, Y.-S. Yang, and C. K. Sun, Opt. Express 16, 11574 (2008). 19. S. P. Tai, W. J. Lee, D. B. Shieh, H. Y. Huang, C. H. Yu, and C. K. Sun, Opt. Express 14, 6178 (2006). 20. C.-H. Yu, S.-P. Tai, C.-T. Kung, I.-J. Wang, H.-C. Yu, H.-J. Huang, W.-J. Lee, Y.-F. Chan, and C.-K. Sun, Opt. Express 15, 11167 (2007). 21. D. Débarre, W. Supatto, A.-M. Pena, A. Fabre, T. Tordjmann, L. Combettes, M.-C. Schanne-Klein, and E. Beaurepaire, Nat. Methods 3, 47 (2006). 22. M. Müller, J. Squier, K. R. Wilson, and G. J. Brakenhoff, J. Microsc. 191, 266 (1998). 23. C. F. Chang, C. H. Yu, and C.-K. Sun, J. Biophotonics 3, 678 (2010). 24. H. Shinoda, Chem. Today (in Japanese) 162, 43 (1984). 25. A. Boyde, J. Dental Research 58, 981 (1979). 26. R. R. Reisz, Cell TREE 12, 218 (1997). 27. Y. Dauphin and C. T. Williams, CrystEngComm 9, 1252 (2007). 28. J. Enax, H. O. Fabritius, A. Rack, O. Prymak, D. Raabe, and M. Epple, J. Struct. Biol. 184, 155 (2013). 29. M. J. Kohn, M. J. Schoeninger, and W. W. Barker, Geochim. Cosmochim. Acta 63, 2737 (1999). 30. D. Spitzer and J. T. Bosch, Calcif. Tissue Res. 17, 129 (1975). 31. H. A. Haus, Waves and Fields in Optoelectronics (PrenticeHall, 1984). 32. B. W. Darwell, Material Science for Dentistry, 9th ed. (Woodhead, 2009).
