1 DIFFERENTIAL LASER INDUCED PERTURBATION SPECTROSCOPY FOR ANALYSIS OF BIOLOGICAL MATERIAL By SARAH ELIZABETH SMITH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUI REMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Sarah Elizabeth Smith
3 ACKNOWLEDGMENTS I thank my advisor, Dr. David. W. Hahn, for his guidance and encouragement throughout my academic career at the University of Florida. Thanks also goes out to Dr. Omenetto and Dr. Mikolaitis for their input and Dr. Angelini for joining my committee late in my studies. Finally, thanks go out to Ray and Jen for their help with the animals and Dr. Castlemen for his he lp with pathology.
4 TABLE OF CONTENTS page ACKNOW LEDGMENTS ................................ ................................ ................................ ............... 3 LIST OF TABLES ................................ ................................ ................................ ........................... 6 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 14 The Properties of Light, Lasers, and Matter ................................ ................................ ........... 14 Properties of Light ................................ ................................ ................................ ........... 14 Properties of Lasers ................................ ................................ ................................ ......... 15 Properties of Matter ................................ ................................ ................................ ......... 16 The Interaction of Light and Matter ................................ ................................ ....................... 22 Light and Matter Interactions on a Bulk Level ................................ ............................... 22 Light and Matter Interactions on a Molecular Level ................................ ....................... 25 Light and Tissue Interactions ................................ ................................ .......................... 27 The Interaction of Lasers and Bulk Matter ................................ ................................ ...... 29 The Interaction of Lasers and Tissue ................................ ................................ ...................... 33 Photochemical Effects ................................ ................................ ................................ ..... 33 Photomechanical Effects ................................ ................................ ................................ 34 Photothermal Effects ................................ ................................ ................................ ....... 35 Coag ulation of t issue ................................ ................................ ................................ 37 Vaporization of t issue ................................ ................................ ............................... 38 Carbonization o f t issue ................................ ................................ ............................. 38 Ablation of t issue ................................ ................................ ................................ ..... 38 L aser Tissue Interaction Examples ................................ ................................ ................. 43 2 SPECTROSCOPIC METHODS ................................ ................................ ............................ 45 Vibrational Spectroscopy ................................ ................................ ................................ ........ 45 Infrared Spectroscopy ................................ ................................ ................................ ...... 48 Raman Spectroscopy ................................ ................................ ................................ ....... 50 Raman m icro spectroscopy ................................ ................................ ...................... 52 Resonance Raman s pectroscopy ................................ ................................ .............. 53 Surface E nhanced Raman s pectroscopy ................................ ................................ .. 56 Raman s pectroscopy of p eptides and p roteins ................................ ......................... 57 Raman s pectroscopy of t issue ................................ ................................ .................. 60 Coupled Raman Spectroscopy and Infrared Spectroscopy ................................ ............. 63 Electronic Spectroscopy ................................ ................................ ................................ ......... 64 Ultraviolet and Visible Spectroscopy ................................ ................................ .............. 65
5 Chrioptic al Spectroscopy ................................ ................................ ................................ 66 Fluorescence Spectroscopy ................................ ................................ ............................. 66 Fluorescence e xcitation e mission m atrix s pectroscopy ................................ ........... 68 Fluorescence r ecovery after p hotobleaching s pectroscopy ................................ ...... 70 Laser I nduced f luorescence s pectroscopy ................................ ................................ 71 Fluorescence s pectroscopy using f luorophore m anipulation ................................ ... 72 Fluorescence s pectroscopy of t issue ................................ ................................ ........ 73 3 EXPERIMENTAL METHODS AND PRELIMINARY RESULTS ................................ ..... 77 Amino Acids Study ................................ ................................ ................................ ................. 85 Introduction to Amino Acids ................................ ................................ ........................... 85 Methods ................................ ................................ ................................ ........................... 88 Results ................................ ................................ ................................ ............................. 88 Conclusions ................................ ................................ ................................ ..................... 93 Coll agenous and Related Biological Materials ................................ ................................ ....... 95 Introduction to Collagenous materials ................................ ................................ ............. 95 Methods ................................ ................................ ................................ ......................... 101 Results ................................ ................................ ................................ ........................... 102 Conclusions ................................ ................................ ................................ ................... 106 Animal Model Study ................................ ................................ ................................ ............. 107 Introduction to Animal Model ................................ ................................ ....................... 107 Methods ................................ ................................ ................................ ......................... 112 Results ................................ ................................ ................................ ........................... 113 Conclusion ................................ ................................ ................................ ..................... 116 4 SUMMARY OF RESEARCH AND PROPOSAL OF FUTURE WORK ........................... 154 Summary of Research ................................ ................................ ................................ ........... 154 Proposal of Future Work ................................ ................................ ................................ ...... 1 60 APPENDIX Appendix ................................ ................................ ................................ ................................ ...... 163 LIST OF REFERENCES ................................ ................................ ................................ ............. 164 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 176
6 LIST OF TABLES Table page 3 1 The band identity, wavenumber, and band assignment for L alanine ......................... 142 3 2 The band identity, wavenumber, and band assignment for L alanine glycine ............. 142 3 3 The band identity, wavenumber, and band assignment for glycine ............................ 142 3 4 The band identity, wavenumber, and band assignment for glycine glycine ................ 143 3 5 The band identity, wavenumber, and band assignment for L proline ......................... 143 3 6 The band identity, wavenumber, and band assignment for glycine L proline ............. 144 3 7 The wav enumber, factor, and band assignment for all traditional Raman studied amino acids ................................ ................................ ................................ ........ 144 3 8 This is the wavenumber, factor, and band assignment for traditional Raman studied peptides ................................ ................................ ................................ ............. 144 3 9 The wavenumber, factor, and band assignment for traditional Raman studied dipeptides ................................ ................................ ................................ ........... 145 3 10 The wavenumber, factor, and band assignment for all DLIPS studied amino acids ..... 145 3 11 The wavenumber, factor, and band assignment for DLIPS studied pe ptides ............... 145 3 12 The wavenumber, factor, and band assignment for all DLIPS studied dipeptides ........ 146 3 13 The predicted class of all materials (i.e., peptides and dipeptides) studied using traditional Raman within the training set ................................ ................................ 146 3 14 The predicted class of all materials (i.e., peptides and dipeptides) studied using traditional Raman within the valida tion set ................................ ............................ 146 3 15 The predicted class of the peptides studied using traditional Raman within the training set ................................ ................................ ................................ ......... 146 3 16 The predicted class of the peptides studied using traditional Raman within the validation set ................................ ................................ ................................ ...... 147 3 17 The predicted class of the dipeptides studied using traditional Raman within the training set ................................ ................................ ................................ ......... 147 3 18 The predicted class of the dipeptides studied using traditional Raman within the validation set ................................ ................................ ................................ ...... 147
7 3 19 The predicted class of all materials (i.e., peptides and dipeptides) studied using the DLIPS method within the training set ................................ ................................ .... 147 3 20 The predicted class of all materials (i.e., peptides and dipeptides) studied using the DLIPS method within the validation set ................................ ................................ 147 3 21 The predicted class of the peptides studied using the DLIPS method within the training set ................................ ................................ ................................ ......... 148 3 22 The predicted class of the peptides studied using the DLIPS method within the validation set ................................ ................................ ................................ ...... 148 3 23 The predicted class of the dipeptides studied using the DLIPS method within the training set ................................ ................................ ................................ ......... 148 3 24 The predicted class of the dipeptides studied using the DLIPS method within the validation set ................................ ................................ ................................ ...... 148 3 25 The band identity, wavenumber, and band assignment for collagen .......................... 148 3 26 T he band identity, wavenumber, and band assignment for fibrinogen ....................... 149 3 27 The band identity, wavenumber, and band assignme nt for IgA ................................ 149 3 28 The band identity, wavenumber, and band assignment for IgG ................................ 149 3 29 The band identity, wavenumber, and band assignment for IgM ................................ 149 3 30 The wavenumber, factor, and band assignment for all traditional Raman technique studied materials ................................ ................................ ................................ 150 3 31 The wavenumber, factor, and band assignment for only the Igs traditional Raman technique studied materials ................................ ................................ .................. 150 3 32 The wavenumber, factor, and band assignment for all the DLIPS method studied materials ................................ ................................ ................................ ............ 150 3 33 The wavenumber, factor, and band assignment for only the Igs DLIPS method studied materials ................................ ................................ ................................ 151 3 34 The predicted class of all materials studied using traditional Raman within the training set ................................ ................................ ................................ ......... 151 3 35 The predicted class of all materials studied using traditional Raman within the validation set ................................ ................................ ................................ ...... 151 3 36 The predicted class of only the Igs studied using traditional Raman within the training set ................................ ................................ ................................ ......... 151
8 3 37 The predicted class of onl y the Igs studied using traditional Raman within the validation set ................................ ................................ ................................ ...... 152 3 38 The predicted class of all materials studied u sing the DLIPS method within the training set ................................ ................................ ................................ ......... 152 3 39 T he predicted class of all materials studied using the DLIPS method within the validation set ................................ ................................ ................................ ...... 152 3 40 The predicted class of only the Igs studied using the DLIPS method within the training set ................................ ................................ ................................ ......... 152 3 41 The predicted class of only the Igs studie d using the DLIPS method within the validation set ................................ ................................ ................................ ...... 152 3 42 The value of the x axis intercept of type IV collagen, fibrino gen, IgA, and IgG ......... 153 A 1 The chemical index of all materials used ................................ ............................... 163
9 LIST OF FIGURES Figure page 3 1 The DLIPS system ................................ ................................ .............................. 118 3 2 The structures of glycine, alanine and proline ................................ ......................... 118 3 3 The structures of glycine glycine, glycine L proline, and L alanine glycine .............. 118 3 4 The general structure of amino acid ................................ ............................... 118 3 5 The traditional fluorescence spectra of the amino acids (i.e. preperturbation) ............ 119 3 6 Key of color that represents each studied amino acid ................................ .............. 119 3 7 The traditional fluorescence spectra of L alanine (blue) and L alanine glycine (red) (i.e. preperturbation) ................................ ................................ ........................... 119 3 8 The traditional fluorescence spectra of glycine (light blue) and glycine glycine (orange) (i.e. preperturbation) ................................ ................................ .............. 120 3 9 The traditional fluorescence spectra of L proline (green) and glycine L proline (violet) (i.e. preperturbation) ................................ ................................ ............... 120 3 10 The DLIPS spectra of the amino acids ................................ ................................ ... 121 3 11 The DLIPS spectra of L alanine (blue) and L alanine glycine (red) after 700 laser pulses ................................ ................................ ................................ ................ 121 3 12 The DLIPS spectra of glycine (light blue) and glycine glycine (orange) after 500 laser pulses ................................ ................................ ................................ ......... 122 3 13 The DLIPS spectra of L proline (green) and glycine L proline (violet) after 600 laser pulses ................................ ................................ ................................ ................ 122 3 14 The DLIPS spectra of glycine (light blue) after 500 laser pulses, L alanine (blue) after 700 laser pulses and L proline (green) after 600 laser pulses ............................ 123 3 15 The DLIPS spectra of glycine glycine (orange) aft er 500 laser pulses, L alanine glycine (red) after 700 laser pulses and glycine L proline (violet) after 600 laser pulses ................................ ................................ ................................ ................ 123 3 16 The traditional Raman spectra of the amino acids (i.e. preper turbation). Each spectrum other than proline pre has been shifted vertically to show distinction between Raman peaks of each amino acid ................................ ............................. 124 3 17 The traditional Raman spectra of L alanine (blu e) and L alanine glycine (red) (i.e. preperturbation) ................................ ................................ ................................ .. 124
10 3 18 The traditional Raman spectra of glycine (light blue) and glycine glycine (orange) (i.e. preperturbation) ................................ ................................ ........................... 125 3 19 The traditional Raman spectra of L proline (green) and glycine L proline (violet) (i.e. preperturbation) ................................ ................................ ........................... 125 3 20 The traditional Raman preper turbation (blue) spectrum of L alanine ........................ 126 3 21 The traditional Raman preperturbation (red) spectrum of L alanine glycine .............. 126 3 22 The traditional Raman preperturbation (light blue) spectrum of glycine .................... 127 3 23 The traditional Raman preperturbation (orange) spectrum of glycine glycine ............ 127 3 24 The traditional Raman preperturbation (green) spectrum of L proline ....................... 128 3 25 The traditional Raman preperturbation (purple) spectrum of glycine L proline .......... 128 3 26 T he DLIPS of the amino acids ................................ ................................ .............. 129 3 27 The traditional fluorescence preperturbation spectra of the collagenous material ........ 129 3 28 The color that represents each studied collagenous material ................................ ..... 130 3 29 A close up view of the absolute fluorescence spec tra of the collagenous material. (Fibrinogen and IgG are not plotted here) ................................ .............................. 130 3 30 The DILPS spectra of the collagenous material after 2000 laser pulses. .................... 130 3 31 A close up view of the DLIPS spectra of the collagenous material after 2000 laser pulses. (IgM is not plotted here) ................................ ................................ ............ 131 3 32 The traditional Raman spectra o f the collagenous material. Each spectrum other than IgM has been shifted vertically to show distinction between Raman peaks of each collagenous material. ................................ ................................ ........................... 131 3 33 The traditional preperturbat ion spectrum of collagen ................................ ............... 132 3 34 The traditional preperturbation spectrum of fibrinogen ................................ ............ 132 3 35 The traditional preperturbatio n spectrum of the IgA ................................ ................ 133 3 36 The traditional preperturbation spectrum of IgG ................................ ..................... 133 3 37 The traditional preperturbation spectrum of IgM ................................ .................... 134 3 38 The DLIPS of Collagenous and related biological material ................................ ...... 135 3 39 The average DLIPS spectra at weeks 1 11 for t he untreated skin .............................. 13 6
11 3 40 The average DLIPS spectra at weeks 1 11 for the DMBA treated skin ...................... 137 3 41 The average absolute fluor escence for the DMBA week 2 ................................ ................................ ................................ ............... 137 3 42 The average absolute fluorescence for the DMBA week 6 ................................ ................................ ................................ ............... 138 3 43 The average absolute fluorescence for the DMBA week 8 ................................ ................................ ................................ ............... 138 3 44 The average DLIPS fluorescence for the DMBA and untreated skin at week 2 ................................ ................................ ................................ ............... 138 3 45 The average DLIPS fluorescence for the DMBA week 6 ................................ ................................ ................................ ............... 139 3 46 The average DLIPS fluorescence for the DMBA week 8 ................................ ................................ ................................ ............... 139 3 47 The H&E stained histology section of skin from an untreated mouse at week 4 ......... 140 3 48 The H&E stained histology section of skin from a DMBA treated mouse at week 4. orthokeratotic hyperkeratosis. (X) highlights the thickening of the stratum spinosum, characteristic of acanthosis. (*) denotes areas of increased dermal collagen density ... 140 3 49 The H&E stained histology section of skin from a DMBA trea ted mouse at week 8. orthokeratotic hyperkeratosis. (X) highlights the thickening of the stratum spinosum, characteristic of acanthosis. (*) denotes areas of increased dermal collag en density ... 141 3 50 The H&E stained histology section through a developed papilloma of skin from a DMBA treated mouse at week 11 ................................ ................................ ......... 141
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DIFFERENTAIL LASER INDUCED PERTURBATION SPECTROSCOPY FOR ANALYSIS OF BIOLOGICAL MATERIAL By Sa rah Elizabeth smith December 2012 Chair: David W. Hahn Major: Mechanical Engineering Improved sensitivity and/or specificity for rapid and accurate biosensing is highly desirable for in situ and in vivo cancer screening, detection of biological pathogen s for biodefense, as well as food and building safely. However, despite the need and motivation, to date the clinical applicability of in vivo sensing schemes has been limited by the large patient to patient variations. The current study seeks to provide a dditional insight as well as increased sensitivity and specificity as compared to the current state of the art optical based sensing methodologies. In this study d ifferential laser induced perturbation spectroscopy (DLIPS) is developed and studied in detai l. DLIPS is a unique optical sensing scheme based on deep ultra violet ( UV ) photochemical perturbation in combination with difference spectroscopy. Applying a sequence of optical probing, UV laser induced perturbation, and repeat optical probing coupled with difference s pectroscopy provides a new spectral signature. This scheme is based on recent research showing that the biological matrix may be altered by low intensity ultraviolet radiation such that the intrinsic fluorescence or Raman scattering response is perturbed. This new spectral dimension based on difference spectroscopy, will be strongly coupled to the local biomolecular
13 matrix. Since the targeted material is optically probed both before and after perturbation with the deep UV light source, the resulting differe ntial response will avoid the major limitation of the current biosensing schemes, namely, the significant variations in the absolute optical response, as generally observed in patient to patient populations. Experiments with amino acids glycine, alanine, p roline, glycine glycine, glycine proline, and alanine glycine are conducted using Raman and fluorescence s pectroscopy as the probes. Results showed that the DLIPS technique can differentiate between all amino acids. Experiments on collagenous and related c onnective tissue materials study involving collagen type IV, fibrinogen, immunoglobulin A, immunoglobulin G, and immunoglobulin M was also conducted. These experiments also showed the sensitivity and specificity that can be achieved with the D LIPS techniqu e using Raman and f luorescence Spectroscopy as the probes. Experiments performed in an animal model study upon mice and hamsters, confirm ed that the DLIPS technique can be implemented into a n in vivo environment while retaining the sensitivity and specific ity this technique offers. Overall, the current set of experiments shows the strong dependence of laser tissue interactions and highlights the need for a complete understanding of the related photo chemical perturbation processes Finally, additional exper iments are proposed to further facilitate this understanding.
14 CHAPTER 1 INTRODUCTION In this dissertation, a unique optical sensing scheme known as differential laser induced perturbation spectroscopy which is based on a sequence of optical probing, ultra violet ( UV ) laser perturbation, and repeat optical probing coupled with difference spectroscopy, is used to analy ze biological material s Analysis of biological material s through an amino acids study, a collagenous material study, and an animal model study, that resu lt in new differential spectral signatures is presented. 1.1 The Properties of Light, Lasers, and Matter 1.1.1 Properties of Light The interaction of a light quanta and a molecule can result in chemical and physical changes within the molecule. The interac tion of light with matter can also tell us about the nature of the matter setting the basis for optical diagnostic scheme. determined that light is a quantized amount of ene rgy that behaves as a particle and a wave. Therefore, light or radiation, p ossess a dual characteristic, and can be viewed as a wave in an electromagnetic field, or a stream of massless particles called photons. A wave can be characterized by its wavelengt h or by the frequency and energy that it carries. Light is known as electromagnetic radiation and an electromagnetic spectrum contains all p ossible waves over a continuum of frequencies and energies. The spectrum spans from radio radiation which has a wave length >10 9 (1 = 1x10 10 m) and a frequency <3x10 9 Hz to gamma ray radiation which has a wavelength <0.1 and a frequency >3x10 19 Hz. (Calvert, 1966) Also, the theory of wave particle duality allows us to assume that photons and electrons follow the sa me general principles. However, a few distinctions must be noted. Electrons must obey Fermi statistics while
15 a photon must obey Bose statistics. In Fermi statistics, two electrons in the same system cannot have the same physical properties (i.e. quantum nu mbers) In Bose statistics there is no restriction so photons can occur in large numbers, having the same energy, and momentum, such as in a pulse of laser light. 1.1.2 Properties of Lasers The amplification of light by stimulated emission of radiation (i. e. LASER light) is a unique form of radiation. Unlike other forms of radiation, light amplification by stimulated emission of radiation, known as laser light, is highly monochromatic, coherent, and collimated. (Niemz, 1996) A laser consists of three main p arts: an energy source or pump, an amplifying medium, and an optical cavity. In the amplifying medium, an atom or molecule is excited and emits spontaneously a photon as it returns to its lower energy state. As these excited state atoms release their photo ns, the photons will interact with other excited atoms to stimulate the release of more photons of the same wavelength, energy and phase. Atoms in the ground state can also absorb photons, and since more electrons are in low energy states than in upper sta tes, the probability of net absorption is larger than a net amplification. For more photons to be produced than removed, more atoms must be in an excited state, thereby creating a population inversion. This is obtained by the energy source or pump deliveri ng energy to a multiple energy level roster (generally three level or four level) and taking advantage of different decay rates. The a mplifying medium will therefore determine the wavelength of the laser radiation. The optical cavity consists of precisely aligned mirrors to direct the photons in phase, through the laser medium to realize amplification as well as to provide directionality and increased intensity with high coherence. The laser beam profile will depend on the geometry of the mirrors, their sep aration, and the optical cavity construction. Therefore, different wavelengths can be emitted from the laser by using different laser systems and mediums
16 As noted above, a laser produces a highly monochromatic, coherent collimated beam. The beam is of a s ingle wavelength or frequency (with some finite linewidth) which is determined by the amplification medium. The laser beam is also considered one of the most coherent forms of light. Coherence defines the phase correlation between different emitted waves a t different times and locations. The waves of light in lasers are in phase with each other in both time and space; hence, they are coherent. The laser beam is highly collimated, the waves of light run parallel to each other, because of the laser cavity des ign and the monochromatic and coherent nature of the stimulated light generated within the cavity. No other light source can generate a beam with precise directionality and minimum angular spread as a laser. Another unique feature of lasers is that they ar e capable of producing short bursts of highly intense light. Laser light, depending on the energy level scheme, can be delivered as a continuous wave (CW) or a pulsed wave. Continuous wave lasers emit a beam of light with near constant intensity over time. Pulsed wave lasers emit a pulse or multiple pluses at a selected interval. The excitation energy can also be stored and released suddenly by operating the laser in a Q switch mode or locking mode. Very short pulses (picosecond or femtosecond) are possible with lasers, but in principle an infinitely short pulse of finite energy will contain all wavelengths with the same power; hence an increase in bandwidth or a reduction in the monochromatic characteristic of the laser light can occur with ultra short puls es. (Pedrotti, 1987) The current research will predominately utilize Q switched lasers with pulse widths in the ns time scale regime, and wavelengths generally in the UV and visible wavelength regime. 1.1.3 Properties of Matter Now that a basic introductio n to light has been presented, the properties of matter will now be briefly reviewed in the context of laser material interactions Generally, k nowing the electron configurations within atoms allows us to better understand and predict the properties of
17 ele ments. A specific group of atoms within a molecule that is mainly responsible for reactions within the molecule is called a functional group. Relevant to biological tissue, a ll organic compounds have a functional group of hydrocarbons. Hydrocarbons are div ided into two main classes based on the structure and are known as aliphatic and aromatic. (Chang, 2003) Aromatic hydrocarbons contain one or more benzene rings where aliphatic hydrocarbons do not contain a benzene ring or benzene group. Also in alkanes, o nly single covalent bonds are present and they contain the maximum number of hydrogen atoms that can bond with the number of carbon atoms present. Like light, matter also possesses a dual nature, with the quantum structure ultimately defining its interacti on with light. The dual nature of electrons can be explained best through the principle states that it is impossible to know, simultaneously, the momentum and the position of a particle with certainty. Therefore, it is not appropriate to imagine the electron in a well defined orbital, circling the nucleus. As a result, particles do not obey deterministic laws of motion. Schrodinger in 1926 determined a mathematical model that can describe the behavior and energies of submicroscopic particles through calculated probabilities. (Chang, 2003) The Schrodinger equation incorporates particle behavior in terms of mass and wave behavior in terms of a wave function. The electr on density gives the probability that an electron will be found in a particular region of an atom. The wave function has no direct physical meaning but the wave function squared is proportional to the probability of finding the electron in a certain region in space (i.e. the distribution of the electron density in space around the nucleus). The wave amplitudes describe the probability that a particle will be found within a region of space during an interval of time. Thus, high wave amplitudes or high electr on densities represent a high
18 probability of locating the electron in that region. This results in acceptable energy levels (most likely) capable of describing electrons moving around the nucleus or the atom. The wave function of an electron in an atom can be thought of as an atomic orbital. Therefore, an atomic orbital has a characteristic energy and electron density distribution. These are known as the various quantized electronic energy levels of an atom. The distribution of electrons in atoms can be des cribed by four quantum numbers, three equation, the principal, the angular moment um, and the magnetic, are used to describe atomic orbitals and to label electrons that reside within the atomic orbitals. The fourth quantum number called the spin quantum number is used to describe the behavior of a specific electron. The four quantum num bers also help to explain electronic transitions and the periodic behavior of the elements. The total energy of an atom with many electrons depends not only on the sum or the orbital energies but also on the energy of repulsion or between the electrons in these orbitals. The magnetic moment due to the spin of an electron and the magnetic field produced by the electrons orbital motion around the nucleus can result in a magnetic interaction and can create spin orbital coupling. The charge of the nucleus (the atomic number) will affect the spin orbital coupling and heavier atoms will have strong spin orbital coupling. Shifting and splitting of the energy levels within the atom can be produced by the spin orbital coupling since strong mixing of spin and orbital properties can occur. Such behavior plays a key role in many absorption and fluorescence schemes.
19 In atoms with several unpaired electrons, strong inter electron coupling between the spin and angular momenta will occur. In lighter elements, the total angul ar momentum of the atom is assumed to be the sum of the net orbital angular momentum of the electrons and the total spin angular momentum. This type of coupling is known as Russell Saunders coupling (L S coupling). For heavier elements, the total angular m omentum of the atom is assumed to be the sum of the individual orbital and spin angular momenta couple for each individual unpaired electron. This type of coupling is known as J J coupling. (Niemz, 1996) In addition to electron levels, molecules also posse ss vibrational leve ls and rotational energy levels, which form the basis for vibrational spectroscopy, such as Raman scattering. At room temperature, most molecules exist in their ground vibrational state. The spacing between two consecutive vibrational le vels is constant and the vibrational frequency is inversely proportional to the square root of the reduced mass, proportional to the square root of the force constant and proportional to the strength of the bond. These relationships help to explain why hea vier nuclei have lower vibrational frequencies. For a molecule with N atoms, the numbers of different vibrational displacements or degrees of freedom are 3N 5 for a linear molecules and 3N 6 for nonlinear molecules. The 5 within the 3N 5 equation comes fro m the fact that three translational and two rotational degrees of freedom occur within linear molecules. The 6 within the 3N 6 equations comes from the fact that there are three translational and three rotational degrees of freedom within nonlinear molecul es. (Chang, 2003) These vibrational modes are coupled and result in simultaneous displacements of all nuclei. The displacements of nuclei and the symmetry of the molecule result in distinct vibrational patterns and are called normal modes of vibration. A s pecific mode of vibration is assigned to the specific bond or angle of the
20 molecule and its corresponding displacement. Such behavior plays a key role in Raman scattering, as discussed later including the selection rules Since a molecule contains more th an one nucleus, determining the origin of the electronic displacement becomes difficult. Two electrons in a spin coupled configuration define the formation of a bond. A diatomic molecule with the same nucleus and the same type of atomic orbitals can combin e to form molecular orbitals. Overlapping of two orbitals in their lowest states can be described by two possible linear combinations of 1s atomic orbitals of each atom. Overlapping wave functions can be described as constructive or destructive interferenc e. The constructive overlapping results in an increased electron density between the two nuclei resulting in a binding between the orbitals and allows the nuclear nuclea r repulsion to be overcome. The comparison with the individual 1s atomic orbitals. This is known as a bonding molecular orbital. The destructive overlapping res ults in a cancellation of the electron density in the region between orbital is higher in comparison with the individual 1s atomic orbitals. This is known as an a nti bonding molecular orbital. (Prasad, 2003) The asterisk symbol as a superscript represents an anti nuclear axis. Only when a multiple bond is formed between two atoms c form. These orbitals are formed by the overlapping of two atomic orbitals and are not as stable as In polyatomic molecules, electrons are either bound to a particular molecule or are quasi free. The motion is rest ricted and these rotational and translation motions become known as lattice vibrations. More than one atomic orbital on one atom is typically involved in bonding and
21 certain atomic orbitals do not mix to form bonds. Also in polyatomic molecules multiple in teractions can occur and these interactions produce a modification of the orbitals of individual molecules. New energy states are spread over both molecules and new energy band formation occurs for both the electronic and vibrational energy levels of each molecule. One of these interactions is known as weak van der Waals interactions (~0.01 eV). (Calvert, 1966) Van der Waals interactions can occur even between neutral molecules. Momentary random fluctuations in the distribution of the electrons of an atom c reate an unequal electric dipole and unequal electric dipoles of two non covalent bonded atoms can attract each other resulting in van der Waals interactions. Another type of interaction is related with specific chemical bonding or chemical association. An example of this type of interaction is hydrogen bonding When a hydrogen atom is bonded to an electronegative atom in one molecule and is bonded to another electron rich atom on another molecule, a weak electrostatic interaction occurs. Having discussed t he general atomic and molecular structure, attention is now turned to more complicated tissue structures. Tissue is composed of specialized cells that are organized to perform specific tasks. Cells interact with each other and their functions are coordinat ed to allow an organism to perform a set of functions. Cells in tissues are generally held together by the extracellular matrix. Within this matrix, cells can interact with each other as well as migrate. The extracellular matrix is composed of a variety of proteins and polysaccharides (long carbohydrate molecules of repeated monomer units joined together by glycosidic bonds) that are assembled into an organized lattice. The chemical structure of protein consists of amino acids that are linked together by a peptide bond. Condensations of the amino groups are of one amino acid with the carboxylic group of another forming the peptide bond. This process involves the NH 2 and COOH groups on
22 both ends of amino acids. (Chang, 2003) There are 20 different amino acids and the functions of each depend on the side chain group. The extracellular matrix not only provides adhesion and binding of tissue but also provides strength and resilience, and cushion for the cells depending on how the matrix organizes. The strength an d resilience of tissue is proportional to collagen content and biological tissue is known to be fragile. Water content, blood circulation and structure all affect the properties of tissue. Living tissue does not have the same structure as dead tissue; dryi ng, freezing, dehydration or soaking in saline of tissue will greatly alter tissue properties as well. In vitro determined tissue properties can also greatly differ from in vivo determined properties. It is obvious the nature of tissue is inhomogeneous and this creates many difficulties in comparing experimental data between samples, notably so for laser tissue interactions, as described in the next section. 1.2 The Interaction of Light and Matter 1.2.1 Light and Matter Interactions on a Bulk Level Matter, when compared to the wavelength size of the incident light is considered bulk matter if the tissue dimension is much larger than the wavelength of light. The wave nature of light leads to two important properties of the interaction of light with matter on a bulk level: refraction and reflection. The absorption and scattering of light are two other properties which are important in this type of interaction but the wave model of light cannot explain these phenomena or the effects which absorbed light produces Therefore, light interacts with bulk matter in three key ways: absorption, reflection and refraction, and scattering. (Pedrotti, 1987) These properties are determined by the sum of the average of the molecular properties that correspond to the bulk matte r. Light will propagate undisturbed until it interacts with a nonhomogenity (e.g. tissue). The interaction can be described as a bulk medium subjected to an external electric field that responds by becoming electronically polarized.
23 Once absorption occurs in matter, the photon energy is partially converted to heat. Energy as well as momentum is involved in the interaction of light with matter, and both must be conserved. The ability of bulk matter to absorb radiation is dependent upon the material compositi on, the wavelength of radiation, and the material properties such as excitable energy state s Absorption can be quantified as the ratio of absorbed intensity to the incident intensity. The amount of light absorbed by bulk material is independent of the inc ident light intensity provided saturation does not occur, and dependent upon the amount of absorbing material through which the light passes. Reflection and refraction are generally dependent upon the material, the incident wavelength, polarization of ligh t and the angle of incidence. In bulk matter, reflection occurs by re radiation (i.e. scattering) from the surface upon which it was incident. Within the plane of incidence, the wave normal of the reflected beam, incident beam, and the reflecting surface n ormal must all satisfy the law s of reflection. A reflecting surface is the boundary between two materials of different indices of refraction, where the index of refraction describes the optical response of a material with respect to propagation of light th rough it. The amount of refraction of light also depends upon the index of refraction which determines the phase as well as the velocity of propagation and strongly depends on wavelength in regions of high absorption only. Refraction usually occurs when tw o mediums of different indices of refraction are within the same reflection surface resulting in a change in speed of the light wave. The change in speed of the light wave results in a change of angle of propagation when entering from on medium to another. and can be difficult to measure in a complex medium like tissue due to simultaneous absorption and scattering.
24 If the frequency of the incident radiation does not corre spond to the frequency of the vibrations of a particle, elastic scattering will occur. In general, depending upon the energy of the incident and scattered photons, scattering can be elastic or inelastic. Elastic scattering occurs by Mie or Rayleigh scatter ing, and the incident and scattered photons will have the same energy (i.e. same wavelength and frequency). Mie scattering is applicable for particles of size comparable to the wavelength of the incident radiation or laser. Mie scattering typically scatter s x radiation wavelength. Rayleigh scattering is applicable only if the particles are much smaller in size when compared to the wavelength of the incident radiation Rayleigh scattering scatters in 4 wavelength. (Pedrotti, 1987) Rayleigh scattering will result in blue light scattering more efficiently than red light as bas ed on scattering alone. In general, the longer the wavelength of light, the deeper it w ill penetrate into bulk matter. In inelastic scattering, the incident and scattered photons will have d ifferent energies. One type of inelastic scattering is Brillouin s cattering. Brillouin scattering occurs by inducing inhomogeneity of the refractive index from an acoustic wave interfering with bulk matter. This type of scattering will result in the frequency of the scattered photons shifting up or down. Another type of inelastic scattering is Raman scattering. Stokes Raman scattering occurs when the scattered photon energy is lower than the incident photon energy, and anti Stokes Raman scattering occurs when the scattered photon energy is higher than the incident photon energy. The difference i n energy between incident Raman scattering radiation is accounted for as energy either added to or removed from vibrational modes of the solid matrix.
25 1.2.2 Light and Matter Interactions on a Molecular Level The stability of molecul es can help explain the nature of the reactions that the molecules may undergo. Energy transfer and excited state formation requires one or more molecules interacting, therefore, a minimum molecular size is required and is known as a dimer. A dimer in an e xcited state is called a n e xcimer (i.e. excited dimer). The interaction of two different molecules with each other, when one is in an excited state is known as an excited state complex on is typically used to explain the interaction between radiation and matter. Stimulated absorption, spontaneous emission, and spontaneous absorption are the three processes that explain this interaction. Absorption of light results in an excitation of cer tain vibrations or electronic states in the molecules of the absorbing material. Like bulk matter, on the molecular level, the ability to absorb radiation is dependent upon the material composition, the wavelength of radiation, the material properties and the energy state. In stimulated absorption, the transition of a molecule from a lower energy initial level to a higher energy level occurs where the photon energy of the incident radiation is equal to the energy gap between the levels. A photon will be abs orbed in this process. Generally, at room temperature molecules are in their lowest state known as the ground state. If the initial level is an excited sate, the absorption is called excited state absorption. Absorption and emission processes can be define d by the Einstein model which states that the number of molecules present in the initial state and the density of the photons are proportional to a transition rate from a lower energy state to a higher energy state. Absorption and emission processes can oc cur at any of the electronic and vibrational states of a molecule, provided that the transition is allowed. Once a molecule is in an excited state, it must find a way to release the extra energy to return back to the ground state. A variety of paths are us ually available for the degradation of the
26 electronic energy. Spontaneous emission may occur whenever molecules are in an excited state. For spontaneous emission, no external radiation is required to start the emission process; therefore the emission rate is only proportional to the number of molecules in the excited state. To radiatively bring the molecule back to the ground state, a photon is spontaneously emitted in a random direction, with photon energy corresponding to the energy gap between the two en ergy levels. In the spontaneous emission process, if external radiation is interacting with molecules, the emitted photon will not generally correspond directionally with the external radiation. In contrast, the process of stimulated emission requires an i ncident photon of an energy that corresponds to the energy gap between the initial and higher level. The incident photon will interact with an atom or molecule in an excited state and stimulates the system to fall to a lower energy level and release a phot on of the same energy, phase, and direction as the incident photon as discussed previously in the context of lasers. Nonradiative processes can also occur where the release of energy can be dissipated as heat or produce a chemical reaction or be passed t o another system. When heat is produced in a nonradiative process, the excess energy is converted to vibrational energy by electronic vibrational state coupling. The vibrational energy is then converted to heat by vibrational relaxation. A combination of r adiative and nonradiative processes can also occur as the molecule returns to the ground state. (Chang, 2003) Even though the entire molecule is in an excited state, the excited state energy is mainly localized for simple transitions and can be assigned to chromophores in a molecule. A chromophore is a molecular unit where an electron being excited is primarily located. Radiative and nonradiative processes are considered photophysical process es which do not lead to an over all chemical change. Excited molec ules can also directly
27 produce new products or forms of free radicals or excited molecules which can lead to physical or chemical changes. The natural state of most molecules is in the ground state and involves paired electrons that, have a total spin of S=0. (Calvert, 1966) Excitation of paired electrons from molecules whose ground states are singlet will result in the electrons staying paired and excited in a singlet state S, or the electrons become unpaired and excited in a triplet state T. Two electro nic states of the same spin multiplicity can cross and this is called an internal conversion. After an internal conversion, vibrational relaxation occurs where the excess energy is converted to heat and the molecule will now be at the lowest vibrational le vel of the electronic state. Next, a photon is emitted and the molecule will return to the ground state. This type of emission (from a high singlet state to a low singlet state of same spin multiplicity) is known as fluorescence. Crossing of excitation can also occur from a singlet state to a triplet state where the two different states have different spin, and this process is known as intersystem crossing and is a nonradiative process. Typically, the energy of an excited triplet state is lower than that of an excited singlet state for excitation of the same orbital state. The transition of the excitation of an electron from the highest occupied orbital centered on one molecule to the lowest unoccupied orbital centered on another molecule results in an inter molecular charge transfer interaction. This type of interaction of energy levels between two different molecules will also produce a shift of their energy levels and is called a Forster energy transfer. A charge transfer transition can also occur in an asy mmetric molecule. 1.2.3 Light and Tissue Interactions The cellular, extracellular, and bulk properties of tissue will produce diverse processes during the interaction of light. Tissue, like other matter, will interact with light through absorption, scatter ing, reflection and transmission; resulting in radiative and nonradiative
28 processes. (Niemz, 1996) Important properties of tissue itself include the structure, water content, blood circulation, thermal conductivity, heat capacity, density, and the presence of melanin, other fluorophores, and chromophores. The inhomogeneity of different samples of biological tissue makes it difficult to correlate its optical properties. If the frequency of the incident wave equals the frequency of the free vibrations of the tissue particles, absorption will occur. Absorption of photons by tissue is typically dominated by proteins, DNA, water, melanin, and hemoglobin absorbing the incident photons. The exact tissue component that absorbs the photons depends on the wavelength o f the incident light, and these components are known as chromophores. A chromophore will absorb energy and become electronically excited. For tissue, this vibrational and translational energy is typically converted to thermal energy or heat. The longer the wavelength, the deeper it will penetrate the tissue up to a wavelength of about 1300 nm. At wavelengths above 1300 nm, penetration is only superficial because of the absorption coefficient of tissue water. (Prasad, 2003) The absorbance of tissue is diffic ult to determine since the photons that are absorbed cannot be used any more for detection and the scattering may be significant. Subtraction of the transmitted, reflected, and scattered intensities from the incident intensity is one method of determine th e absorbed intensity. In general, the most evident effect in light and tissue interactions is scattering. Once light enters tissue, scattering occurs because of the heterogeneous structure of tissue which varies in particle size and index of refraction bet ween different parts of the tissue. The depth of penetration is also limited since scattering in such a turbid medium results in backward and forward scattering. It should now be obvious that the scattering process in tissue is complex. In biological tissu e, inelastic scattering via Brillouin scattering is weak, becoming significant only under shockwave generation. Neither Rayleigh nor Mei scattering can completely describe the
29 elastic scattering of light by tissue. Raman scattering of light by tissue will produce excitation of the molecular vibrations and may be significant in tissue. Scattering within tissue will also affect the abs orption of light by the tissue. In most tissues, absorption and scattering processes will take place simultaneously, and will occur with varying ratios. 1.2.4 The Interaction of Lasers and Bulk Matter The destructive interaction of lasers with matter is facilitated by the generation and then acceleration of free electrons. For such light and matter interactions, a free electron i s produced by the simultaneous absorption of photons or by the kinetic energy of an impacting free electron to exceed the band gap energy. However, for the interaction of lasers with matter, the exceeded band gap energy requirement is replaced by the effec tive ionization potential requirement since (Niemz, 1996) At least one free seed electron produced by multi photon ionization or by impact ionization is req uired to start a cascade ionization effect. Photoionization for different laser field strengths and frequencies can result in multi photon ionization. In impact ionization, the kinetic energy of the impacting electrons is used to overcome the effective ion ization potential requirement. The excess energy that remains after the impact of electrons is distributed among the electrons involved in the collision. This results in the electrons needing to gain less energy to reach the critical level. The energy carr ied by the free electron within the material, for a time interval that depends on the material and laser settings, will originally induce a primary thermalization effect. Thermoelastic stresses caused by the rise in temperature will stay confined to the ir radiated area, creating a rise in pressure. If the laser pulse duration is longer than the thermalization effect, the pressure within the confined medium will reach a maximum, and a stress wave will be emitted that contains a compressive and tensile stress Even if damage is not produced by the increase in
30 temperature (thermal damage) the tensile stress wave can induce fractures in the material. In an aqueous media, rupture by tensile stresses under isothermal conditions is called cavitation and bubble form ation by heating under isobaric conditions is called boiling. However, under confined stress conditions where heating and stretching are both occurring distinction becomes difficult therefore the term optical breakdown is used. Also, if the laser pulse dur ation is in the nanosecond range or shorter and has an extremely high spatial density of photons an optical breakdown will result. During optical breakdown, temperatures of several thousand Kelvin can be reached. Optical breakdown can include bubble format ion, plasma luminescence, and ionized regions of extremely high electron density. Material removal via volumetric processes is known as ablation. Ablation processes can remove molecular clusters, molecular fragments, and single molecules as well as forming bubbles or cracks through the fracture of chemical bonds. The ablation of deeper layers will continue as long as the critical energy is supplied and this process is known as phase explosions. Photochemical, photomechanical and/or photothermal processes ca n lead to vaporization, molecular fragmentation and void formation through phase transitions. When a pulsed laser beam is focused, plasma formation results in catapulting, while for a pulsed unfocused laser beam, confined thermal ablation results in catapu lting. Plasma formation initiated by linear absorption can be used for micro dissection and for surface cleaning. Conventional cleaning of surfaces typically has a negative result involving direct contact and chemical reactions. Surface cleaning by lasers is capable of removing submicroscopic particles and hydrocarbon films of contaminants from surfaces. Since the ability to control material removal in surgery and electronic manufacturing is important, ablation at 248 nm has been studied by Kuper and coll e a g u es in 1990. For many
31 polymers, the dominating contribution of ablation by UV e xcimer lasers is unknown. Poly methyl meth acrylate (PMMA) can be used as a model for quantitative and qualitative analysis of the photolysis products. A well defined photochem ical fragmentation pattern of PMMA has been found and the major primary process is the side chain scission under the formation of double bonds. The degradation process upon heating of PMMA indicated the existence of highly vibrational excited molecules and can be considered as the photothermal contribution. Finally, the high measured quantum yield showed the photochemical contribution to the ablation process. It was concluded that thermal and photochemical processes can produce an increase in the specific v olume to cause ablation of material. Nucleation and pressure generation processes at a liquid solid interface induced by a pulsed laser can create a surface plasmon. A surface plasmon can be used as a probe which can provide accurate information on the ear ly stages of the nucleation processes, nucleation thresholds, transient pressure generation, absolute pressure amplitudes and growth velocities by rapid bubble growth. The surface plasmon will be scattered by the bubbles formed resulting in the surface pla smon resonance broadening and shifting and hence temporarily increasing the reflectance. As the bubbles collapse, the surface plasmon resonance becomes narrower and the reflectance decreases. The resonance shift is caused solely by the rise in temperature at the interface and this is supported by numerical computations for the temperature evolution at the interface. It was also determined that the bubbles remained confined within the superheated liquid layer and the size does not exceed the thickness of th e superheated liquid layer. Knowledge about the plume dynamics and the stress distribution below the target surface helps to explain the kinetics of phase transitions during ablation. The visualization of ablation plumes is difficult since the refractive i ndex variations are usually smaller in gases than in
32 liquids or solids and the plumes are small in size, so interactions are short with light illumination that produces only small deflections. Campillo in 2006, using a Schlieren filter, which requires a po int source and white light, studied plume dynamics. The Schlieren technique works on the principle of deflection of light beam to produce dark field images. Using this technique, the plume was determined to consist of an upper and lower part. With the uppe r part consisting of water vapor that results from complete vaporization of the targets top layer, and the lower part consisting of vapor and droplets that result from a phase explosion. It was also shown that the collision between the plume and ambient ai r produced an internal shock wave that traveled back through the center of the plume. Any simple relation between radiant exposure and ablation depth as predicted by a blow off model or stead state models can only be an approximation for a limited range of radiant exposures. Therefore, different models have been developed in an attempt to describe UV laser ablation of polymers by Arnold and Bityurin in 1999. In a photochemical model, electronic excitation will result in the direct breaking of bonds without thermalization. In models where bonds are broken thermally, the ablation rate, threshold of the laser pulse rate, and pulse length are all dependent and supports that the ablation process is thermal in nature. In photophysical models, both thermal and non thermal features are taken into account. Models of ablation can also be divided into volume and surface models. In volume models, the decomposition of material is assumed to take place within the bulk material. In surface models, material removal takes pla ce within several monolayers from the surface. Combining multiple models can predict a sharp explosive ablation threshold as well as the near threshold for ablation occurring significantly after the end of the laser pulse. Also, the small ablated depth per laser pulse will result in high values of ablation velocity.
33 1.3 The Interaction of Lasers and Tissue The interaction of laser light with tissue depends upon the properties of the tissue and the parameters of the laser. The magnitude, exposure time, and p lacement of the deposited heat inside the tissue will determine the spatial extent and degree of tissue reaction. The optical properties of the tissue that effect this interaction are reflection and refraction, scattering, and absorption. Laser beams can b e used to open, seal, cut, vaporize, coagulate, cauterize, rupture and weld human tissue. Biophysical or biochemical processes can also be monitored it real time using very short pulses of laser light. However, once the beam enters the tissue, scattering o ccurs resulting in the rapid loss of collimation of the beam. This will result in a loss of the initial directionality of the laser beam as well as a defocusing of the beam spot. The scattering process in tissue is quite complex and involves several mechan isms since the structure of tissue will initiate scattering of the beam and the region within the tissue will be irradiated from all directions by multiple photons. The fundamental principles, depending upon the interaction time and the effective power den sity of the interaction of lasers with tissue can be summarized as photochemical, photothermal, and photomechanical effects. For photochemical and photothermal processes, a quasi continuous wave irradiation can be used. Photomechanical processes can be ach ieved with short, high powered laser pulses. Interaction types that occur adjacently cannot always be strictly separated. 1.3.1 Photochemical Effects Photochemical effects will occur at long interaction times, ranging from seconds to continuous wave, and a t low power densities, typically 1 W/cm 2 Light is absorbed by the tissue with no primary heating of the tissue and will induce chemical effects and reactions within the tissue. Almost all biological relevant photochemical reactions are dependent upon the generation of reactive oxygen species. Photochemical interactions of lasers with tissue are complicated and
34 lengthy. Therefore, photochemical interactions will be discussed when appropriate in the following chapter. 1.3.2 Photomechanical Effects Photomecha nical effects will occur at power densities greater than 100 W/cm 2 Laser light is converted into kinetic energy by strong electric fields generated from the high irradiance leading to either a breakage of the intracellular structure, or a dissociation or ionization of the tissue involved. Photomechanical effects on tissue are determined by the magnitude of the tensile forces, shear forces and their duration since tissue must strain before it ruptures. For tissue, while the strain at fracture does not chang e significantly with strain rate, the ultimate tensile strength (UTS) increases. Rapid deformation conditions will result in the increase in UTS since significant viscous dissipation occurs between matrix elements. I n general, i ncreased tissue strength is associated with increased collagen content and phase transitions of water within tissue are strongly dependent upon the mechanical properties of tissue. In soft tissues, high intensity irradiation produces plasma and the generation of shockwaves in the ti ssue disrupts the tissue through mechanical effects. The mechanical energy of a laser is used to create shockwaves and cavitation in tissue with minimal thermal damage is known as photodisruption. Photodisruption will result in these effects spreading into adjacent tissues. Catapulting driven by plasma formation will result in strong shear forces when tightly focused laser pulses are used. Also, the re cultivation rate of catapulted cells is much higher when focused pulses are used. Time resolved photograph y of UV exposed live cells proves the side effects by heat will play a minor role when compared to the damage by mechanical effects. Another photomechanical effect that occurs at power densities above the threshold value is the photoablative effect. The en ergy of a single UV photon is sufficient to cause dislocations in the photoablative effect. The photoablative effect differs from photodisruption and plasma
35 induced ablation by occurring on a shorter time scale, typically < 1 ns. Photodisruption and therma l induced ablation will occur on a time scale of 1 s down to 1 ns. Photomechanical effects will occur at relatively the same energy densities therefore, exposure time is the main parameter used to distinguish between different effects. Photoablation occur s so quickly that thermal conductivity and thus thermal damage effects have almost no impact. The laser energy is used to cause a phase transition resulting in the tissue evaporating The applied electric field, when comparable in size to the average atomic or intramolecular Coulomb electric field, will cause tissue breakdown. Again, the extent of this process depends on the tissue properties. This type of ablation is localized within the laser beam spot and can be used for tissue sculpturing. Power densitie s below the threshold value cannot result in spontaneous tissue removal and the absorbed ener gy will be converted into heat. 1.3.3 Photothermal Effects Photothermal effects can be induced in tissue during laser tissue interactions. Photothermal effects inc lude a large group of interaction types where increased temperature is the main parameter. The degree and extent of photothermal effects is determined by the optical and thermal properties of the tissue as well as the laser parameters. Heat generation with in the tissue is mainly determined by the wavelength dependency absorption of the tissue. The temperature rise within tissue is associated with the photon energy converting to kinetic energy through a combination of nonradiative processes and vibrational r elaxations. Thermal effects will result, in most biomolecules, since a large number of accessible vibrational states are available to facilitate absorption as well as a large number of channels are available for the rmal decay and deactivation. Many lasers used for surgical applications fall within this category of converting laser light into thermal energy. When tissue is irradiated by a laser beam, heat is generated inside the tissue and the absorptivity of the tissue can change. In the example of the corn ea exposed to
36 UV radiation, this change is associated with both the formation of radicals (whose absorption cross section exceeds that of the original macromolecule links) and the development of light scattering as result of water in the cornea starting to boil. If heat moves faster than molecules can diffuse, thermal energy can be deposited into high absorbance regions through short laser pulses, and this is known as selective heating. During selective heating, the laser energy is absorbed by a targeted ch romophore, leaving the surrounding tissue largely undamaged. The time needed for the targeted chromophore to cool after laser irradiation will be proportional to the strength and number density of the chromophore s, and is known as thermal relaxation. Therm al relaxation time during thermal decomposition is important since it measures the thermal susceptibility of the tissue. The temperature that tissue reaches locally during and after exposure to a laser determines the location and extent of thermal effects with the tissue. Heat transport within the tissue is associated with the thermal tissue properties, and l osses of heat during laser tissue interactions also need to be considered. For most laser application the loss of heat is associated with conduction. H owever, estimation of the spread of energy by thermal conductivity is difficult since tissue layers can strongly differ in structure or complicated geometries are involved. Lowering water content will generally result in a decrease in heat conductivity and heat capacity. Living cells have a highly ordered structure that is stable at body temperature (~37 C). As the tissue temperature increases, the duration and peak value of the tissue temperature will govern the photothermal effects achieved. These effect s include coagulation, vaporization, ablation, and carbonization of tissue. Inflammation, cellular injury and cellular enzyme deactivation can occur to tissue as the temperature is raised by ~5 10 C. Tissue generally, can reach up to a temperature of ~45 C before irreversible damage occurs if exposure time is not
37 limited. Older tissue has a higher density of cross linking; therefore, higher temperatures are required to undergo these transitions. Further heating of tissue to ~45 50 C will result in molecu les within the tissue entering into an energy activated state and transitioning into an irreversible damaged state. Absorption by tissue at this range will result in weak hydrogen bonds and van der Waals interactions that stabilize the helical configuratio n of the chains in collagen being overcame. Shrinkage of collagen fibers in tissue will occur as a tensile stress is developed along the fibrils as well as enzymatic changes and the development of edema. Spatial arrangement of protein molecules will be los t and membrane alterations can also occur at this temperature range. 18.104.22.168 Coagulation o f t issue Tissue will begin to coagulate and denatur alize as temperatures approach ~50 60 C. At this temperature range, the helical structure of the chains in collage n transform into a denatured structure that is randomly coiled. Most protein molecules will lose their ability to function in the cell and also become denaturalized. In thermal denaturation, both the magnitude and duration of thermal exposure can be interc hanged to coagulate tissue. Either high temperatures at low heating times or long heating times at low temperatures will produce the thermal energy for tissue temperature to become ~50 60 C. It is exactly the denaturation of protein structures that determ ines the morphology of laser cuts. Successive laser pulses will result in the coagulated zone deepening within the tissue. Heating of tissue beyond 60 C will result in a disintegration of collagen where the coagulated tissue becomes necrotic, hydrogen bon d breaking and retraction will occur. After a relaxation of the stresses that develop during shrinkage, the tissue structure will experience total mechanical failure. Above 70 C denaturation of DNA occurs and at temperature higher than 80
38 C equilibrium o f chemical concentration is destroyed and membrane permeability is drastically increased. Depending on the nature of the tissue, there may be a delay in tissue necrosis. 22.214.171.124 Vaporization of t issue Once tissue reaches a temperature of 100 C, vaporizati on will occur. Water within tissue and cellular protoplasm will vaporize and produce thermal ablation of the irradiated tissue. As water within the tissue tries to expand in volume as it vaporizes, an increase in pressure will result and lead to localized micro explosions. The vapor generated will carry away excess heat. Phase transition of the tissue will be accompanied by bubble formation (micro cavitation) and thermal decomposition of the tissue fragments. 126.96.36.199 Carbonization of t issue Carbonization of tissue will occur once tissue reaches approximately 150 C. Tissue will char as its organic elements convert into carbon. Carbon is released and blackening of the tissue occurs. Carbonization of tissue currently has no benefits and results in irreversible damage. Since tissue will become necrotic at lower temperature and carbonization reduces visibility, it should be avoided in any case. When tissue reaches temperatures beyond 300 C, melting can occur. Melting of tissue can be used for tissue welding. 1.3 .3.4 Ablation of t issue Tissue temperatures between 300 C and 1000 C are achieved during ablation of tissue with smoke generation. Ablation as a pure thermal effect is produced by the pressure buildup from steam formation. The tissue matrix elastic resto ring forces and the surface tension can be overcame by the vapor pressure within the tissue. Bubble growth will occur within the tissue as the water within the tissue boils. The temperature of the tissue is above the saturation temperature and bubbles will continue to grow. As the pressure continues to increase, tissue that cannot withstand the deformation and stresses associated with the phase explosion will result in
39 material ejection since the ultimate tensile strength of the ECM is exceeded. Evaporation followed by the expulsion of the evaporated material is the ablation process. Thermal decomposition includes ablation. In ablation, the tissue is torn open and ejected by the expansion of steam leaving behind craters in the area. Ablation events will dep end of the laser irradiance, radiant exposure, optical tissue properties and mechanical tissue properties. The vigor of the ablation process is correlated to the water content of the tissue. Lateral tissue damage away from the ablation site can be created by blood vessels transporting heat. Damage to surrounding tissue can also occur through scattering. Ablation with little to no thermal damage can also result, notably tissue ablation with UV lasers. The ablation efficiency at high radiant exposures will be limited by the scattering and absorbing of the incident light by the ablation plume which reduces the amount of energy that can be deposited to the target tissue site. A recoiled induced tissue expulsion can occur after the primary tissue ejection and is largely determined by the mechanical strength of the tissue. Fo r soft tissue ablation occurs via a phase explosion of the tissue water since the tissue matrix is weaker than the pressure developed during the phase separation. When the ablation front reach es a depth where the temperature is below the stability limit of the superheated tissue water, ejection will stop. In mechanically weak tissue, ejection can be followed by a recoil induced material expulsion and by a dramatic mechanical deformation of the non ablated material. A recoil pressure can be induced in soft tissue by the primary ejection of ablation products resulting in a second expulsion of the tissue. This secondary tissue expulsion increases the ablation efficiency and the side effects of abla tion. If the speed of ablation is more important than precision and damage to surrounding tissue, ablation efficiency can also be increased in soft tissue by reducing the laser pulse duration. Recoil stresses can result in a delay in ablation of
40 deeper tis sue layers and ablation continuing beyond the laser pulse. In mechanically strong tissue, like skin, no secondary expulsion effect will occur. The tensile strength of the tissue matrix in mechanically strong tissue is higher than the stresses resulting fro m the phase explosion. Ablation of strong tissue occurs only after pressure is built up through confined boiling. The ablated tissue can be re deposited onto the surface and is typically unwanted. Flow components parallel to the ablated tissue surface can be monitored to prevent re depositing. A primary application of ablation is phototherapy or photothermal ablation (PTA). Phototherapy is used to treat various medical conditions. Photochemical ablation is one of the most successful techniques for refractiv e corneal surgery, where the refractive power of the cornea is altered in myopia, hyperopia, or astigmatism. Cancer cells can be directly destroyed by heat generated through the absorption of light within the tissue. PTA can be performed repeatedly without the accumulation of toxic side effects and resistance. PTA is a localized physical treatment and therefore would have fewer side effects than conventional cancer treatments. Photothermal effects have been shown to be enhanced significantly, therefore incr easing the efficiency of PTA, when a light absorbing material is applied to the target tissue. Photothermal coupling agents are used to mediate increases in the local temperatures far above the threshold at which irreversible cell death occurs. A study by animal PET of tumor damage induced by photothermal ablation with Cu Bis DOTA 2011) showed that CuS nanoparticles can be used in partnership with a NIR laser to cause significantly greater cell death to BT474 tumor ce lls than exposure to a NIR laser alone. In 2011, Song and coworkers investigated the potential application of small molecular weight Cu labeled bis DOTA Hypericin to increase the efficiency of PTA. Hypericin is a naturally occurring photodynamic agent that induces cell necrosis when combined with light.
41 Hypericin will bind to phospholipids that are exposed to the exterior of the cell when the cell membrane is damaged, which is the case in cell necrosis. Since phospholipids normally reside in the interior of the cell membrane, the authors hypothesized that hypericin derivative can be used for early assessment of PTA induced necrosis. MRI was used to provide real time temperature sensitive images during the therapy. Breast tumor (BT474) cells where treated wit h photothermal ablation therapy In vitro results showed that treated BT474 cells had a higher uptake of Cu Bis DOTA Hypericin than untreated cells. Autoradiography and histology results also showed that radiotracers where selectively applied to the necrot ic zone of the tumor that was induced by PTA. Therefore, this study was able to show that Cu Bis DOTA Hypericin can potentially be used to image tumor cell damage induced by photothermal ablation therapy. Thermotherapy carried out by the transpupillary rou te is known as transpupillary thermotherapy (TTT). Transpupillary thermotherapy is an accepted conservative treatment for choroidal melanoma. In TTT, the laser beam is directed on the apex of the tumor through a dilated pupil and is performed at temperatur es that fall within the coagulation zone. Therefore, TTT can result in an irreversible cytotoxic effects and damage to the tumor cells. The interaction of neural tissue and a flexible CO 2 laser fiber for the application of neurosurgery was examined and com pared by Ryan and coworkers in 2010. The wavelength of the CO 2 laser is 10.6 m and is highly absorbed in tissue and water, independent of pigment. A major application of CO 2 laser energy is for resection and ablation of tumor tissue. A rapid conversion of light to heat in very small volumes of tissue with minimal penetration to surrounding areas is therefore possible. Experiments were conducted using a swine cranium; its large animal brain tissue can simulate humans. The flexible CO 2 laser was also compare d to conventional methods for neural tissue incisions by the authors. The thermal tissue injury lateral
42 to the incision was minimal and did not increase significantly with higher power settings of the flexible CO 2 laser. Clean pia (delicate innermost layer of the membranes surrounding the brain and spinal cord) in cisions without hemorrhaging were produced with the focused CO 2 laser and the depth increased as a function of wattage. Wide charred pia incisions were produced when the CO 2 laser was unfocused. At high magnification, the desiccated tissue was observed to be blood cell debris free. When compared to bipolar cautery, the total tissue area affected by thermal energy was much higher than with the CO 2 laser. When axons suffer thermal injury, they develop swelling and balloons and these nerve fibers are vulnerable to degeneration and death. Therefore thermal injury must be kept to a minimum and the minimal injury to adjacent tissue was produced by the flexible CO 2 laser. In corneal ablation, high precision as well as minimal adjacent tissue damage are very solid state laser 213 nm and e xcimer laser 193 nm: A randomized, contralateral, comparative, experimental study UV lasers since corneal tissue strongly absorbs in the far UV region (190 200 nm). A solid state 2 13 nm laser was compared to an e xcimer 193 nm laser. Only small differences were observed between the two. For the e xcimer laser, ablated samples showed keratocyte dehydration, vascularization, and pronounced fragmentation of the cell bodies. Also, the e xcimer laser samples demonstrated normal epithelial thickness with mild hyperplasia and slightly thicker pse udo membrane than the solid state laser samples. The solid state laser ablated specimens demonstrated normal appearance of keratocytes across the entire corneal thickness. Also, the solid state laser samples demonstrated thinner than normal epithelium.
43 1.3 .4 Laser Tissue Interaction Examples Forme fruste keratoconus is an abortive state of keratoconus where the progression process has stopped at a certain point. Keratoconus occurs when the cornea thins and gradually bulges outward causing blurred vision and may cause sensitivity to light and glare. Since LASIK (laser assisted in situ keratomileusis) may trigger the halted progression of keratectasis, crosslinking of the cornea was introduced as an approach to stop the progression of keratoconus into a forme fruste keratoconus cornea. Surface ablation has shown good refractive results and is less invasive regarding the biomechanical stability of the cornea in keratoconus cases. Kaller and collagens in 2005 published clinical results of customized topography gu ided surface ablation in patients with forme fruste keratoconus. Rehabilitation of the eyes by means of customized topography guided surface ablation was the main goal of this study. A significant reduction in the refractive error, corneal irregularity, an d ghosting for all patients was observed in comparison to their preoperative statues. Femtosecond laser pulses for nanosurgery of cells and biological tissue has been reported by Vogel and coworkers in 2005. These authors used low density plasma to help ex plain the mechanisms behind femtosecond laser pulses. Femtosecond laser pulses, because of the high precision and the very low energy threshold optical breakdown, create very fine effects with a spatial extent below the optical diffraction limit. Experimen tal parameters used for cell surgery and computed numerical results were compared and two modes of dissection were differentiated. In one mode, they determined that dissection using long pulses at MHz repetition rates are mediated by free electron induced chemical decomposition and not related to heating or thermoelastic stresses. Bubbles will cause dislocations far beyond the laser focus spot. In the second mode, they determined that intercellular dissection at kHz repetition rates rely on thermoelastic st resses that induce the formation of small transient cavities. Both modes of
44 femtosecond laser pulses for nanosurgery showed to achieve better precision in cell surgery than surgery performed using continuous wave pulses. A comparison between the thresholds for optical breakdown and dissection using nanosecond and femtosecond laser pulses has been published by Vogel in 2007. The author determined that for nanosecond pulses, the optical breakdown threshold depends mainly on the linear absorption at the laser focus. At nanosecond pules, plasma luminescence usually serves as an experimental breakdown criterion. For femtosecond pulses, no plasma luminescence in the visible region occurs. Also, for femtosecond laser pulses, the thresholds for optical breakdown and dissection are the same and the required irradiance is approximately two orders of magnitude larger than for nanosecond pulses. Also, the breakdown energy required during femtosecond laser pulses is approximately three orders of magnitude smaller in compa rison to nanosecond laser pulses. One advantage of femtosecond pulses over nanosecond pulses is when wavelengths in the IR are used; they can penetrate deep into tissue without compromising the precision of tissue effects because the wavelength dependence of femtosecond breakdown is weak. Finally, it was concluded that any thermal denaturation of biomolecules will always be mixed with free electron induced chemical effects and the latter will probably dominate.
45 CHAPTER 2 SPECTROSCOPIC METHODS Spectroscopy techniques can provide a large amount of molecular structure information rapidly. Since the absorption and emission of atoms and molecules must occur between well defined energy levels, every element has a unique emission spectrum. Each spectroscopic meth od can be distinguished by how it interacts with the sample material, the energy used and the component it determines. Nuclear magnetic resonance, vibrational and electronic spectroscopy methods all involve the absorption of electromagnetic energy and the promotion of the sample molecules from a ground state to an excited state. The following sections describe in detail the spectroscopic methods of vibrational and electronic schemes Infrared (IR) Raman, Ultraviolet (UV) Visible (VIS) and Fluorescence spectroscopy have all been proposed for biochemical and clinical analysis for in vivo screenings of diseased, (thereby avoiding a biopsy), population screening, early diagnosis, prognosis, monitoring of therapy and many others. Therefore, these techniques will be covered in more detail. 2.1 Vibrational Spectroscopy Raman and Infrared (IR) spectroscopy are techniques that exploit the normal vibrational modes that occur naturally and predictably within molecules. The photon energies associated with this region of the electroma gnetic spectrum can cause vibrational excitation, but are generally too small to result in electronic excitation. Molecules will experience a wide range of vibrational motions based upon the atomic composition, ( as mentioned in the previous chapter ). In a normal mode of vibration, each atom executes a simple harmonic oscillation about its equilibrium position. The center of gravity of the molecule does not move but all the atoms move in phase with the same frequency. Stretching, bending, rocking, scissoring and twisting are all
46 types of vibrational motion. The mass of the component atoms and the strength of the bonds involved will determine the frequency which a given vibration will occur. For a clear explanation, consider two completely free identical diato mic molecules. They will vibrate with identical frequencies but when the two diatomic groups are part of a molecule, they can no longer vibrate independently of each other because the vibration of one group causes displacement of the other atoms in the mol ecule. These displacements are transmitted through the molecule and interact with the vibrations of the second group. The resulting vibrations are in plane and out of plane combinations of the two diatomic vibrations. Unless the groups are widely separated in the molecule, the two frequencies may not be resolved since the coupling is very small. When this occurs, usually the higher frequency mode is the anti symmetric (out of phase) vibration and the lower frequency mode is the symmetric (in phase) vibratio n. Unless a combination of force constants and mass effects result in similar uncoupled frequencies, the vibrations are not coupled. The distinction between symmetric and anti symmetric modes is accomplished by means of depolarization measurements, where s ymmetric modes produce polarized Raman bands and anti symmetric modes produce depolarized Raman bands. The symmetric modes are always active in the Raman spectrum because they produce a change in the polarizability of the molecule. However, anti symmetric modes are not active in Raman scattering for molecules possessing a center of symmetry. Raman and i nfrared spectra of a large number of compounds that contain a particular functional group have certain features that appear at generally the same frequency f or every compound containing the group. These spectral features are then associated with the complementary functional group. Also, the exact position of the group frequency within the range can provide further information on the environment of the function al group. A substitute
47 atom or group creates changes in the distribution of electrons in a molecule and can be detected in the vibrational spectrum. Some changes involve electron distribution in a molecule and cause changes in the force constants, and as a result creates changes in group frequencies. Separate signals for each 3N 6 vibrational modes of a molecule are not always displayed in the spectrum. The signals observed can be a combination of vibrations, increased or decreased because of molecular inte ractions or decreased by molecular symmetry or spectrometer limitations. Dissimilar modes such as bending and stretching vibrations can be coupled if the frequencies of the vibrations are similar and the two groups involved are adjacent in the molecule. An example of this is in amides. The C N stretching vibration is similar to that of the N H bending mode and the interaction of these similar frequencies results in two bands in the spectrum. One band occurs at a higher frequency and one band occurs at a low er frequency than the uncoupled frequencies of the N H bending vibration and C N stretching vibration. These bands are known as Amide II and Amide III bands. As a note, the Amide I band is associated with the C=O stretching vibrational mode. Fermi resonanc e is another special case of coupling. Fermi resonance occurs from coupling of a fundamental vibration with a combination and results in a shift of group frequencies and the introduction of extra bands. When Fermi resonance occurs, there is a sharing of in tensity and this effect can be strong. Typically stretching frequencies are higher than the corresponding bending frequencies since a bond can bend easier than it can stretch or compress (i.e. less energy) The stretching frequencies of hydrogen bonds are higher than those to heavier atoms. Also, excluding hydrogen bonds, triple bonds have higher stretching frequencies than the corresponding double bonds, which have higher frequencies than corresponding single bonds.
48 A vibrational spectrum will display the transmittance on the vertical axis and the energy in inverse centimeters (cm 1 ) on the horizontal axis. The inverse centimeter is also known as the wavenumber since it is the number of wave cycles in one centimeter, and importantly, is directly proportion ed to energy. 2.1 .1 Infrared Spectroscopy Infrared (IR) spectroscopy is a rapid and simple method for obtaining the identity and structure of molecule s including organics. The absorption spectrum is a plot of the percentage of IR radiation that passes thr ough the sample vs. the wavelength (or wavenumber) of the radiation. The position and relative amplitudes of the absorption peaks or bands gives information about the molecular structure. There are two main types of instruments for IR spectroscopy, those a re Fourier transform (FT) and dispersive. FT IR spectrometers are based on the Michelson interferometer. Interferometers record the spectrum in the time domain, resulting in an interferogram which must be transformed to the frequency domain by means of a F ourier transformation to obtain the IR spectrum. In IR dispersive instruments, the spectrum is directly recorded in the frequency domain using a dispersive device. For a particular vibrational mode to absorb IR energy, a change in the dipole moment should occur with displacement, and a large change in dipole moment will result in strong absorption bands. Practically all organic compounds absorb IR radiation. IR spectra can be obtained from samples in any phase, but care must be taken to insure that the samp le solute or sample substrate does not obscure the sample spectral regions. For example, glass absorbs IR radiation and should not be used. Vibrations that are often weak in the Raman spectrum typically have strong IR absorption characteristics. Infrared s pectroscopy produces spectra of samples that absorb radiation in the infrared region of the electromagnetic spectrum, from about 200 to 4000
49 cm 1 IR absorption occurs when energy is transferred from the incident radiation to the molecule, and a quantum me chanical transition between two vibrational energy levels occurs. The difference in energy between the two vibrational energy levels is directly related to the frequency of the electromagnetic radiation. A feature observed in the spectrum is an expression of an absorption band in the sample. The most important transitions are from the ground state to the first excited levels and these transitions usually give rise to strong absorption bands in the IR. For example, the method of static and time resolved infr ared difference spectroscopy has been used to investigate carboxylic groups in greater detail using [4 13C] aspartic acid labeled bacteriorhodopsin (BR) by Engelhard and coworkers in 1985. In this study, data provided assignment of the helices in the amino acid sequences based on the neutron scattering experiments. For aliphatic amino acid side chains, hydrogen bonding, protonation state, and coordination of cations are the dominate factors for determining band positions. Infrared spectroscopy is one of a f ew techniques that can define the protonation state of side chains which has an important consequence for the electrostatic interactions in proteins. Also, infrared spectroscopy can observe side chains and the protein backbone in the same experiment, there fore allowing the comparison of the kinetics of the backbone structural changes with the side chain signals. Molecules adsorbed on surfaces as well as their structures can be identified using infrared spectroscopy based on the surface infrared selection ru le. This rule states that only vibrational modes with a component of the dynamic dipole moment along the surface normal will be observed. Interpretation of the spectra traditionally relies on comparisons with spectra of known molecular species or using the ory to assign surface vibrations to particular molecular vibrations.
50 Theoretical spectra reproduce the general qualitative features of the experimental spectra in terms of approximate frequencies and relative intensities. The use of theory to assign the su rface vibrational spectra greatly enhances the reliability of those assignments and provides additional insights that cannot be obtained from only the experimental spectra. However, when using theory to assign surface vibrational spectra, the relative inte nsities are generally not in complete agreement with experiments. This is mainly because the theoretical spectrum does not take into account the intermolecular forces within the molecule. Comparisons between IR spectra are often difficult. Baseline distort ion has been shown to make it difficult to differentiate between small bands in a spectrum and background noise. Also, small deviations in band positions and band intensities can produce artificial bands in the region of symmetric vibration. 2.1 .2 Raman Sp ectroscopy Raman spectroscopy is also a vibrational spectroscopic technique. When the energy of an incident photon is unaltered after interaction with a molecule, the photon is scattered and has the same frequency as the incident photon, this is known are Rayleigh scattering or elastic scattering. If energy is transferred from either the molecule to the photon or from the photon to the molecule, the scattered photon has less or more energy than the incident photon, respectively, in a process is known as Ram an scattering (i.e. inelastic scattering). When energy is transferred from the photon to the molecule, the molecule will be left in a higher energy level given rise to Stokes lines. When energy is transferred from the molecule to the photon, the molecule m ust already be in an excited state so that it can return to a lower state after giving up energy to the photon, giving rise to anti Stokes lines. Accordingly, if a photon has a lower frequency (i.e. lower energy) than the incident light, a change in the vi brational mode of the sample molecule will occur, this is known as the Stokes Raman shift. Simultaneous absorption of an incident
51 photon and the emission of a Raman scattered photon with a shift in wavelength will correspond to vibrational level energy dif ferences in the molecule (Raman shift). As noted above, Raman scattering occurs at wavelengths shifted from the original incident light by the energies of the vibrational mode of the sample. The change in the polarizability of an induced molecular dipole m ust result from vibrational motion for Raman scattering activity. Raman spectroscopy provides a characteristic signal for certain molecular bonds. The position of the band depends on specific vibrational modes and is a very sensitive technique for determin ing differences in molecular structure. The plurality of Raman shifts in the region from about ~1000 to 1800 cm 1 is usually complex and unique to a particular material, d intensity as a function of the energy difference between the scattered photons and the incident photons are shown. Since each molecule has its own characteristic Raman spectrum, the Raman spectrum can determine the composition of the sample. The typical spectrum shows a series of peaks shifted by a unique characteristic vibrational frequency associated with a specific chemical compound. The concentration of the molecule is generally proportional to the intensity of the bands. Raman spectroscopy can be in vivo or in vitro and is a nondestructive vibrational technique for providing dynamic and structural information on molecules. A monochromatic laser beam is typically used as the incident light for Raman spectroscopy. The gain or loss in photon energy corr esponds to the difference in the final and initial vibrational energy levels of the molecules that are participating in the interaction. The Raman shifts are represented in wavenumber with the unit of inverse centimeter (cm 1 ), noting that the Raman shifts (i.e. positions) are independent of the excitation source wavelength. Raman peaks are associated with the vibration of a particular chemical bond or a single
52 functional group in the molecule and are specific and provide direct information about the compos ition One advantage of using Raman spectroscopy is it can provide information on the structure, concentration and interaction of biochemical molecules in their environments within intact cells and tissues. Raman spectroscopy has the ability to provide the above information in a non destructive way and without homogenization, extraction, or the use of dyes, labels of other contrast enhancing agents. Raman spectroscopy can also be used to study the interaction of biological molecules and nanoparticles. Nanop articles can pass through cell and nuclear membranes and are used to affect cell structure and cause damage to specific cells. Through nanoparticles designed to bind to particular biological molecules, individual molecules can be monitored as well as destr oyed in the case of cancer treatment. Raman spectra of solids when compared to those solids in solution are more complex and almost always sharper. In solid samples, the frequencies of the bands are shifted as compared to those in the liquid phase. When mo lecules are in a solution, the intermolecular forces are random and are variable since adjacent molecules are continuously colliding. The violence of the collisions and the shift of the frequencies of the intermolecular vibrations result in broadening of b ands. The weakness of Raman spectrum of water and the relative ease with which low frequencies can be measured are two advantages in molecules of biological importance of Raman spectroscopy for measuring vibrational spectra. 2.1 .2.1 Raman m icro spectroscop y Raman micro spectroscopy, in which both the excitation and the collection of Raman scattered light are done with the aid of a microscope objective, can be used to identify and grade human skin pilomatrixoma. Pilomatrixoma is a benign skin tumor derived f rom the hair matrix. Cheng and Liu in 2005 used Raman micro spectroscopy to confirm that collagen is the
53 predominate compound in normal skin and Pilomatrixoma samples contain a higher content of tryptophan than normal skin samples. In one study conducted b y Zhang and coworkers in 2008, Raman micro spectroscopy was used to investigate the regular variations in organic matrix composition of small yellow croaker otoliths (small concentrations composed mainly of calcium carbonate found in the middle ear of fish ) and their relation to fish calcification process. Otholiths structures are used for age determination on a daily or annual scale. By understanding the otholith calcification process, calcified biominerals such as shells and coral skeletons can be better understood. The results showed the versatility of Raman micro spectroscopy technique and provided insight into the calcification process. A combination of multimodal multi photon imaging and Raman micro spectroscopy can be used to characterization of struc ture constituent correlation of skin without labeling. Multimodal multi photon imaging allows the visualization of dermatological features and Raman micro spectroscopy allows the identification of the molecule. Contrasts between dermatological features and molecular specificity are produced without the use of fixation, sectioning or staining. Immediate illustration of detailed histological changes on intact tissues can be visualized using these techniques in combination. Therefore, the changes in response t o repeated physical or chemical injuries can be imaged subsequently on the same area. The ability to visualize multiple tissue structures and characterize the tissue on a molecular level allows monitoring of tissue changes associated with disease and could allow the outcome of different therapies to be investigated. 2. 1 .2.2 Resonance Raman s pectroscopy A particularly sensitive method of Raman spectroscopy is known as resonance Raman spectroscopy. This method is often used when the vibrations are very weak i n regular Raman as
54 these weak vibrations are strongly enhance and measured. Resonance Raman spectroscopy uses absorption of energy from the excited source at or near an electronic transition in the molecule. Raman intensity will increase as resonance is ap proached and the electronic transition closest to the excitation frequency should contribute the most for similar transition moments. Also, the relationship between resonance enhancement of a vibration mode, the atomic displacement, and the molecular elect ronic transition can also be examined. Resonance Raman spectroscopy can also be used to obtain a better understanding of electronic excited state. Coupling between peptide amide groups in proteins may reveal the roles of peptide bonds through which ground state electron transfer or photo induced transfer occurs in proteins. Photochemical studies of aqueous dipeptides suggest that a photo induced transfer from the carboxylate group to the peptide bond occurs upon UV irradiation, as related below. Kinetic UV resonance Raman spectroscopy measurements of the pH dependency of the activation barrier for trans cis isomerization of glycine glycine (Gly Gly) has been reported by Holtz, Li, and Asher in 1998. The charge state dependence of the trans cis ground state e nergy difference was determined. Even the presence of a cis amide linkage was found to significantly affect the protein structure. Cis trans isomerization can also slow the folding and refolding dynamics in proline and non proline containing proteins. Thei r results suggested only a modest impact of the charge state on energy differences and activation barriers for simple amides. This was surprising since a significant difference in the activation barrier between the different charge states due to entropic d ifferences in water ordering was expected. The lack of differences may have resulted from canceling effects. The insertion of drugs to DNA has been studied by Benevides in 2008, using resonance Raman spectroscopy. Resonance Raman spectroscopy was chosen be cause it is sensitive to the
55 structure of drug/DNA compounds, and it is selective, permitting only the observation of bands corresponding to the vibrations of the chromophore framework. This Raman technique has two main limitations. Those are: the laser wa velength required to make the excitation can only occur in the electronic absorption band of the compound and the fluorescent effects of the samples can sometimes over shadow the Raman peaks. Single crystal UV near resonance Raman spectroscopy can be used to probe electronic transitions. Vibrations coupled to the electrons involved in a particular electron transition can be probed by excitation of the Raman spectra in resonance with a particular electronic transition. If the particular vibrations probed are localized to a particular region of the molecule, the atomic localization of an electronic transition can be detected. In one study, (Chen, 1996) single crystal UV near resonance Raman spectroscopy was used to investigate the dipeptide Gly Gly. They grew a crystal of hydrated Gly Gly, determined its crystal structure and measured its Raman tensor close to resonance. A general method to use pre resonance single crystal Raman measurements to determine the orientation of the transition moments of the resonant excited state and the next closest excited state as well as the direction of the molecular electronic transition moments was developed. For most non resonance spontaneous Raman scattering experiments, pulsed laser sources should be avoided whenever a cont inuous wave (CW) laser source is available. This is because pulse lasers have much higher peak powers and this may result in significant non linear optical phenomenon such as, increased thermal, photochemical, and photo thermal damage, especially in solid samples. The heat from pulsed lasers can also lead to sample decomposition, artificially changed samples, and phase transitions. For solid samples, semiconductors, and biomolecules, as well as samples that strongly absorb the laser, this problem might be e ven more overwhelming.
56 Even for CW lasers if the power is too high or the measuring time is too long thermal damage problems can occur. However, pulsed lasers could be used as the excitation source for Raman spectroscopy. For thermally unstable solid sampl es, pulsed lasers can be used as long as the sample has no strong electronic absorption at the laser wavelength and the energy of the laser is kept below a certain threshold. In one study conducted by Sieler and c oworkers, the depolarization ratio of Raman bands was used to infer the electronic transitions which correspond to the coupled molecular vibrations. Polarized non resonance and resonance Raman and FTIR spectroscopy were used to measure model peptides in H 2 O and D 2 O at pH/pD values between 1.5 and 1 2.0 with visible, near UV, and far UV excitation wavelengths. In Fourier transform Raman instruments the sample replaces the source lamp of the near IR spectrometer. The most important results were for Gly Gly and showed that all amide bands as well as the symmetric carboxyl stretch band are doublets. The amide I doublet resulted from vibrational coupling of the delocalized H 2 O bending mode with the internal coordinates of the amide I mode. The amide III doublet resulted from vibrational coupling between th e twisting mode of the C alpha group and the internal coordinates of the amide III mode. 2. 1 .2.3 Surface enhanced Raman s pectroscopy Raman spectroscopy has also shown to be a valuable instrument in the field of surface science. Greatly enhanced cross secti ons of Raman scattering from molecules adsorbed on a metal surface was an important discovery that lead to the technique known as surface enhanced Raman spectroscopy. Surface enhanced Raman spectroscopy (SERS) has an advantage over other techniques since i t can provide high quality spectra of samples in water solution and there is a reduction in the fluorescence background which is often a problem in Raman signals of
57 biological molecules. Detailed analysis of spectra can be compared and Raman bands can be a ssigned for specific functional groups contained in these molecules. SERS also has an advantage in biological applications since it reduces the fluorescent background which is often a problem in regular Raman scattering of biological molecules. The mechani sm of SERS is not completely understood but can be treated as an electromagnetic or chemical enhancement. When considered as an electromagnetic mechanism, the resonant interaction of the incident laser light with the surface plasmon of the SERS active subs trate is thought to be the cause. When considering SERS as a chemical enhancement, the charge transfer transition between the metal surface and molecules adsorbed on it is thought to be the cause. SERS however, does have a few disadvantages. The experiment al conditions of SERS, such as metal particle size, shape of the metal, and the surface potential of metal colloids are hard to control. As well as the structure of adsorbed molecules or the nature of adsorption and distribution of the electromagnetic fiel d on and near the metal surface. Also, vibrational modes are not intensified equally in SERS. Surface enhanced Raman spectroscopy, however, does allow the measurement of most samples at very low levels of detection and provides the opportunity to investiga te the orientation of molecules. Surface enhanced resonance Raman spectroscopy (SERRS) is a technique used for the identification of components of a mixture without separation. SERRS has been shown to be a more sensitive and specific detection method for t he analysis of DNA when compared to fluorescence. In the detection of DNA with fluorescence is limited due to the broad and overlapping of the spectra. 2. 1 .2. 4 Raman s pectroscopy of p eptides and p roteins Vibrational analysis of dipeptides containing alanin e and serine has been studied by Mukhopadhyay in 2008, using Raman spectro scopy. The spectra recorded contain several
58 overlapping bands, so the authors used theoretical geometries for all the dipeptides investigated. The dipeptides were constructed from a library of amino acids and the dihedral angels were varied to obtain an energy minimum. Significant differences between the experimental and calculated frequencies were obtained as well as limited success in reproducibility of the spectra. The lack of corr elation between calculated and experimental frequencies in the amide I and amide II regions could indicate that these frequencies are influenced greatly by hydrogen bond and intermolecular coupling effects, showing a strong argument for non bond interactio ns occurring. Finally, Gauss Cauchy product functions were used to identify the correlating band assignments and the frequencies. Amide III modes have been proven to be a combination of NH in plane bending and CH deformation vibrations. Amide III vibration s also produce a strong Raman scattering effect and can be related most directly to conformational features of peptides. glycine at different pressures in 2011 by Sharma and coworkers. To pressures up to 23 GPa, no structural phase transitions were observed. The changes observed occurred in the slope of NH 3 torsional and CO 2 bend ing modes across 3 GPa suggests subtle structural rearrangements. This results in changes in the nature of the intra layer hydrogen bond interaction across this pressure. As in all biological materials, changes in the hydrogen bonding interactions play a s ignificant role in structural conformation, though not always resulting in a structural phase transition. Splitting, broadening and intensity redistribution of various Raman modes under pressure may be due to changes in the intermolecular coupling. Amino a cids, peptides, and proteins can be modified without any damage to the molecular structure or conformation using a nucleophilic addition reaction. Raman spectroscopy can be used to discern site specific reactions on the amino acids alpha amines or side gro up
59 amines. Determining where the modification has taken place is a significant step in the study of all protein modifications. Quantitative chemical results can also be achieved using Raman spectroscopy, thus, offering a distinct advantage over ultraviolet visible assay by showing exactly the reaction sites. Another method for accurately describing peptide conformation behavior involves assuming the solvent is continuous. However, this model may not be adequate for the description of peptide conformational behavior because of the lack of the explicit directional hydrogen bonds. Also, for dipeptides glycine proline (Gly Pro) and alanine proline (Ala Pro), a small increase of population in Pro isomers is observed with rising temperature. The presence of the pr oline ring provide an opportunity to study conformation of peptide side chains which are often undetectable by other spectroscopic techniques; however, there is a limited influence of the side chains on the backbone conformation. Raman spectroscopy can pro vide diagnostic evidence for reverse turn structure when the amide I and amide II bands are analyzed. Reverse turn structures can provide important conformational components of protein structures. In the amide I region, reverse turns have transitions which helical structure. Also, in small sheet structures. Portable Raman instruments are also available for protein ana lysis. Portable Raman instruments are designed to acquire data in field conditions. Pure amino acids in field conditions have been studied be Culka and coworkers in 2010. This study collected Raman spectra of astrobiologically important amino acids using t wo portable Raman instruments in outdoor conditions. The noise level in the spectra of both instruments was similar when measured during
60 handheld use. Signal to noise was significantly reduced in both instruments when a custom made holder was used. Their r esults showed that both instruments provided good quality Raman spectra that can be used without additional spectral manipulations for the detection of relevant amino acids. The bandwidth of the Raman bands was larger in the portable instruments than in ty pical laboratory instruments, but not so large as to lose spectral information. Insight into problems in protein structure and misfolding of individual peptide bonds in diverse environments can be obtained by Raman spectroscopy. A general strategy has been proposed by Burth in 2001 which is based upon isotope edited Raman difference spectroscopy. Using Raman spectra of unlabeled human insulin, three set of crystals containing C in different locations in the insulin chain has been studied. By subtracting the spectra of known labeled and unlabeled crystals and labeling two individual peptide links, the precise Amide I location and the local secondary structures of those bonds were pinpointed. This strategy can be used for other proteins providing that part of the polypeptide chain is in an ordered environment. turns which are important conformational features of proteins and peptide s. Normal vibrations of peptide potential ener gy distributions and the corresponding Jacobians allow the composition of the observed amide frequencies to be established. 2. 1 .2.5 Raman s pectroscopy of t issue Raman spectroscopy has been utilized in the investigation of pathophysiological transformations in tissues and life cycles of single cells by Volkmer in 2005, since it allows the characterization of molecular structures. Coherent anti Stokes Raman scattering (CARS) microscopy has be developed as a new method of multi photon imaging. This form of ima ging is based on signals that are specific to a particular molecular vibration. CARS imaging therefore creates a chemical contrast without the use of labeling. Simultaneous acquisition of images based
61 on multi photon imaging can allow the visualization of various tissue components with well defined localization and contrast. Multi photon microscopy is non destructive, deep penetration and is a powerful tool for biomedical imaging. Some other examples of multi photon microscopy are: two photon excited fluore scence (TPEF), second harmonic generation (SHG) microscopy, and sum frequency generation (SFG). Cell proliferation can also be monitored by Raman spectroscopy. Cultures in the plateau (non proliferating) and exponential (proliferating) phases of growth and estimation of the relative amount of biochemical components in cells and nuclei can be determined through the use of Raman spectroscopy. Relative amounts and ratios of biochemical components can be used to detect and quantify reproducible differences and to determine specific biochemical changes that can be assigned to differences due to cell proliferation. Raman spectroscopy has the ability to obtain biochemical information from proteins, nucleic acids, lipids, and carbohydrates in a single acquisition. W hen appropriate choices of laser wavelength and power are made, Raman spectroscopy is a noninvasive and nondestructive technology that can be applied to clinical patients in vivo The analysis of single cells in vitro can also be performed since Raman spec troscopy uses high power focusing objectives, confocal optics, and sub micrometer resolution stepping stages. Biological and biochemical differences between cells in a given population is expected since cells exist at different points in their cell cycle a nd cells grown in cultures can have different confluences. These differences result in spectral variations and in the past, was either averaged during spectral acquisition or post processing. Spectral differences between cell populations can be very subtle and single cell studies have focused on the differences in Raman spectra between different cell types for example, healthy vs. tumorous. Other Raman microscopy
62 studies have been performed for analyzing a population of a single type of cell for biochemical differences. Detected spectral changes from cell death via apoptosis and necrosis as well as spectral differences between live and dead cells and exponentially growing and plateaus phase cells have all been reported by Notingher and Henery in 2006. Raman spectral study of single live cells can be conducted to discriminate between cells synchronized in the G0/G1, S, and G2/M phases of the cell cycle and accurately detected the difference in G0/G1 cells from S and G2/M cells. Raman microscopy has the capabil ity to detect inherent sources of biochemically based spectral variability between single cells of a human tumor cell line cultured in vitro The most significant source of Raman spectral variability between cells in a culture can be confidently attributed to biochemical changes arising from the progression of individual cells through their mitotic cycle. Another application of Raman spectroscopy is that it can be used to identify molecular changes associated with diseased tissue, creating a unique tool for research in areas related to cell biology and natural tissues. Real time diagnosis of lung cancer has been performed using near infrared Raman spectroscopy and has been reported by Huang and coworkers in 2003. The Raman spectra differed significantly betw een normal and malignant human tumor lung tissues as well as a difference in parts of lung carcinoma. The tumors showed higher percentage signals for nucleic acids, tryptophan, and phenylalanine, in comparison to normal tissue. The tumors also showed lower percentage signals for proline, phospholipids and valine, in comparison with normal tissue. Basal cell carcinoma is the most common type of skin cancer and has been studied by Gniadecka in 2008 using Raman spectroscopy. Normal skin and basal cell carcinom a verified by histopathology was harvested and analyzed in one study, using near infrared Fourier transform
63 Raman spectroscopy. In the skin cancer samples, alterations in proteins and lipid structures were apparent. Spectral changes were observed in protei n bands, amide I, amide III and CC stretching and in lipids, CH 2 scissoring vibration and CH 2 in plane twist vibration. Some disadvantages of Raman spectroscopy include the effect of stray light, which if not removed, can produce a high background signal t hat can saturate the Raman spectrum. Also, weak Raman bands can be masked by background scattering from small suspended particles in solution or liquid samples. If the sample fluoresces under visible light, the Raman spectrum may be totally obscured. Howev er, the problem of fluorescence emission from the sample can be reduced or eliminated when the laser excitation is in the near infrared region or the red region. Also, when the laser beam focus is increased, the Raman intensity increases but this may damag e the sample. 2. 1 .3 Coupled Raman Spectroscopy and Infrared Spectroscopy Vibrational frequencies may give rise to IR absorption, Raman scattering or both. When vibrational frequencies give rise to both, observed wavenumbers will be numerically the same in Raman and IR, but the intensities will in general be very different. The intensity of Raman line is a linear function of concentration and the intensity of IR absorption band is a logarithmic function of the concentration. Raman spectroscopy can provide in formation complementary to that obtained from infrared spectroscopy. The use of experimental Raman and infrared data as well as normal coordinate analysis has been shown to provide vibrational assignments to frequencies of the lowest energy non planar tran s conformer molecule structure. This idea has been applied to study phenylurea and phenylthiouea molecules by Badawi in 2008. Also in this study, it was reported that the effect of the steric interactions between the CH 3 and NH 2 groups in destabilizing the planar structure should be considered. The CH 3 group seems to hinder the conjugation between the C=C and
64 C=O groups and therefore weaken the bond character of the C C bond in the methyl substituted compound. Raman scattering and infrared absorption techni ques can also be used to determine molecular sites in which interactions occur between peptides and other molecules. Hydrophobic amino acids play an important role in protein folding. In soluble proteins, a hydrophobic pocket is created that is hardly acce ssible by the surrounding water. Information on the secondary structure of peptides and proteins in aqueous solutions can be analyzed by the Amide spectral regions. The Amide spectral regions are highly sensitive to peptide backbone motions. Understanding protein structure and folding can be accomplished by determining conformation preferences of dipeptides in aqueous solution. Depolarized Rayleigh scattering and Raman spectroscopy has been used by Avignon and coworkers in 2004 to study glycine and alanine dipeptides in aqueous solution. As well as hydrophobic interactions between side chains, residue steric interactions, screening of backbone electrostatic interactions and side chain conformation entropy by solvents have all been suggested to be the cause o f conformational preferences of residues in polypeptides. It was shown that the only electrostatic screening model was able to explain the conformation preferences of dipeptides. The frequency of the Amide I band is the most accurate determinable vibration al spectra parameter for peptides in an aqueous solution. The Amide I band depends on the position and intensity of the components which are determined by the dipeptide conformation. This vibration can be used to distinguish other conformations. 2. 2 Electr onic Spectroscopy When the energy of the incident radiation equals the electronic transition of the molecule, absorption occurs and this is known as electronic excitation. Electronic excitation typically involves a single electron moving from an occupied t o an unoccupied orbital resulting in the
65 molecule moving from a ground state to a higher electronically excited state. Many different electronic transitions are possible and generally electrons will be promoted from the highest occupied molecular orbital t o the lowest unoccupied molecular orbital. Electronic spectroscopy utilizes the interaction of electromagnetic radiation with the electronic transitions of the molecule being studied. Three types of electronic spectroscopy are discussed below. Those includ e ultraviolet and visible, chrioptical, and fluorescence spectroscopy. 2.2 .1 Ultraviolet and Visible Spectroscopy Ultraviolet (UV) and Visible (VIS) spectroscopy is a highly sensitive technique used in structure determination to detect the interaction betw een groups and to study chemical reactions when functionality is altered. When an electron is promoted from a ground state to an excited state, the absorption of UV VIS is measured and typically displayed as a plot of absorbance intensity versus the wavele ngth. All organic compounds will absorb light in the UV region and some absorb light in the VIS region, therefore this method of spectroscopy can be helpful when organic compounds are studied. Molecules that are likely to absorb light in the UV VIS region are known as electron functions or atoms where non bonding valance electrons are found. These areas of light absorbing groups or functions are known as chromophores. UV VIS spectroscopy is the best method for identifying and detecting chromop hores. When UV VIS light is absorbed, dipole dipole coupling of chromophores at their excitation state can produced a delocalized excitation. Delocalized excitation can cause shifting, broadening and splitting of the spectra. These spectra changes can then be associated with the relative orientation and intensity of the dipole moments associated with the UV VIS absorption bands of each chromophore. Overlapping delocalized excitation transitions occurs when the chromophores interact.
66 2.2 .2 Chrioptical Spectr oscopy Optical rotatory dispersion (ORD) and circular dichroism (CD) spectroscopy are both types of chrioptical spectroscopy. Once an electron is promoted from a ground state to an excited state orbital, absorption is measured in ORD and CD spectroscopy by measuring the absorption of left and right circularly polarized light or measuring the refractive indices and taking the difference. CD spectroscopy is used to detect functional groups and to determine conformation of molecules. The spectrum is displayed as wavelength on the horizontal axis and the vertical axis can be displayed as positive or negative. CD spectroscopy has been used for chemical analysis, biophysical studies, and secondary structure of proteins studies. Chrioptical spectroscopy is typicall y limited to only optically active compounds. 2.2 .3 Fluorescence Spectroscopy Some species are capable of absorbing electromagnetic radiation and this causes the electrons to enter an excited state by absorbing the photons energy. From the ground state, if enough energy is available, the electron moves to one of the many vibrational states in the excited electronic state. As the specie transitions from a singlet excited state back to the ground state, energy is released in the form of light and this is know n as fluorescence (phosphorescence is produced if the transition is from an excited triplet state). If the electron spins are parallel, a triplet state is possible and usually forms by intersystem crossing from an excited singlet state rather than by direc t excitation from the ground state. If the electron spins are antiparallel the state is singlet. The energy gap between the excited state and the state that the electron falls will determine the wavelength of light that is emitted. The emitted photons will have different energies and therefore different frequencies. The wavelength of light used to excite the species and the emitted wavelength by the species is tailored to that specific species. Typically an ultraviolet light is used to excite the electrons in the species. The light being emitted from the
67 species occurs at a lower energy and this can be within or outside the visible spectrum. Also, the wavelengths of fluorescence are longer than vibrational absorbance. This is because vibrational relaxation o ccurs in the excited electron state and the electronic relaxation occurs from the ground state of the excited electron, causing lower energy electronic relaxations than energy of absorption. The different frequencies and relative intensities of the emitted light are used to determine the structure of the different vibrational levels. Very few species in a sample will fluoresce at a given wavelength, following a specific excitation wavelength, making fluorescence a very sensitive technique for detecting part icular species. At low concentrations the intensity generally will be proportional to the concentration of the fluorophore. Specie concentrations as low as one part per trillion can be detected under ideal conditions, making fluorescence spectroscopy more sensitive than absorbance spectroscopy, which are often used as complementary techniques. Successful single specie detection experiments, a high degree of selectivity achieved using a variety of photochemical and instrumental approaches, and flexibility in measuring a range of solid and liquid samples are all advantages in fluorescence spectroscopy. Fluorescence spectroscopy is a nondestructive and sensitive technique that allows the identification of different aromatic fluorophores with electron donating f unctional groups as well as the study of dynamic properties of the sample under different environmental conditions. The broadness and position of fluorescence peaks are related to structural components such as molecular size and content of functional group s. Conventional fluorescence techniques are helpful in the investigation of the overall composition of humic substances in terms of aliphatic and aromatic groups showing different structure.
68 Fluorescence spectroscopy is also well established for the analys is of molecular interactions. Because of its sensitivity to concentration, measurements of peptide solutions can be obtained at much lower concentrations than NMR measurements, thus allowing direct measurements of the peptide monomer. Fluorescence spectros copy can be used to follow conformational dynamics on a timescale down to a few nanoseconds. The interaction between the fluorophore and quencher interactions occur by reorientation through local diffusion inside the volume set by the molecular linker and yield an upper limit to the observable rate constants for polymers end to end contacts. The interactions can be used as evidence of conformational changes in macromolecules or macromolecular complexes. In aqueous solutions typically many fluorophore quench er pairs can induce formation of hydrophobic complexes. 2.2 .3.1 Fluorescence e xcitation e mission m atrix s pectroscopy When the light emitted by a sample is measured and the excitation light is at a constant wavelength, an emission spectrum is collected. An emission map is created when emission spectra from a range of excitation wavelengths are combined together. This results in emission intensity as a function of excitation and emission wavelengths. When the emission light wavelength is held constant and the excitation light is scanned through different wavelengths an excitation spectrum is collected. A widely used method for analyzing water samples of different origin and dissolved organic matter (DOM) in aquatic environments is fluorescence excitation emiss ion matrix (EEM) spectroscopy. DOM in ground and surface water consists of a variety of organic compounds. Humic substances and protein related substances are the primary sources of fluorescence in dissolved organic matter. Fluorescence spectroscopy gives a wide spectrum of information in very short time and is a very sensitive and selective technique. Fluorescence is
69 very sensitive to low DOM concentrations and to differences in properties and distributions of fluorophores between similar samples. The exci tation and emission ranges reported for humic substances are 280 380 nm and 350 600 nm respectively. Because of fluorescence sensitivity and selectivity, it can be applied to small samples without pre concentration or isolation and is a rapid, nondestruct ive method that allows subsequent analyses to be conducted after spectra are collected. Fluorescence spectroscopy can be used for determining qualitative aspects of organic matter character and functional properties. In one study conducted by Weishaar in 2 003, fluorescence spectroscopy was used to assess the effect of changing coagulation pH on organic matter character, removal and composition. Fluorescence excitation emission matrix (EEM) spectra display the intensity of fluorescence within the sample agai nst the excitation wavelengths and the wavelength at which excited fluorophores emit light. Natural fluorescence (from humic and fulvic like) and microbial derived organic matter fluorescence (from tryptophan and tyrosine like) occurs at shorter emission w avelengths. Chlorophyll fluorescence is found at longer wavelength and is indicative of algae activity. Results of this study show that fluorescence spectroscopy is an effective technique for the characterization of organic matter in a range of different t reatment conditions. For optimized coagulation conditions, a great reduction in fluorescence intensity between raw and clarified water was observed. Fluorescence EEM spectroscopy has been used to determine protein like, fulvic acid like and humic acid like substances. At low concentrations these materials are directly proportional to fluorescence intensity and thus fluorescence EEM spectroscopy can be used to assess compost maturity. The combination of fluorescence spectroscopy with regional integration ana lysis can also be used as a tool for determining compost maturity. Unfortunately, classical fluorescence
70 spectroscopy alone does not accurately measure the maturity of compost because it cannot differentiate from humic and non humic compounds. Therefore, a combination of fluorescence EEM with regional integration analysis was used to determine the degree of humic substances in the compost and hence the compost maturity. The role of quinones in redox reactivity of DOM is important for the cycling of metals a nd the breakdown of natural and contaminated organic matter in anoxic environments. In the reduction of humic substances by microorganisms, quinone moieties act as electron acceptors. Quinone like components are shown to contribute to the EEM spectrum. Uni que excitation and emission curves can also be used to characterize fluorophores. Fluorophores have been compared with model Quinone. Quinone like components are shown to contribute to the EEM spectrum. In general, every fluorescence spectrometer gives dev ice dependent signals. Research has been conducted to standardize signals based on emission and excitation corrections as well as quenching and normalization effect corrections. Standardization of data is important so comparisons of results can be performe d. Spectral correction, Raman normalization, and inner filter effect correction can be applied and compared to fluorescence excitation emission matrix measurements by different spectrometers. 2. 2 .3. 2 Fluorescence r ecovery after p hotobleaching s pectroscopy Fluorescence recovery after photobleaching (FRAP) experiments can provide information about the mobile fraction of fluorescence molecules and the rate of mobility in a defined compartment, which can be related to the diffusion time of molecules. FRAP can a lso be used to determine protein localization, activity, interactions, conformations changes and dynamics, as well as following events during cell division and signaling within living cells. In FRAP experiments, fluorescent molecules are photobleached in a small area of the cell irreversibly by a high powered focused laser beam. Diffusion of surrounding molecules that were not bleached
71 into the bleached area leads to a recovery of fluorescence and this is recorded at a lower laser power. FRAP can be used to measure the continuity of membrane compartments and the behavior of proteins during mitosis. FRAP spectroscopy can also be applied to measurements of membrane protein laterally diffused in single cells. Measurements of the lateral diffusion of integral me mbrane proteins labeled with fluorescein can show the rate of recovery of fluorescence after bleaching. Fluorescence has shown to rapidly return to spots bleached by a laser beam in the continuous fluorescence of cultured cells labeled on the surface. FRAP spectroscopy has been incorporated in one experimental system by Rabut and Ellenberg in 2005 for the characterization of structural and molecular mobility. In the experimental system the authors proposed, transport dynamics are measured by monitoring the transient behavior of an inhomogeneous distribution of concentration. A high laser power is exposed to the sample, irreversibly bleaching the fluorophores, producing a spatial distribution of fluorophores. The recovery of fluorescence is monitored with the same laser at a lower power as to minimize further bleaching. This technique was applied most successfully to highly purified solutions of biological macromolecules for measurement of molecular translational diffusion coefficients. 2. 2.3.3 Laser i nduced f luorescence s pectroscopy Laser induced fluorescence (LIF) spectroscopy can be used for the determination of amino acids and amines in plasma, cells, urine, serum, cell extracts, and biological tissue. This technique takes advantage of the sensitivity provi ded by laser excited fluorescence and high atom density and is useful for detection of a number of elements. In general, LIF is differentiated from fluorescence spectroscopy as discussed above, by the use of specific, targeted excitation/emission pairs.
72 In 2008, Kaneta and coworkers used LIF spectroscopy coupled with cyclodextrin modified capillary electrophoresis to determine amino acids in urine. Three types of amino acids were used in one study without the need of pretreatment and with the recovery range of 76 117 %. Conventional analysis was also performed in this study with the limit of quantitation to be ~10 M for all of the amino acids measured. For analysis of amines and other amino acids not presented, optimization of separation including buffer co ncentration, pH, and additives are necessary. A variation in fluorescence emission by distance dependent fluorescence quenching between a fluorophore and a quenching functional group can be monitored by photo induced electron transfer spectroscopy. As ment ioned above, PET requires van der Waals contact for efficient quenching. PET quenching can occur though dynamic quenching, occurring transiently through molecular collisions, or static quenching, occurring in molecular complexes that are stable for multipl e excitation emission cycles. The interaction between the fluorophore and quencher interactions occur by reorientation through local diffusion inside the volume set by the molecular linker and yield an upper limit to the observable rate constants for polym ers end to end contacts. The interactions can be used as evidence of conformational changes in macromolecules or macromolecular complexes. In aqueous solutions typically many fluorophore quencher pairs can induce formation of hydrophobic complexes. 188.8.131.52 Fluorescence s pectroscopy using f luorophore m anipulation Some biologically important molecules do not exhibit, to any great extent, analytically useful properties, such as a high molar absorptivity or electrochemical activity. For these molecules, sensiti ve direct detection is difficult. When a sample does not exhibit a measurable chemical characteristic, manipulation of the sample can be performed. For example, selective detection of proline is very important and useful for the diagnosis of proline relate d diseases.
73 High levels (10 15 times higher than normal) of proline in the blood can cause seizures or intellectual disability and is known as hyperprolinemia. A new fluorescence emitting compounds has been synthesized and characterized by Lerouge and cowo rkers in 2011. Photo luminescent neutral and anionic di carboranyl and tetra carboranyl fragments were bonded through CH 2 units on different organic functional groups and their influence on the final molecular properties have been studied. A carborane is a cluster composed of carbon and boron atoms. Two classifications of carborane are closo which represents a complete polyhedron and nido which represents a polyhedron that is missing one vertex. All the closo carborane and nido carborane derivatives exhib it a blue emission at room temperature under ultraviolet excitation wavelengths in different solvents. Also, the common fragment in all of these compounds is the CH 2 bridge that occurs between the aromatic ring and the carborane cage. This seems to prevent the key to photoluminescence properties, especially since compounds without this functional group do not display this feature. The closo carborane derivatives showed that the maximum emission depends on several factors such as: the solvent polarity, the substituent (phenyl or methyl) bonded to the C cluster, and the organic unit bearing the carborane clusters (benzene or pyridine). Emission results at room temperature showed that the phenyl closo carb oranyl exhibit maximum emission intensities between 369 nm and 371 nm, independent of the solvent and the methyl closo carboranyl exhibit maximum emission intensities between 333 nm and 363 nm, dependent on the solvent. 184.108.40.206 Fluorescence s pectroscopy of t issue Fluorescence, Raman, IR, and reflection/absorption spectroscopy are all techniques for in situ detection of organic and biological materials in real time that are non contact and noninvasive. Native fluorescence has been shown to differentiate clas ses or groups of organic
74 molecules and biological materials when excitation occurs at specific wavelengths in the region of deep UV. Excitation wavelengths between 200 400 nm with emission wavelengths between 270 500 nm can be used to compare fluorescence excitation emission matrix spectra of pure organic materials, microbiological samples, and environmental background materials. Excitation wavelength greater than 350 nm have shown to provide little spectral information about biological samples. Optimal exc itation wavelength(s) necessary for detection and differentiability of many organic and biological materials in different environments can be determined. In fluorescence spectroscopy and spectral imaging of tissue, several fluorophores present in tissue pr ovide important biochemical bases for optical contrasts. The fluorescent properties of most fluorophores present in tissue are correlated with certain pathological conditions and this allows tissue diagnosis through fluorescence spectroscopy. Some fluoroph ores that are in tissue include collagen, elastin, porphyrin, flavin adenine dinucleotide (FAD), and reduced nicotinamide adenine dinucleotide (NADH). FAD and NADH are two fluorophores that are electron carriers involved in the metabolic process of cells. Their fluorescence can be used to determine the reduction oxidation state of cells, as proven by Drezek and coworkers in 2005. NADH and FAD also usually have relatively low fluorescence quantum yields and exhibit relatively weak fluorescence at biological concentrations. Previous studies of cervical tissue show an increase in NADH fluorescence within the epithelium with progression of abnormal tissue. Tryptophan fluorescence was also reported to be useful in discriminating malignant metastatic cancer cells from non metastatic cells and inflammation from cancer. A simple method to rapidly measure fluorescence signals and spectral fingerprints of complex mixtures like tissue is known as synchronous fluorescence spectroscopy. Conventional
75 fluorescence spectrosc opy uses either a fixed wavelength excitation to produce an emission spectrum or a fixed wavelength emission to record an excitation spectrum, while in synchronous fluorescence spectroscopy a fixed wavelength excitation emission spectrum and the fixed wave length emission excitation spectrum are recorded simultaneously. A constant wavelength interval is maintained throughout the synchronous fluorescence spectrum. The synchronous method can effectively examine the fluorescence peaks of multiple tissue fluorop hores in one scan, reduce the overlap between fluorescence peaks, and suppress/express the effect of specific absorption peaks on the shapes of the fluorescence spectra due to hemoglobin by selecting an appropriate wavelength interval in tissue like phanto ms and ex vivo rat brains. One limitation of this technique is that the background fluorescence from fiber optic probes from un rejected excitation light could exert a great impact on the spectral shape for small wavelength intervals (< 40 nm) when compare d to conventional fluorescence spectroscopy. However, synchronous fluorescence spectroscopy has shown that it can greatly simplify qualitative interpretations of fluorescence spectra from tissue or tissue like samples. Fluorescence imaging in vivo is simil ar to fluorescence microscopy. They both use appropriate filters and low light camera to collect fluorescence emission light from samples. However, in vivo fluorescence imaging works at a macroscopic level and whole body small animals instead of cells in c ulture dishes or slides are imaged. This allows the visualization of fluorescent emission from fluorophores of biology in intact and native physiological states. Fluorescent organic, biological and inorganic nanoparticles can be used for in vivo fluorescen ce imaging. These nanoparticles serve as a platform to build multifunctional probes for multi modality imaging although traditional small molecule NIR dyes are currently being used.
76 Increases in the availability of fluorescent probes and reports of this te chnique are showing to be an important translational tool between basic research and clinical applications. Reabsorption by the sample can also result in changes in the spectrum. If a part of the sample absorbs at the wavelength that the fluorophores emits radiation, the sample may absorb some or all of the photons emitted. Photodecomposition can also occur and will decrease the intensity of the spectra. Scattered light (from Rayleigh or Raman scattering) effects must also be considered when fluorescence sp ectroscopy is used.
77 CHAPTER 3 EXPERIMENTAL METHODS AND PRELIMINARY RESULTS Using the understanding of light, light matter interactions, and spectroscopic methods, this dissertation work seeks to develop and apply new spectroscopic tools for analysis of bi ological systems. As discussed in Chapter 1, a targeted molecular sample may be altered by low intensity (subablative) ultra violet ( UV ) laser radiation such that the resulting optical response, as measured here by fluorescence emission or Raman scattering, is perturbed. Probing the targeted sample with conventional optical schemes before and after UV perturbation affords the opportunity to look for a differential response in the optical signal. The key concept of this work is the permanent molecular disruption (i.e. bond breakage) of the macro molecular target and how that alters the optical probe response, as observed with difference spectroscopy. The differential laser induced perturbation spectroscopy (DLIPS) scheme seeks to exploit the ultraviolet (UV) laser material interaction with difference spectroscopy to realize a new spectral dimension that may be used for materials classification or identification of specific structures (e.g. tissue pathologies). In this dissertation, deep UV laser perturbation of a targeted mo lecular system in combination with Raman and fluorescence spectroscopy techniques is explored in order to extract additional information tied to the UV laser material coupling. Through deep UV laser perturbation of a targeted molecular system in combinatio n with traditional spectroscopic probes, a difference spectrum is used to extract a novel spectral signature following laser perturbation The methodology using Raman and fluorescence spectroscopy has been described earlier in Smith and coworkers in 2011.
78 The DLIPS technique incorporates three complementary techniques to improve upon previous fluorescence based biosensing strategies: laser induced fluorescence emission, ultraviolet (UV) laser perturbation of tissue, and difference spectroscopy. The perturba tion pulses from the deep UV e xcimer laser (193 nm, hv= 6.4 eV) are strongly absorbed by biological tissue and used to cleave molecular bonds within the extracellular matrix (ECM). Irradiation of biological matrices at 193 nm can cause photoionization, inc luding strand breakage, locally denatured sites, interstrand cross linking, reactions via photo dimers, and other products. (Smith, 2010) In the current work, despite being well below the intensity threshold for tissue ablation, permanent alter ation of the underlying tissue structure is induced, with resulting changes within the fluorescence spectrum, specifically with respect to photoreactive biomolecules, as made apparent with the DLIPS scheme. Permanent photochemical changes are induced despi te being below the critical photon flux (~50 mJ/cm 2 for ~10 ns pulse width ) to affect material removal. Additionally, because the pre and post perturbation spectra are combined (see Equation 3.1) into a difference spectrum, the DLIPS technique removes unw anted contributions from unperturbed tissue fluorophores, broadband fluorescence, and importantly, variations in the emission bands which are unique to the sample, but not necessarily to the targeted pathology. Equation 3.1 shows the DLIPS spectral respons e, namely (3.1) where Em pre ( d Em post represent the fluorescence emission intensity recorded at each wavelength before (pre) and following (post) perturbatio n by the UV e xcimer laser, respectively. A negative DLIPS signal corresponds to a reduction in intensity following the photo perturbation step, which is generally attributed to the destruction of a corresponding fluorophore. In contrast, a positive DLIPS signal corresponds to an increase in fluorescence intensity
79 following perturbation, which may indicate destruction of a fluorescence quenching species and/or the destruction of a concomitant absorbing compound, thereby allowing more light to reach the actual fluorophore by removing competing species The DLIPS system is shown schematically in Figure 3 1. For all measurement s in this work, fluorescence excitation was accomplished using a Q switched, frequency tripled Nd:YAG laser, while the UV perturbation was a ccomplished using a 193 nm ArF e xcimer laser. The 355 nm fluorescence excitation beam was aligned coaxially with the 193 nm perturbation laser using a dichroic mirror (193 nm) as a beam combiner. Prior to the beam combiner, a variable att enuator was used to reduce the e xcimer laser to the desired perturbation intensity. The 355 nm beam diameter was about 30 % less in di ameter than the 193 nm beam (top hat beam profile) at the target plane, to ensure complete perturbation of the entire fluorescence probe volume. The co linear laser beams were passed through a pierced mirror positioned at 45 o to the target plane. With the 355 nm beam incident on the target, fluorescence emission was redirected by the pierced mirror and then focused onto a fiber optic bundle using a combination of two 50 mm diameter UV grade lenses, where it was passed to a 0.3 m Czerny Turner spectrometer a nd recorded with an intensified CCD (ICCD) array detector. Prior to entering the fiber optic, two sharp edge filters were used in series: a high pass filter designed to block the residual 355 nm light, and a low pass filter designed to block any residual 5 32 nm light remaining from the frequency tripling process. With this system, fluorescence emission was successfully collected in the spectral window between about 390 500 nm. A digital delay generator was also used to synchronize and control both the ICCD and the 355 nm laser with a 200 nm window used to collect the emission
80 For each target, the Nd:YAG laser (355 nm) was used to excite fluorescence from thin, solid films. Typically 100 to 200 s pectra were recorded and averaged for the sample. This procedu re was applied for analysis of multiple individual spots spread over multiple films. For each spot, a fluorescence spectrum was recorded, referred to as the pre pert urbation spectrum. Then an ArF e xcimer laser (193 nm) aligned coaxially to the 355 nm laser was used to deliver laser pulses to t he same spot. The 100 J/pulse e xcimer laser energy (3 mJ/cm 2 ) was sufficiently low such that no ablation of the film was realized, as verified by visible microscopy, but the to disrupt the molecular structure via direct bond breakage. Such a high photon energy equals or exceeds most bond energies, including C N (3.1 eV), C O (3.6 eV), C C (3.6 eV) and C=C (6.3 eV). F ollowing perturbation with the e xcimer laser, and a minimum t ime lag of 30 seconds, an additional fluorescence spectrum was recorded, referred to as the post perturbation spectrum. These two spectra were then subtracted to generate the difference spectrum at each location. This procedure, which permanently disrupts the molecular matrix of the targeted material, and thereby alters the optical response, has been applied to selected amino acids, collagenous materials and an animal model in the current work These experiments are described in the sections that follow. Th in films deposited on UV grade quartz flats were also examined using a confocal micro Raman spectrometer (JY Horiba LabRam) with 632.8 nm excitation. A series of pre perturbation Raman spectra were recorded, baseline subtracted, normalized to the strongest Raman band, and then averaged to produce a single, pre perturbation Raman spectrum. The same spots were then each exposed to pe rturbation pulses from the ArF e xcimer laser (100 J/pulse, 3 mJ/cm 2 ), after which the treated spots were again analyzed with mi cro Raman, and the same measurement procedure applied to produce a single, post perturbation spectrum. The post
81 perturbation spectrum was then subtracted from the pre perturbation spectrum to yield the DLIPS spectrum. This procedure has been applied to ami no acids and collagenous materials and are described in the sections that follow. A chemometric software package produced by InfoMetrix called Pirouette 4.5 was used in this dissertation for exploratory analysis and classification modeling of the amino aci ds and collagenous materials studied. Chemometrics is a data collection task, typically involving many measurements made on many samples, used to explore patterns of association in data, track properties of materials on continuous basis, and prepare and us e multivariate classification models. Using the chemometric software, hidden patterns in complex data can be revealed by reducing the information to a more comprehensible form. Possible outliers can also be exposed and whether there are patterns or trends in the data can be indicated. Exploratory analysis of multivariate data sets can reveal computationally and graphically, associated patter ns within the data. R educing large and complex data set s to a suite of best views can also expose possible outliers. I n Pirouette, samples are stored as rows in a matrix and the number of columns is equal to the number of variables. Transforms operate only on the independent variables while preprocessing operates on the independent and dependent variables. Before running a multivariate algorithm, it is often necessary to adjust a data set. Three transforms were performed upon the amino acids and collagenous materials for both the absolute Raman and DLIPS Raman data sets. The transforms are a baseline correction, sm oothing correction and division correction. The baseline correction transform corrects offset by subtracting a profile from each spectrum, in all cases a unique derived curve fit profile was used. The smoothing transform in Pirouette is based on a Savitzky Golay p olynomial filter. A convolution is applied to the independent variables in a window containing a center data point
82 and n points on either side. A weighted second order polynomial is then fit to the 2n+1 points. The center point is then replaced by the fitt ed value. In all cases a 13 point smoothing transform was applied. The division (i.e. divide by ) transform used on all data sets is the divide by Sample 1 Norm transform which is also known as area normalization. The normalization factor in this transform is the area under the sample profile which is calculated by summing the absolute values of all included variables for each sample. Since several multivariate algorithms compute results that are driven by variance patterns in the independent variables, prep rocessing is also necessary. Preprocessing was also performed upon the data sets to mitigate the influence of variable ranges and magnitudes. To all data sets, mean centering is the preprocessing technique applied. In mean centering, the origin of the data is shifted without the relative relationships between samples being altered. In Pirouette, a mean is computed for each variable and then subtracted from each data value to produce a mean centered matrix. Principal Component Analysis (PCA) was used as an e xploratory technique implemented by Pirouette. PCA finds linear combinations of the original independent variables that account for the maximum amount of variation. PCA reduces the dimensionality of the data sets by retaining only the principal components Thus, capturing relevant variations and disregarding certain combinations of irrelevant variations. PCA provides a view of variability which reveals if there is natural clustering in the data and if there are outlier samples. Before successful PCA models can be built, outliers must be excluded. The detection of outliers helps to optimize the number of factors since the original variables with the most loading has the largest impact upon the factor. Samples falling outside the critical Mahalanobis distance and critical Sample Residual thresholds are considered potential outliers. Samples exceeding
83 only one threshold slightly are considered normal but samples lying either significantly beyond one threshold or beyond both are considered outliers. All samples that did not meet the critical Mahalanobis distance and/or critical Sample Residual thresholds were examined closely before discarding and rerunning PCA excluding those samples. The sample residual threshold is based on a 95 % probabilit y limit. The criti cal sample residual is defined as: (3.2) where the model residual variance are shown in equation 3.3 and equation 3.4, respectively. (3.3) (3.4) where is a vect or, the ith row of the residual matrix E, defined as the difference between the : (3.5) The Mahalanobis distance is the distance from the multivariate mean computed from the k factor score ve ctor: (3.6)
84 where, S is the scores covariance matrix and is the mean score vector. Assuming that the Mahalanobis distance is normally distributed, using an F distribution, a critical Mahalanobis distance is defined as: (3.7) The classific ation model used here is based on similarity. Similarity techniques in classification modeling assume that the closer samples lie in spa ce, the more likely they belong to the same category. The classification algorithm used here is called K nearest neighbo r (KNN). The multivariate distance used in KNN is known as the Euclidean distance The general expression for this Euclidean distance is: (3.8) where a and b are the data vectors of two samples. The KNN algorithm also requires that a class variable be activated and a maximum k value selected. KNN constructs a model using samples pre assigned to a category which is derived from preexisting knowledge of the data. Using the classification model in which categories are defined by the training se t, predictions of the unknown data set are performed and each unknown is assigned to one and only one category. For each data set, half of the samples were randomly placed in a training set and half were designated as the validation set. The categories ass igned to the training set were defined based on clusters found during the exploratory analysis. During model building, the training set is validated by the agreement between the measured class (m class) and the predicted class (p class) produced from the K NN algorithm. Agreement between the two suggests that the m class
85 are reasonable. When the two disagree, the cause may be an outlier or an inappropriate k value. The value of k is the largest number of neighbors to be polled during classification and was s et as less than twice the size of the smallest category. Using this model to predict classes in the validation set and validate the pre assigned classes of the training set were performed. By calculating the Euclidean distance (Equation 3.8) separating eac h pair of samples, an individual sample can be assigned to a class which is the class to which most of the nearest neighbors belong. The number of neighbors polled is defined by the k value. When the training and validation subsets are both representative of the investigated population, a quality and reliable model can be built. 3.1 Amino Acids Study 3.1.1 Introduction to Amino Acids Using Raman and f luorescence spectroscopies, DLIPS was used here to evaluate glycine (C 2 H 5 NO 2 ), L proline (C 5 H 9 NO 2 ), L alanin e (C 3 H 7 NO 2 ), glycine glycine (C 4 H 8 N 2 O 3 ), glycine L proline (C 7 H 12 N 2 O 3 ), and L alanine glycine (C 5 H 10 N 2 O 3 ). Since amino acids are the building blocks of all species, their specific spectral signatures can be used to analyze other complex biological assemb lies. Proteins play a key role in nearly all biological processes by facilitating a wide range of functions, such as transport and storage of vital substances, coordinated motion, mechanical support a nd protection against diseases. In the human body, it is estimated that 100,000 different kinds of proteins exist. Each protein has a specific physiological function that is specified by the proteins chemical composition and structure. Proteins in every species are assembled from the same twenty amino acids. Th e type, number, and sequence of amino acids in a given protein will determine the proteins structure. An amino acid is a compound that contains at least one carboxyl group ( COOH) and one amino group ( NH 2 ). Proteins are formed during a polymerization
86 proc ess that links amino acids. The amino acids link together by a peptide bond formed by condensation of the amino group of one amino acid with the carboxylic group of another. This process continues on both ends (involving the NH 2 and COOH groups) to produce the polymeric chemical structures of a protein. Proteins are ubiquitous throughout the cell. pleated can be helix structure an overall rod like shape is observed th rough the stabilization by intramolecular hydrogen bonds between the NH and CO groups if the proteins main chain. All CO groups in the main chain of amino acids are hydrogen bonded to NH groups. pleated structure an over sheet shape is observed si nce almost all of the chain is fully extended while each chain forms many intermolecular hydrogen bonds with adjacent chains. pleated structures are known and are called parallel and antiparallel. There are however, many proteins w pleated structures. Most proteins have a certain amount of flexibility even though their forces create a stable structure. Amino acids in humans are classified as essential or nonessential. Essential am ino acids must be consumed in our diet, while nonessential amino acids are made by the body. Collagen is the most abundant protein in humans, consisting of mainly aliphatic amino acids; where roughly every third amino acid in its polypeptide chain is glyci ne. Alanine and proline are also nonessential amino acids that are found in abundance in the body, more so than other amino acids. The structures of glycine, alanine and proline can be seen in Figure 3 2 Also, the structures of the dipeptides glycine glyc ine, glycine L proline, and L alanine glycine can be seen in Figure 3 3.
87 Alpha amino acids are the important amino acids found in living organisms. Peptides and amino acids, the COOH and NH 2 groups are attached to the same carbon atom which is called an alpha carbon atom. The general structure amino acids can be seen in Figure 3 4 The R groups in amino acids will determine the structure/function relationship of peptides and proteins as well as allowing enzyme reactions to occur. Optically active amino acids that are found in living organisms are the L isomeric forms (with the exception of phenylalanine that appears in both L and DL isomeric forms). Amino acids found in proteins can also perfor m additional functions. For example, glycine helps to trigger the release of oxygen during cell making processes as well as being important in the manufacturing of hormones that are responsi ble for a strong immune system. Protein structure can be divided i nto four levels: primary, secondary, tertiary, and quaternary structure. The primary structure refers to the unique amino acid sequence of the polypeptide chain. The secondary structure refers to the parts of the polypeptide chain that are stabilized by in tramolecular forces. The tertiary structure involves the three dimensional structure stabilized by intermolecular forces that may occur through interactions between amino acids that a far apart in the polypeptide chain. For protein molecules that contain m ore than one polypeptide chain, the quaternary structure must be taken into account to describe the overall arrangement of the polypeptide chains. Amino acids that are not associated with proteins or peptides are found free in the body or in combined state s preform specialized functions. In addition to forming proteins, enzymes and other body tissue, amino acids are also involved and play a crucial role in a variety of chemical reactions, energy transfer, muscle activity, and energy production cycles.
88 3.1.2 Methods In this study each amino acid (see Appendix ) was mixed with deionized water at a 1:1 ratio, stirred for 24 hours, and then deposited onto UV grade quartz flats. The films dried for 24 hours (mg amino acid: ml DI H 2 O) The DLIPS analysis of 10 i ndividual spots, spread over multiple films of each amino acid was conducted as described above. An Nd:YAG laser (355 nm) with energy set at 10 d to excited fluorescence. The e xcimer laser (193 nm) was used to deliver a total of 500 laser pulses for amino acids glycine and glycine glycine, 700 laser pulses for amino acids L alanine and L alanine glycine, and 600 laser pulses for amino ac ids L proline and glyc ine L proline. Analysis of 4 individual spots for glycine and glycine glycine duced at this wavelength alone in the absence of any 193 nm perturb ation Analysis of 20 individual spots spread over multiple films of each amino acid was con ducted as described above. The e xcimer laser (100 J/pulse, 3mJ/cm2) was used to deliver a total of 800 laser pulses for L alanine, L alanine glycine, L proline, an d glycine L proline, 500 laser pulses for glycine, and 700 laser pulses for glycine glycine. Analysis of 4 individual spots for glycine and glycine glycine using only t he micro Raman spectrometer with an excitation wavelength of 632.8 nm was also conducted to determine if any effect was induced at this wavelength alone. 3.1.3 Results The traditional preperturbation fluorescence spectrum of each sp ecimen is located in Figure 3 5 Throughout section 3.1, all amino acids are color coded and there corr espondent s can be seen in Figure 3 6 To show a better comparison between alanine, proline, and glycine and these amino acids when they are attached to glycine (i.e. L alanine glycine, glycine glycine, and glycine L proline), Figure 3 7, Figure 3 8, and Figure 3 9 were plotted. Figure 3 7 shows the
89 traditional fluorescent preperturbation spectra of L alanine and L alanine glycine. Figure 3 8 shows the traditional fluorescent preperturbation spectra of glycine and glycine glycine. Figure 3 9 shows the traditional flu orescence preperturbation spectra of L proline and glycine L proline. The spectra are not corrected for neutral spectral response. In the traditional fluorescence spectra of the studied amino acids, as graphed above, it can be seen that no significant diff erence was present. The DLIPS fluorescence spectrum of each specimen is located in Figure 3 10 To show a better comparison between alanine, proline, and glycine and these amino acids when they are attached to glycine, Figure 3 11 Figure 3 12 and Figure 3 13 were plotted. Figure 3 11 shows the DLIPS spectra of L alanine (blue) and L alanine glycine (red) aft er 700 laser pulses. Figure 3 12 shows the DLIPS spectra of glycine (light blue) and glycine glycine (orange) after 500 laser pulses. Figure 3 13 shows the DLIPS spectra of L proline (green) and glycine L proline (violet) after 600 laser pulses. Also, to better compare alanine, proline and glycine, their DLIPS sp ectra are plotted in Figure 3 14. Figure 3 15 shows the DLIPS spectra of glycine glycine, L a lanine glycine and glycine L proline. Interesting results for the studied amino acids occurs before and after 430 nm. For each amino acid as a d istinctive spectral shape has e merged and this can be seen in Figure 3 10 When comparing the DLIPS spectra of the dipeptides with their peptide counterparts, it can be seen that the slopes after 430 nm, of L alanine glycine, glycine glycine, and glycine L proline are less than L alanine, glycine, and proline. Before 430 nm, another effect is seen in Figure 3 14 Th is graph shows the DLIPS spectra of glycine glycine, L alanine glycine and glycine L proline The slope of glycine L proline decreases from 380 nm 420 nm, while the slopes of glycine glycine and L alanine glycine increases. With regard to the or zero cros sing point, an important point needs to be noted. In the perturbation spectra, a negative value is considered a
90 loss in fluorescence while a positive value represents an increase in fluorescence at that particular wavelength as discussed above in the con text of chromophores, quenchers and absorbers. The traditional preperturbation Raman spectrum of each specimen is located in Figure 3 16 Also, each spectrum other than proline has been shifted vertically To show a better comparison between alanine, proli ne, and glycine and these amino acids when the y are attached to glycine, Figure 3 17, Figure 3 18, and Figure 3 19 were plotted. Figure 3 17 shows the Raman preperturbation spectra of L alan ine and L alanine glycine. Figure 3 18 shows the Raman preperturba tion spectra of gl ycine and glycine glycine. Figure 3 19 shows the Raman preperturbation spectra of L proline and glycine L proline. The traditional preperturbation, Raman spectra of each specimen are located in Figures 3 21 thru Figure 3 26 For each spec imen, located on its graph, are labeled DLIPS peaks and their associated wave number, and band assi gnment can be found in Tables 3 1 thru Table 3 6. Also, shown in Figure 3 26 are the DLIPS of all materials studied. In the absolute Raman preperturbation sp ectra of L alanine and L alanine glycine, two common bands were found in both. Those bands are a CC stretch and a CN bend that occurs at 870 cm 1 and a CN stretch and CH 3 perpendicul ar rock that occurs at 1024 cm 1 In the absolute Raman preperturbation s pectra of glycine and glycine glycine only one band was common in both. That band is a CH bend that occurs at 1453 cm 1 In the absolute Raman preperturbation spectra of L proline and glycine L proline, a few bands were found to represent these two amin o a cids. Those are: an NH 3 rock and a CH bend that occurs at 1185 cm 1 a CC stretch and a CH 2 rock that occurs at 989 cm 1 and a CH bend and a NH rock that occurs at 1270 cm 1
91 In the absolute Raman preperturbation spectra of glycine and L proline, a common band at 930 cm 1 occurs and is assigned a CC stretch and a CO bend. Also, in comparing the absolute Raman preperturbation spectra of glycine and glycine L proline, two bands are were found to be present in both. One band occurs at 1054 cm 1 and is assigne d to a CN stretch. The second band occurs at 1413 cm 1 and is assigned to a CO stretch. In comparing the absolute Raman preperturbation spectra of glycine, alanine, and glycine glycine, a CH bend occurred in all three at 1159 cm 1 A CC stretch that occur red at 912 cm 1 was apparent in the absolute Raman preperturbation spectra of glycine glycine and L proline. In the absolute Raman preperturbation spectra of glycine glycine and L alanine glycine, a common band occurred in both at 1260 cm 1 with the band a ssigned to a CC stretch and a CO bend. The absolute Raman preperturbation spectra of glycine glycine, L alanine glycine, and glycine L proline showed to all shared a common bond that occurs at 1665 cm 1 and is assigned to a C=O stretch and a NH 2 bend. In a ddition, the glycine L proline and L alanine glycine absolute Raman preperturbation spectra also shared a common band that occurred at 924 cm 1 which is assigned to a CC stretch and a CN bend. Partial component analysis (PCA) was performed for all the tra ditional Raman and DLIPs spectra of the amino acids (peptides and dipeptides), as well as for only the peptides and only the dipeptides. The wavenumber, factors, and band assignment for all the materials studied, only the peptides, only the dipeptides using traditional Raman spectroscopy are located in Table 3 7, Table 3 8, and Table 3 9 respectively. The wavenumber, factors, and band assignment for all the materials studied, only the peptides, only the dipeptides using the DLIPS method are located in Table 3 10, Table 3 11, and Table 3 12 respectively.
92 When comparing the traditional Raman technique PCA between the dipeptides and peptides greatest factors, the only wavenumber in common is at 867 cm 1 (which also appears in the PCA of all the materials). In c omparing the traditional Raman technique PCA between all the materials studied and only the peptides, the common bands include the following: 855 cm 1 867 cm 1 897 cm 1 and 1328 cm 1 In comparing the traditional Raman technique PCA between all the mate rials studied and only the dipeptides, the common bands include the following: 867 cm 1 937 cm 1 971 cm 1 1485 cm 1 and 1666 cm 1 Also, 1655 cm 1 was only found in the traditional Raman technique PCA of all studied materials. In the traditional Raman technique PCA of only peptides, the unique wavenumbers are 548 cm 1 912 cm 1 1323 cm 1 and 1455 cm 1 In the traditional Raman technique PCA of only dipeptides, the unique wavenumbers are 1278 cm 1 and 1447 cm 1 When comparing the DLIPS method PCA betw een the dipeptides and peptides greatest factors, the y have no wavenumber in common. In comparing the DLIPS method PCA between all the materials studied and only the peptides, the common bands include the following: 538 cm 1 836 cm 1 1323 cm 1 1340 cm 1 1422 cm 1 In comparing the DLIPS method PCA between all the materials studied and only the dipeptides, the co mmon bands include only 963 cm 1 Also no unique wavenumbers were found in the DLIPS method PCA of all studied materials. In the DLIPS method P CA of only peptides, the unique wavenumbers are 1001 cm 1 and 1367 cm 1 In the DLIPS PCA of only dipeptides, the unique wavenumbers are 554 cm 1 662 cm 1 867 cm 1 893 cm 1 926 cm 1 975 cm 1 1070 cm 1 1246 cm 1 and 1666 cm 1 Comparison between the DLIPS method and the traditional Raman technique PCA of all materials studied shows no common factors. For the only peptide cases, similar factors are
93 located at 538 cm 1 and 1323 cm 1 For the only dipeptide cases, similar factors are located at 867 cm 1 and 1666 cm 1 Using the information gathered from the PCA, the K nearest neighbor (KNN) was performed for a training set and validation set for (i) all the materials studied (N=6) (ii) only the peptides (N=3) and (iii) only the dipeptides (N=3) using a tr aditional Raman technique and the DLIPS method presented within this dissertation. In Table 3 13 thru Table 3 18 the training sets and validation sets for the three cases mentioned above using a traditional Raman technique can be found. In Table 3 19 thru Table 3 24 the training sets and validation sets for the three cases mentioned above using the DLIPS technique can be found. The training set of all the materials studied using the traditional Raman technique, shown in Table 3 13 has no misclassifications. The validation set in Table 3 14 of all the materials studied sho ws a misclassification of 5 % of the Proline samples, 10 0 % of the Ala Gly samples, 15 % of the Gly Gly samples, and 45 % of the Gly Pro samples. The training set of only the peptides studie d and only the dipeptides studied using the traditional R aman technique, revealed no misclassifications. The validation set of the peptides studied using the traditional Raman technique shows a misclassification of 5 % of the Proline samples. In the valida tion set of the dipeptides using the traditional Raman technique shows a misclassification of 5 % for Gly Pro. In the training sets and validation sets of all materials studied, only the peptides studied, and only the dipeptides studied using the DLIPS met hod has no misclassifications. 3.1.4 Conclusions Interesting results occur before and after 430 nm for the studied amino acids in comparing their DLIPS spectra. For each amino acid as a d istinctive spectral shape has e merged. Before 430 nm, another effect occurs in the DLIPS spectra of glycine glycine, L alanine glycine and glycine L proline The slope of glycine L proline decreases from 380 nm 420 nm, while the
94 slopes of glycine glycine and L alanine glycine increases. Also, in comparing all absolute Rama n preperturbation spectra, glycine glycine and L alanine glycine had a common band that occurred at 1260 cm 1 with the band assigned to a CC stretch and a CO bend. G lycine glycine and L alanine glycine, also showed to have a significant DLIPS value associa ted with an Amide III. Therefore, the increasing slope of glycine glycine and L alanine glycine DLIPS spectra from 380 nm thru 420 nm may be attributed to the CC stretch and CO bend that occurs at 1260 cm 1 or the Amide III which had a significant DLIPS va lue associated with it, both of which did not occur in glycine L proline. The decreasing slope of glycine L proline from 380 nm thru 420 nm could also be attributed to a CH 2 bend, a CH 2 wag and/or a NH 3 symmetric bend, all of which were present in the abso lute Raman preperturbation spectrum and have a significant DLIPS value in glycine L proline alone. When comparing the DLIPS spectra of the dipeptides with their peptide counterparts, it can be seen that the slopes after 430 nm, of L alanine glycine, glycin e glycine, and glycine L proline are less than L alanine, glycine, and L proline. Again, in comparing these two groups, the absolute Raman preperturbation spectra of glycine glycine, L alanine glycine, and glycine L proline showed a common bond that occurs at 1665 cm 1 and is assigned to a C=O stretch and a NH 2 bend that was not present in L alanine, L proline, or glycine. In addition, the glycine L proline and L alanine glycine absolute Raman preperturbation spectra shared a common band at 924 cm 1 which is assigned to a CC stretch and a CN bend. Therefore, the reason for the dipeptides having smaller positive slopes after 430 nm than the peptides may be because of the common bond that occurs at 1665 cm 1 and is assigned to a C=O stretch and a NH 2 in the d ipeptides. A key finding is the discrimation power of DLIPS as compared to traditional Raman alone. Table 3 14 shows significant misclassification of the dipeptides with traditional Raman,
95 while Table 3 20 shows perfect classification of the dipeptides. In Table 3 7 the predominant factors involve hydrocarbons while Table 3 10 shows the largest factor at 538 cm 1 and is assigned to an Amide IV. This suggests that the peptide bond was perturbed. It should be not ed, however, that a more complete spectral anal ysis needs to be obtained before the slopes of the DLIPS spectra can be directly associated with a photochemical response that are occurring in the studied amino acids when perturbed by a 193 nm laser. Nevertheless, it is clear that the DLIPS technique has been successive in differentiating between L alanine, L alanine glycine, glycine, glycine glycine, L proline, and glycine L proline. Overall, it is concluded that the DLIPS does provide a new spectral dimension when examining the fundamental building bloc ks of biological tissue, showing promise of the DLIPS scheme for analysis of more complex specimens. 3.2 Collagenous and Related Biological Materials 3.2.1 Introduction to Collagenous materials Collagen and elastin are the two major structu ral protein mol ecules that form connective tissues used to protect and support organs. Collagen type IV is found primarily in the deepest layer of tissue, the basement membrane, which secures the overlying layers of stratified epithelium. This type of collagen contains a bout 1400 amino acids and is longer than the fibril helical structure. Collagen type IV antibodies can be used in the detection of basement membrane loss in carcinomas and the vascular na ture of neoplasms as well as in the classification of soft tissue tumors. One of type IV gene coding mutations lead to Alport syndrome. Fibrinogen is one of the three main components of plasma (globulins and albumin are the other two) however, only 3 % of plasma is
96 made up of fibrinogen. During blood coagulation, fibrinogen is synthesized by the liver and converted by thrombin into a protein called fibrin. Through processes in the coagulation cascade that activate zymogen pro thrombin to the serine protease thrombin, which converts fibrinogen into fibrin, fibrin is cross linked by factor XII and a clot is formed. The fibrin is stabilized further by fibrinolysis inh ibitors and binding to several adhesive proteins of various cells. Fibrin can function as temporary inhibitor of these enzymes by binding specifically to activate coagulation factors and entrapping them. Injury to collagen and elastin by inflammation is co mmon in patients with connective tissue diseases. s. Therefore, many connective tissue diseases are inflammatory diseases. Abnormal reactions of the immune system can result in inflammatory and autoimmune diseases. Abnormal immune system activity arises in autoimmune diseases through the immune systems in appropriate immune response of the body against species normally present in the body. For a disease to be regarded as an autoimmune disease direct evidence from transfer of pathogenic antibody or pathogenic T cells, indirect evidence based on reproduction of the autoimmune disease in experimental animals and circumstantial evidence from clinical clues must be presented. Autoimmu ne diseases can be classified based upon the type of reaction produced by the normal immune system. Another prominent feature of ma ny connective tissue diseases is vascular damage and the circulation autoantibodies. Circulating autoantibodies have been detected in the serum of patients suffering from systemic lupus erythematosus, (SLE), rheumatoid arthritis (RA), systemic vasculitis ( SV), Kawaskai syndrome (KS), and many other connective tissue diseases. One
97 widely studied autoimmune disease is SLE. SLE is a chronic autoimmune disorder that can affect every organ system. A lupus band test one of the tests performed to determines if a p atients has SLE. A lupus band test detects I mmunoglobulin G (IgG) antibodies located just below the outer layer of the tissue sample, the dermoepidermal junction. A positive result would indicate that IgG was present. An antigen is a substance that is reco gnized as foreign and provokes an immune tissue structure the greater the immune response against those antigens. The most commonly encountered antigen presenting cells are macrophages and dendritic cells. These cells engulf antigens and degrade them into small fragments that eventually, bind to major histocompatibility complex (MHC) molecules (cell surface proteins). The antigen MHC complex will stimulate lymphocyt es to attack that particular antigen. Lymphocytes attack specific antigens through two different mechanisms. One group called B lymphocytes produce proteins called antibodies. Antibodies circulate in the bloodstream and penetrate into extracellular fluids binding to the antigen that induced the immune response. A second group is called cytotoxic T lymphocytes (CTLs) bind directly to the cells that are exhibiting antigens on their surface. The targeted antigens are killed by the CTLs causing them to burst. Natural killer (NK) cells are also a type of lymphocytes. This type attacks a broad spectrum of abnormal cells but only makes up a small amount of the total lymphocyte population. In many types of connective tissue diseases, patients will have unexplained immunoglobulin abnormalities. For example, high levels of IgA can occur in patients with rheumatoid arthritis, SLE, and some liver diseases, high levels of IgG can mean a chronic infection is present, such as AIDS, high levels of I mmunoglobulin M (IgM) ca n mean
98 mononucleosis, rheumatoid arthritis, or kidney damage and high levels of fibrinogen are associated with cardiovascular disease, but can also occur in the presence of inflammation. Immunoglobulins are globular proteins, one of the two main protein cl asses, are mostly soluble in aqueous solutions and are globe lik e in shape. Immunoglobulins (Ig s) are secreted and expressed by activated B lymphocytes. The different classes of Igs in humans are IgA, IgD, IgE, IgG, and IgM. Igs can be found in the serum, other body fluids, and tissue fluids. All of these immunoglobulins share a similar structure, but differ in size, charge, amino acid sequence and glycan content. Igs also contain a Fab region and a Fc region. The Fab region is concerned with antigen bindin g and the Fc region is concerned with the effector functions. Igs are products of the immune system and bind to antigens through specific sequences on their Fab regions. Once an Ig is bound to an antigen, a direct effect can occur in which the antibody and antigen are clumped and the antigen is neutralized or prevented from invading cells or an indirect effect can occur in which effector functions are released via the Fc regions. Fc effector functions are exclusive to each heavy chain class present in the I g. Effector functions include activation of complement and cell binding of various types. In IgA, IgD, and IgG the Fab and Fc region are separated by a flexible joint region. In IgE and IgM, instead of a flexible joint, the Fab and Fc region are separated by an additional CH domain. This results in a restriction of the overall motion of IgE and IgM. All Ig molecules consists of two polypeptide chains, heavy chains (H chain) and light chains (L chain), that are linked together by disulphide (S S) bonds and n on covalent forces. The chains are composed of Ig domain regions which consist of ~105 130 amino acids arranged in a sheet structure. A variety of amino acid sequences, glycan content, and location and number of
99 glycosylation sites occur in immunoglobuli ns. Each H Immunoglobulin classes are distinguished by these heavy chain differences. L chains exist in chains which are called Type chains which are called Type L but not both. The L chains are regulated differently dependent upon the antigen stimulus. Thus, immunological disturbances that result in abundant insight and correlation with diseases and disease progression. The three most common immunoglobulins found in serum are IgA, IgG, and IgM. Immunoglobulin A (IgA) is an antibody that plays a critical role in mucosal immunity in its secretory form. IgA antibodies protect the body surfaces that are exposed to outside substances. IgA is also the most abundant immunoglobulin in the body overall. Therefore, IgA is found in saliva, tears, synovial fluid, intestinal fluid, and respiratory tract secretions and uses the mucosal surfaces as a first line of defense in infections that enter through mucosal surfaces. IgA can also be found in blood, as made by plasma cells, but only in small amounts. Serum IgA is a monomer but IgA found in secretions is a dimer. chain of each IgA is disulfide bridged on the hinge region. Non covalently bonded to the Fc region is a glycosylated polypeptide chain called a secretory component (SC). Secreted and circulating IgA antibodies ar e known to have antiviral properties and can bind to some cells. Immunoglobulin G (IgG) is an antibody found in all body fluids. They are the smallest antibody but occur in the largest abundance than any other antibody (75 % 80 %) in serum. IgG is the majo r antibody involved in the secondary immune response and serves to activate the complement system. IgG antibodies are very important in fighting bacterial and viral infections. IgG is also the only immunoglobulin that can cross the placenta in a pregnant w omen and is the
100 major component of passive maternal antibody transfer via colostrum and yolk. Interestingly, patients diagnosed with rheumatoid arthritis (RA) have an increased level of IgG (approximately 3 times normal IgG serum levels) except during preg nancy when the disease goes into remission and IgG serum levels return to normal. IgG molecules are monomers, categorized into subclasses whose structures differ in the number of disulfide bonds and length of the hinge region. Binding of IgG can enhance ph agocytosis and bind specifically to Fc receptors on cells can result in an increased activation of antibodies released. Immunoglobulin M (IgM) is an antibody found in blood and lymph fluids. IgM is the first immunoglobulin made by the fetus and is the thir d most common immunoglobulin in serum. IgM is also the first immunoglobulin made by virgin B cells, once stimulated by antigens. IgM appears early in the course of an infection and usually reappears in a lesser quantity if exposure continues. IgM is produc ed by B cells and is the primary antibody against A and B antigens on red blood cells. As with IgG, once IgM is bound to an antigen, the complement system is activated. IgM antibodies are the largest found in the circulatory system and are mainly responsib le for clumping together microorganisms for eventual eliminating from the body. IgM antibodies are mainly responsible for red blood cells clumping in blood transfusion of ates a recent infection and the clarification of the specific IgM antibodies can be used to diagnosis uterine infection. IgM exists as a pentamer in serum normally, but can also exist as a monomer. IgM as a pentamer has all identical heavy chains and all identical light chains. In a pentameric form, IgM has a hinge region that is formed from a protein covalently bound via an S S bond. IgM in a monomer form functions as a receptor for antigen on B cells and lacks the hinge region and but
101 has an additional linkage of 20 amino acids. Monomeric IgM is non covalently bonded to two Ig. These additional proteins allow contact to occur between surface immunoglobulin and an antigen, which is required to start antibody secretion. 3.2.2 Methods Abnormalities in immunoglobulins, collagen and fibrinogen content are strong indicators of autoimmune diseases. Autoimmune disease s involve a large category of diseases in which may of the symptoms overlap. Associating these abnormalities with a specific disease or the progression of a specific disease is of great clinical value. A correlation could result in earlier prognoses and ea rlier treatment (if applicable), therefore, decrease d morbidity rates and increase d quality of life. In collagen diseases that take a chronic course, increased IgA and IgG and a normal or diminished IgM concentration occur frequently. However, no distinct association has been found between immunoglobulin disbalance and a specific disease. Collagenous materi al concentrations are not only a ffected by specific pathological processes, they can also differ widely between patients with the same disease. The main goal of this study was to differentiate between the three most common immunoglobulins (IgA, IgG, and IgM), collagen and fibrinogen through differential laser induced perturbation spectroscopy (DLIPS), as described above, with the ultimate goal of producing an optical tool that will provide clinicians additional insight into autoimmune diseases. Each collagenous material was examined using fluorescence and Raman spectroscopy. By comparing the fluorescence intensity before and after perturbation, a difference spectroscopy was found for each material. In this study, each col lagenous material (See Apandix 1 ) were mixed with deionized water, stirred for 24 hours, and then deposited in thin films onto quartz flats. The films dried for
102 24 hours. For each film, 20 i ndividual spots were examined by the DLIPS technique described above for a fluorescence probe The e xcimer delivered 2000 laser pulses to the each of the spots on the film. The difference between the two spectra was computed for each spot and then averaged among the ten spots to obtain a spectral signature for each collagenous material. To determine if the 355 nm laser was altering the samples, measu rements were taken without the e xcimer laser perturbation Raman spectra were also obtained for each film, yi elding a unique spectral signature for each material. A confocal micro Raman spectrometer, as described above, was used. In this constrained from 300 to 2000 cm 1 The e xcimer delivered 1500 laser perturb ation pulses to each of the spots on the film. The data manipulation for the ten spots incorporated pixel correction, baseline subtraction, and normalization to the strongest Raman band; the spots were then averaged to obtain a spectral signature for each antibody. 3.2.3 Results The absolute preperturbation fluorescence spectrum of each spec imen is presented in Figure 3 27. Figure 3.29 shows a closer view of the collagenous materials that are difficult to see in Figure 3 27 because of the difference in inte nsities of the each absolute preperturbation spectrum. Throughout section 3.2, all collagenous materials are color coded and there corre spon dents can be seen in Figure 3 28 No general distinguishable feature was observed in the absolute fluorescence (this denotes simply the pre perturbation fluorescence (i.e. unperturbed) and not the fluorescence spectrum corrected for instrument response) spectra from the collageno us material study. In Figure 3 2 7 interestingly, the absolute preperturbation fluorescence spectra of fibrinogen and IgG have high intensities in comparison to collagen, IgA, and IgM. The DLIPS fluorescence
103 spectrum of each spe cimen is located in Figure 3 30 Figure 3 31 shows a closer view of the collagenous materials that are d ifficult to see in Figure 3 30 because of the difference in intensities of the each DLIPS spectrum. After 2000 laser pulses, the DLIPS spectra of the studied collagenous materials, as shown i n F igure 3 30 seem to follow a similar pattern occurring at different intensities except for IgM which occurs at a much higher intensity with a completely different spectral shape. In Figure 3 30 which shows all specimen except for IgM, a pattern becomes apparent between collagen, fibrinogen, IgA, and IgG. To obtain a better understa nding of the collagenous materials structures as well as to enrich their related DLIPS spectra Raman spectroscopy was used. The absolute preperturbation Raman spectrum of each sp ecimen is located in Figure 3 3 2 Again, all collagenous materials are color coded and there correspondents can be seen in Figure 3 28 Also, each spectrum other than IgM has been shifted to show distinction between Raman peaks of each collagenous material. Each collagenous material is also plotted separately and for each specimen, located on its graph, is labeled peaks. The associated peak label, wave number, and band assignment can be found in the table following each Ram an preperturbation spectrum. Figur e 3 33 and Table 3 25 are associated with collagen. Figure 3 34 and Table 3 2 6 are associated with fibrinogen. Figure 3 35, Figure 3 36, and Figure 3 37 show the absolute preperturbation spectra of IgA, IgG, and IgM, respectively. Also, Table 3 27, Table 3 28, and Table 3 29 are associated with IgA, IgG, and IgM, respectively. The DLIPS of all materials studied are located in Figure 3 38. Partial component analysis (PCA) was performed for all the traditional Raman and DLIPs spectra of the collagenous materials studied, and only the immunoglobulins. The wavenumber, factors, and band assignment for all the materials studied, and only the immunoglobulins using
104 traditional Raman spectroscopy are located in Table 3 30 and Table 3 31 The wavenumber, factors, and band assignment for all the materials studied, and only the Igs using the DLI PS method are located in Table 3 32 and Table 3 33 In comparing the traditional Raman technique PCA between all the materials studied and only the Igs common bands include the following: 360 cm 1 411 cm 1 489 cm 1 735 cm 1 1008 cm 1 1130 cm 1 1302 cm 1 and 1447 cm 1 Also, no unique band was found in the traditional Raman technique PCA of all studied materials. In the traditional Raman technique PCA of only Igs, the unique wavenumbers are 1240 cm 1 In comparing the DLIPS method PCA between all the materials studied and only the Igs, the common bands include the following: 360 cm 1 392 cm 1 444 cm 1 489 cm 1 806 cm 1 1008 cm 1 1099 cm 1 1130 cm 1 1240 cm 1 1302 cm 1 and 1447 cm 1 Also, in the DLIPS method PCA of all materials studied, the unique wavenumber is 735 cm 1 In the DLIPS PCA of only Igs, the unique wavenumber is 1062 cm 1 Comparison between the DLIPS method and the traditional Raman technique PCA of all materials studied shows similar wavenumbers at 360 cm 1 489 cm 1 735 cm 1 1008 cm 1 1130 cm 1 1302 cm 1 and 1447 cm 1 Also, for only the traditional Raman technique PCA of all materials studied, unique band at 411 cm 1 and 1062 cm 1 appeared. For only the DLIPS method of all materials studied, unique bands are located a t 392 cm 1 444 cm 1 806 cm 1 1099 cm 1 and 1240 cm 1 For the only Igs cases, similar factors are located at 489 cm 1 1008 cm 1 1062 cm 1 1130 cm 1 1240 cm 1 1302 cm 1 and 1447 cm 1 For only the traditional Raman technique PCA of only the Igs stu died shows unique factors at 360 cm 1 411 cm 1 and 735 cm 1 Also, for only the DLIPS method of only the Igs studied shows unique factors at 392 cm 1 444 cm 1 806 cm 1
105 and 1099 cm 1 As observed with the amino acids, the CN and NCC bands play key role s in the DLIPS analysis. Using the information gathered from the PCA, the K nearest neighbor (KNN) was performed for a training set and validation set for all the materials studied, and only the Igs using a traditional Raman technique and the DLIPS method presented within this dissertation. In Table 3 34 thru Table 3 37 the training sets and validation sets for the cases mentioned above using a traditional Raman techn ique can be found. In Table 3 38 thru Table 3 4 1 the training sets and validation sets for the cases mentioned above using the DLIPS technique can be found. The training set of all the materials studied using the traditional Raman technique shown in Table 3 34 has no misclassifications. The validation set in Table 3 35 of all the materials stud ied shows a misclassification of 11.7 % of the IgA samples, and 33.3 % of the IgG samples. The training set of only the Igs using the traditional Raman technique, shown no misclassifications. The validation set of the Igs studied using the traditional Rama n technique shows a misclassification of 17.65 % of the IgA samples. In the training set of all materials studied using the DLIPS method shows a misclassification of 5 % of the IgG samples and 22.22 % of the fibrinogen samples. In the validation set, a mis classification of 10.5 % of the collagen samples, 21 % of the fibrinogen samples and 21 % of the IgG samples are shown in Table 3 39 The training set of only the Igs using the DLIPS methods shows no misclassifications. The validation set in Table 3 40 of the Igs only studied shows a misclassification of 21 % of the IgG samples. Overall, the DLIPS performed better for the Ig samples, while traditional Raman did better for the collagen and fibrinogen when polling all data.
106 3.2.4 Conclusions The high intensit ies of the absolute preperturbation fluorescence spectra of fibrinogen and IgG as well as both of these specimens being the only two affected by the 355 nm laser, may be attributed to a CC stretch that occurs at 1016 cm 1 and a CH bend and CH 2 stretching t hat occurs at 1454 cm 1 These band assignments at these wavelengths were not detected in type IV collagen, IgA, nor IgM. The DLIPS spectrum of IgM, which occurs at a much higher intensity and with a completely different spectral shape than the other studi ed collagenous material, may be attributed to the specific band wavenumbers and assignments that were only found in IgM. Those are a CC bend that occurs at 422 cm 1 a CN stretch that occurs at 524 cm 1 and a CH bend and a CN stretch that occurs at 806 cm 1 and a CC stretch that occurs at 920 cm 1 a CH bend and a CN stretch that occurs at 1035 cm 1 a CN stretch that occurs at 1156 cm 1 an Amide III that occurs at 1260 cm 1 an Amide II that occurs at 1480 cm 1 and a C=O stretch and CO 2 that occurs at 1594 cm 1 However, it should also be noted that IgM films used in this study was the only collagenous material studied which was diluted to a 1:10 ratio to deionized water (1:1 ratio to deionized water was used for all others presented in this study). Fur ther testing could be done to determine if the concentration of IgM in the thin film affects the fluorescence intensity. As mentioned above, the DLIPS spectra of the studied collagenous material seem to follow a similar pattern occurring at different inten sities, except for IgM which occurs at a much higher intensity with a completely different spectral shape. The cause of this specific pattern may be related to specific wavenumber and band assignments of molecules found only in type IV collagen, fibrinogen IgA and IgG. Those bands are a CN bend that occurs at 394 cm 1 a NH 2 rock that occurs at 735 cm 1 and a CN stretch and NH 3 bend that occurs at 1392 cm 1
107 However, the x axis intercept of the DLIPS fluorescence spectra of type IV collagen, fibrinogen, I gA and IgG, (i.e. where the intensity changed from negative to positive), was unique. These x axis intercepts could be used as a means of differentiation and are listed in Table 3.2.6. The cause of the different x axis intercepts may be attributed to a CH bend and CN stretch that occurs at 1139 cm 1 which was only seen in fibrinogen, a CC stretch that occurs at 1227 cm 1 which was only seen in IgA, or to an Amide I and C=O stretch at 1676 cm 1 which was only observed in IgG. In conclusion, DLIPS has sho wn to be a sensitive technique in which differentiation between 5 different collagenous materials namely type IV collagen, fibrinogen, IgA, IgG, and IgM, was achieved. However, it should be noted that a CH bend and a CH 2 twist that occurs at 1310 cm 1 was present in the absolute preperturbation Raman spectra of only type IV collagen, fibrinogen, IgA, and IgM and a CC stretch that occurs at 1070 cm 1 in the absolute preperturbation Raman spectrum of only fibrinogen and IgM, have yet to be associated with a ny part of their DLIPS spectra. Clearly the 193 nm excimer strongly couples to the peptide bones and associated CN bonding. Further investigations should be conducted to associate these repeating Raman peaks with a DLIPS spectral response so that a complet e picture of type IV collagen, fibrinogen, IgA, IgG, and IgM can be obtained. 3.3 Animal Model Study 3.3.1 Introduction to Animal Model Cancer or a malignant neoplasm is a disease whose properties change with time. Cancer cells divide and grow uncontrollab ly, forming malignant tumors. Cancer can be caused by the direct damage of genes or by existing genetic faults within cells. Cell alterations that commonly occur in cancer cells is a decrease in specialized cell surface structures which are composed of
108 pro tein and are known as gap junctions. In cell cell communication, gap junctions join adjacent cells together in a way that allows small molecules to pass directly from one cell to another. Once gap junctions disappear from some cells, the cells revert to un controlled growth. In cancer cells, another altered property is the ability of the cells surface functional groups to move more readily and increases the rate at which they can bind to sugar binding proteins that are highly specific to their associated sug ar functional groups called lectins. A single lectin molecule can link two cells together by binding to carbohydrate groups exposed on the surface of each cell. Therefore, the tendency of cancer cells to clump together when exposed to lectin proteins is en hanced. Cancer cells can also spread to neighboring tissues and throughout the body. When considering only the size of circulating cancer cells, which are larger than blood cells, they will most likely become lodged in the capillaries. The lodged cancer ce lls can now adhere to or penetrate through the capillary walls and begin the formation of a metastatic tumor. For most primary tumors, the first such capillary bed will be in the lungs, which may help explain why the lungs are a relatively frequent site of metastasis for many kinds of cancer. The pattern of blood flow in the circulatory system will also determine where cancers cells, floating in the bloodstream, are likely to accumulate. Cancer cells can also spread to other parts of the body through the ly mphatic system. The bodies flow patterns are important but do not always explain observed metastatic distributions. The behavior of malignant tumors depends not just on the traits of cancer cells, but also on interactions between cancers cells and normal c ells of the surrounding host tissue. In healthy humans, destruction of newly forming cancer cells through the immune system is a routine event. Cancer simply reflects the failure of an adequate immune response against
109 abnormal cells or the ability of the a bnormal cells to exhibit unique antigens that can inhibit an immune response. However, determining the specific cause of cancer is complex. Cancer exhibits a variety of antigenic changes which can be present in differing concentrations but are not unique t o a specific cancer. With a particular focus on the early detection of cancer, advanced optical techniques have been developed to probe tissue. The interactions between light and biological samples are used currently in diagnostic medicine with the use of optical microscopy to study tissue samples. Fluorescence based techniques have also been developed to probe specific tissue fluorophores, for example, collagen, nicotinamide adenine dinucleotide (NADH), fl avins, and porphyrins. Drezek and coworker in 2001 s tudied cervical tissue and showed an increase in NADH fluorescence within the epithelium with progression of abnormal tissue. Epithelial cell layers are separated from the underlying tissue by a distinct boundary structure called the basal lamina layer, w hich is a thin dense layer of protein containing material that makes it relatively easy to determine whether a cancer has begun to spread. Current methods of cancer detection include screening tests, medical imaging and certain signs and symptoms. Once can cer is detected, it can be diagnosed by examination of a tissue sample of the suspect lesions followed by histopathological analysis under visible light microscopy which remains the gold standard for disease state confirmation in a large proportion of can cers. Cancer is typically treated with surgery, radiation therapy, and chemotherapy. When a cancer has metastasized to unknown locations, a stimulated immune response against the cancer cells, wherever they may be located can be employed. Cytokines which a re proteins produced by the body to stimulate immune responses against infectious agents, can be used to treat cancer.
110 Antibodies can also be used to target cancer cell destruction either by themselves or by linking them to radioactive substances, chemothe rapeutic drugs, or other kinds of toxic substances. Antibodies by themselves, the antibodies bind to the cancer cell surface, where their presence triggers an immune attack that destroys only those cells to which the antibody is attached. Antibodies when l inked to toxic substances, allows the toxins or radioactivity to be selectively concentrated at tumor sites by the antibody without accumulating to toxic levels elsewhere in the body. The basic strategy to develop monoclonal antibodies is to immunize anima ls with human cancer tissue and then select those monoclonal antibodies that bind to antigens on the cancer cell surface. When they are injected into individuals with cancer, these antibody molecules would be expected to circulate until they encounter canc er cells. Radioactive antibodies are also useful for determining where cancer cells are localized for monitoring changes in tumor cell numbers in response to treatment. Chances of survival vary greatly by the type, location, and degree of cancer proliferat ion the patient demonstrations at the beginning of treatment. Cancer can affect people of any age and the risk of developing cancer typically increases with age, though a few types of cancer are more common in children than adults. Molecules are correlated with certain pathological conditions and have unique distributions in tumors. Classes or groups of these organic molecules and biological materials can be differentiated when excitation occurs at specific wavelengths. Because the excitation/emission wavel ength pair is a combination of the molecular structure and the overall molecular environment (e.g. biological matrix), fluorescence emission may provide discrimination among emitting materials. Common to spectra taken from complex samples, high intensity, broadband fluorescence responses from tissue often obscure the rich biomolecular information found in lower intensity
111 fluorescence emission bands. As this broadband response could potentially be from a tissue fluorophore unrelated to disease, targeting spe cific biomarkers of disease has been challenging using these techniques. However, the largest hurdle to surmount in translating these spectral strategies to the clinic has been the significant patient to patient variation in fluorescent properties. This ha s been shown to be associated with race, age, sex, air temperature, and even deformation of the tissue when applying the probe. Some patient and sample variability includes: fluctuations in absolute emission intensity, emission peak shifts, and changes in the scattering and absorption properties of the tissue, etc. Brookner and coworkers in 2003, for example, reported variations in peak fluorescence intensities varying by more than a factor of five between patients while the intra patient coefficient of var iation was less than 25 %. While fluorescence signals are directly linked to the local molecular structure of the targeted system, there clearly remains a need for further improvement in optical based sensing schemes to specifically address the variations realized with absolute fluorescence intensity. A s mentioned above, the patient to patient variability is a known confounding factor for clinical in situ fluorescence measurements. Although multiple standardization and normalization techniques have been pro posed to combat this issue, it necessarily comes at the cost of lost information from the collected spectra. In the case of DLIPS, the spectra represents a differential response measurement from the same location and is in that sense self referencing in or der to specifically remove variation associated with inter sample differences. The DMBA as described below, model of pathogenesis has been well characterized, particularly with regard to in vivo models for the evaluation of fluorescence based techniques t o detect and diagnosis cancerous and pre cancerous lesion formation.
112 3.3.2 Methods The main goal of this study was to investigate the DLIPS potential as a precancerous diagnostic technique. The study was terminated at 11 weeks, prior to the emergence of sq uamous cell carcinomas (SCCs) as the goal of the study was not to investigate the biological difference between cancerous and normal tissue but rather to investigate DLIPS potential as a precancerous diagnostic technique. Thus tissue comparisons were made between healthy tissue and tissue during the course of SCC formation due to 7,12 dimethylbenz(a)anthracene ( DMBA ) initiation and promotion as reported recently in Kozikowki, et al. 2001 Tumor formation on female athymic nude mice, six to eight weeks old, was induced and promoted by topical application of 7,12 dimethylbenz(a)anthracene (DMBA) in mineral oil at a concentration of 0.5 % w/w applied topically to the dorsal skin. Typically acetone is used as a solvent for DMBA but to avoid skin irritation due to repeated acetone administration alone, mineral oil was used. Application was repeated two to three times per week throughout the 11 week of this study. DMBA application was discontinued if mice began exhibiting signs of systemic toxicity, particularly w eight loss. The untreated mice received topical application of mineral oil alone. In this study, analysis of pathological changes in a murine model using the DLIPS technique as described above was conducted. Prior to all spectral measurements, each mouse w as cleaned with an alcohol and anesthetized using a ketamine/xylazine solution (Phoenix Pharmaceutical; Lloyd Laboratories) at 10 ml/kg of body weight. Starting in week two of this study, fluorescence spectra and DLIPS spectra were collected from two spots on each side of two untreated mice and two DMBA treated mice (4 spots x 2 mice). An Nd:YAG laser (355 nm) d to excited fluorescence. The e xcimer laser (193 nm) was used to deliver a total of 2500 laser pulses for each spot. Spectra were collected once a week for 11
113 weeks and despite emerging lesions on the DMBA treated mice by week eight, from skin regions that appeared to be pathology free under visual inspection (i.e. visible lesions were avoided). At the end of this study all remaining mice were euthanized using an intraperitoneal injection of Euthasol (Virbac AH). Tissue comparisons were also made during this study between the untreated and DMBA treated tissues. One mouse was selected from each group and was euthanized a s described above. Immediately following euthanasia, the dorsal skin was excised all the way down to the anteroposterior axis and fixed in formalin (Sigma Aldrich). Histological slices were either prepared immediately from these samples or the fixed skin w as stored in a 30 % sucrose solution for later use. Fixed samples were paraffin embedded, sliced into 5 m sections, and stained with hematoxylin and eosin (H&E). The sections were then mounted on glass and analyzed by a veterinary pathologist Dr. Castlem an of the University of Florida 3.3.3 Results At the conclusion of the study, all spectra were visually inspected and 22 individual spectra (~14% of the total collected) were rejected due to (1) excessive spectral noise and/or no observable difference in the pre and post perturbation spectra; indicative of low Nd:YAG laser and/or low perturbation laser intensity during the collection process, or (2) due to atypical spectral appearance (e.g. unusually large negative or positive perturbation) which is indic ative of mouse movement (e.g. twitching) between the pre and post perturbation measurements. Given the considerable variation in absolute fluorescence intensity as compared to the degree of actual perturbation, any spatial movement of the target during th e DLIPS procedure wil l generate a spurious response. Only spectra with unambiguous outlier characteristics were omitte d during the screening process.
114 Over the weeks of DMBA application, noticeable changes in shape of DLIPS spectra emerged (DMBA treated ver sus untreated), while these changes are not directly observable in the fluorescence spectra alone. The average DLIPS spectra at weeks 1 11 for the untreated and DMBA treated skin can be seen in Figure 3 39 and Figure 3 40 respectively. Across all acquired spectra, the most important regions in explaining the differences are: below 400 nm, a peak at 420 nm, and a region of increasing importance at wavelengths above 460 nm. Additionally, when the untreated spectra and DMBA treatment spectra are analyzed sepa rately, unique band regions emerge to potentially explain the observed pathological progression to lesion formation in the DMBA model as opposed to the skin from the untreated mice. Primarily, the DMBA treated skin has an important band between 400 and 420 nm, while band is below 400 nm. Figure 3 39 thru Figure 3 41 shows the average traditional fluorescence spectra of untreated and DMBA treated skins at week 2, week 6, and week 8. Figure 3 4 4 thru Figure 3 46 shows DLIPS spectra of the D MBA treated and untreated skins at week 2, week 6, and week 8. Spectral samples at the same time points of the study, illustrate the ability of the DLIPS technique to exploit internal normalization of the spectra relative to the tissue, and thus revealing information about the underlying pathology. Both figure sets mentioned above show the emergence of a noticeable band in the lower wavelengths of the DLIPS spectra throughout the pathological progression of the DMBA model. This information is either obscure d or not detected using the pre perturbation fluorescence alone, as the spectra from the skin (both untreated and DMBA treated) are dominated by the broad, correlated fluorescence peak centered on 460 nm. Across all acquired spectra, the most important reg ions in explaining the differences are: below 400 nm, a peak at 420 nm, and a region of increasing importance at wavelengths
115 above 460 nm. The entire data set showed significant variability above this wavelength. This implies that this region may specifica lly account for variability between control and DMBA treated skin using the DLIPS technique. In Figure 3 4 7 thru Figure 3 50 representative photomicrographs of the H&E sections of the skin at four, eight, and eleven weeks of the study are shown. Each phot omicrograph is at a magnif ication of 150x. Figure 3 45 is from an untreated mouse after week four of the study. There is normal epidermal thickness with several layers in the stratum corneum. Hair follicles and sebaceous glands are in normal density within dermal collagen and subcutis. Subsequent weeks of skin samples from control mice were comparable to the samp le at week four. Figure 3 46 shows the skin after four weeks of DMBA treatment which displayed moderate epidermal changes including acanthosis and orthokeratotic hyperkeratosus, characterized by thickening of the stratum spinosum, stratum granulosum, and stratum corneum. Hair follicles were reduced in density and the dermis in treated areas was almost devoid of sebaceous glands compared to skin from control mice, while at the same time dermal collagen density was increased. After eight weeks of topical DMBA application, skin was characterized by more severe acanthosis, orthokeratotic hyperkeratosis and dermal fibrosis than in skin after four week s, as shown in Figure 3 49 Hair follicles were infrequently found in treated areas, and sebaceous glands were almost completely absent. Squamous cell papillomas were occasionally present in the epidermis (although not shown in this section). At the conclusion of the study, week 11, the DM BA treated skin (Figure 3 50 ) showed more severe epidermal acanthosis and hyperkeratosis. There was also an increased, but still low density, scattering of squamous cell papillomas throughout the epidermis. Dermal fibrosis was slightly more severe than in samples from mice at eight weeks of DMBA treatment. This observed histopathological progression was as previously described
116 when using this model, with inflammation and hyperplasia occurring early, dysplastic lesions forming at around six weeks of treatment, and later the emergence of papillomas prior to carcinoma formation. 8), and 11) segments, based upon the visual observation of simil ar pathological states within those time periods for the DMBA treatment group was considered pre pathological as both observational and histological evidence showed only the earliest stages (inflammation and hyperplasia) of d were considered definite pathological because dysplastic lesions and papillomas were visible during those periods of the study. 3.3.4 Conclusion As has been shown before, the strong fluorescence peak centered on 460 nm for the fluorescence spectra can most reasonably be assigned to the tissue fluorophore NADH. While this is an important tissue fluorophore, and is a primary target of fluorescence tissue assessment, it serves as a marker of increased cellular metabo lism, which is a hallmark of dysplasia, but not necessarily unique to it. In conjunction with increased cellular metabolism, one of the earliest markers of pre cancerous progression is cellular infiltration and the release of growth factors and cytokines. This unchecked cellular signaling, results in the proliferation of fibroblasts, increased collagen synthesis, and suppression of collagenase production, with the overall effect of restructuring the local ECM. These effects suggest that techniques which tar get collagen remapping might provide direct diagnostic coupling of the spectra to the pathology. In the spectral region up to 420 nm, is the most prominent region in explaining the variance when using DLIPS, and is in the range for emission due to collagen This agrees w ith previous work showing that e xcimer laser pulses can break collagen amide bonds, resulting in a pronounced difference
117 spectrum band. The proximity to the 460 nm NADH peak could indicate it is simply a residual region from that emission. A lternatively, it could be related to differential melanin accumulation in DMBA treated skin versus untreated samples as this has been shown to be a marker for some skin cancers. It is also well known that melanin interacts with UV light; hence a unique dif ferential response under DLIPS interrogation is entirely possible. While this is an interesting finding, further investigation of individual tissue fluorophores using DLIPS will be necessary in understanding this response band and its diagnostic merit. Th e severe drop in performance for identifying pathology during weeks five through eight as further quantified in a detailed chemometic analysis (Kozikowki, et al. 2012) is likely due to emission fluctuations associated with the data collection process or t he particular spots being interrogated on the back of particular mice. However, the detection performance improved uniformly with the maturation of the DMBA induced pathology. In the age of computer aided medical diagnostics, techniques that add new inform ation to the decision making process (i.e. orthogonal sensing) are invaluable. These findings show preliminary support for applying this technique as a stand alone medical diagnostic tool or as a complementary technique to traditional fluorescence spectros copy for the detection of pathology in vivo. Of particular note were the findings that the DLIPS spectra emphasize different spectral regions and the endogenous pathology detection potential for DLIPS shows superior performance in this study. This techniqu e couples specifically to morphological changes in the ECM, which is unique compared to laser induced fluorescence.
118 Figure 3 1 T he DLIPS system Figure 3 2 The structures of glycine, alanine and proline Figure 3 3 T he structures of glycine g lycine, glycine L proline, and L alanine glycine Figure 3 4 T amino acid
119 Figure 3 5 T he traditional fluorescence spectra of the amino acids (i.e. preperturbation) Figure 3 6. Key of color that represents each studied amino acid Figure 3 7. T he traditional fluorescence spectra of L alanine (blue) and L alanine glycine (red) (i.e. preperturbation)
120 Figure 3 8. T he traditional fluorescence spectra of glycine (light blue) and glycine glycine (orange) (i.e. preperturbation) Figure 3 9. T he traditional fluorescence spectra of L prolin e (green) and glycine L proline (violet) (i.e. preperturbation)
121 Figure 3 10 T he DLIPS spectra of the amino acids Figure 3 11 T he DLIPS spectra of L alanine (blue) and L alanine glycine (red) after 700 laser pulses
122 Figure 3 12 T he DLIPS spectra of glycine (light blue) and glycine glycine (orange) after 500 laser pulses Figure 3 13. T he DLIPS spectra of L proline (green) and glycine L proline (violet) after 600 laser pulses
123 Figure 3 14. T he DLIPS spectra of glycine (light blue) after 500 laser pulses, L alanine (blue) after 700 laser pulses and L proline (green) after 600 laser pulses Figure 3 15 T he DLIPS spectra of glycine glycine (orange) after 500 laser pulses, L alanine glycine (red) after 700 laser pulses and glycine L proline (violet) after 600 laser pulses
124 Figure 3 16. T he traditional Raman spectra of the amino acids (i.e. preperturbation). Each spectrum other than proline pre has been shifted vertically to show distinction between Raman peaks of each amino acid Figure 3 17 T he traditional Raman spectra of L alanine (blue) and L alanine glycine (red) (i.e. preperturbation)
125 Figure 3 18 T he traditional Raman spectra of glycine (light blue) and glycine glycine (orange) (i.e. preperturbation) Figure 3 19 T he traditional Raman spectra of L proline (green) and glycine L proline (violet) (i.e. preperturbation)
126 Figure 3 20 The traditional Raman preperturbation (blue) spectrum of L alanine Figure 3 21 The traditional Raman preperturbation (red) spectrum of L alanine glycine
127 Figure 3 22. The traditional Raman preperturbation (light blue) spectrum of glycine Figure 3 23 T he traditional Raman preperturbation (orange) spectrum of glycine glycine
128 Figure 3 24 The traditional Raman preperturbation (green) spectr um of L prolin e Figure 3 25 The traditional Raman preperturbation (purple) spectrum of glycine L proline
129 Figure 3 26. T he DLIPS of the amino acids Figure 3 2 7 The traditional fluorescence preperturbation spectra of the collagenous material
130 Figure 3 2 8 T he col or that represents each studied collagenous material Figure 3 2 9 A close up view of the absolute fluorescence spectra of the collagenous material. (Fibrinogen and IgG are not plotted here) Figure 3 30 T he DILPS spectra of the collagenous material aft er 2000 laser pulses
131 Figure 3 31 A close up view of the DLIPS spectra of the collagenous material after 2000 laser pulses. (IgM is not plotted here) Figure 3 3 2 The traditional Raman spectra of the collagenous material. Each spectrum other than IgM has been shifted vertically to show distinction between Raman peaks of each collagenous material.
132 Figure 3 3 3 The traditional preperturbation spectrum of collagen Figure 3 3 4 T he traditional preperturbation spectrum of fibrinogen
133 Figure 3 3 5 The t raditional preperturbation spectrum of the IgA Figure 3 3 6 The traditional preperturbation spectrum of IgG
134 Figure 3 3 7 The traditional preperturbation spectrum of IgM
135 Figure 3 38. The DLIPS of Collagenous and related biological material
136 Figure 3 39 T he average DLIPS spectra at weeks 1 11 for the untreated skin
137 Figure 3 40 T he average DLIPS spectra at weeks 1 11 for the DMBA treated skin Figure 3 41 T he average absolute fluorescence for the DMBA treated week 2
138 Figure 3 42 T he average absolute fluorescence for the DMBA week 6 Figure 3 4 3 T he average absolute fluorescence for the DMBA week 8 Figure 3 4 4 T h e average DLIPS fluorescence for the DMBA week 2
139 Figure 3 4 5 T he average DLIPS fluorescence for the DMBA week 6 Figure 3 4 6 T he average DLIPS fluorescence for the DMBA nd untreated skin at week 8
14 0 Figure 3 4 7 T he H&E stained histology section of skin from an untreated mouse at week 4 Figure 3 4 8 T he H&E stained histology section of skin from a DMBA treated mouse at week 4. e stratum corneum, symptomatic of orthokeratotic hyperkeratosis. (X) highlights the thickening of the stratum spinosum, characteristic of acanthosis. (*) denotes areas of increased dermal collagen density
141 Figure 3 4 9 T he H&E stained histology section o f skin from a DMBA treated mouse at week 8. orthokeratotic hyperkeratosis. (X) highlights the thickening of the stratum spinosum, characteristic of acanthosis. (*) denotes areas of increased dermal collagen density Figure 3 50 T he H&E stained histology section through a developed papilloma of skin from a DMBA treated mouse at week 11
142 Table 3 1. T he band identity, wavenumber, and band assignment for L alanine L Alanine Band Identity Wavenumber (cm 1 ) Band Assignment A 661 CC stretch, C=O bend B 781 CO bend C 870 CC stretch, CN bend (normalization peak) D 1024 CH 3 rock, CN stretch E 1125 NH 2 rock, CH 3 rock F 1159 CH bend G 1319 CH bend H 1374 CH 3 sym. bend I 1422 CH 3 asym. bend Table 3 2. T he band identity, wavenumber, and band assignment for L alanine glycine L alanine glycine Band Identity Wavenumber (cm 1 ) Band Assignment A 603 CO bend B 870 CN stretch, CC stretch (n ormalization peak) C 942 CN stretch, CC stretch D 1024 CH 3 asym. rock CN stretch E 1260 Amide III F 1291 CH bend G 1355 CO bend, CH bend H 1413 CH 3 sym. bend I 1477 Amide II J 1665 C=O stretch, NH 2 bend Table 3 3. The band identity, wavenumber, and band a ssignment for glycine Glycine Band Identity Wavenumber (cm 1 ) Band Assignment A 706 CH bend B 897 CC stretch C 930 CC stretch, CO bend D 1054 CN stretch E 1159 CH bend F 1330 CH bend ( normalization peak) G 1413 CO stretch H 1453 CH bend
143 Table 3 4 T he band identity, wavenumber, and band assignment for glycine glycine Glycine glycine Band Identity Wavenumber (cm 1 ) Band Assignment A 732 NH 2 rock B 912 CC stretch C 975 CC stretch ( normalization peak) D 1014 CH bend, CN stre tch E 1142 CH bend, CN stretch F 1159 CH bend G 1247 Amide III H 1260 CC stretch, CO bend I 1453 CH bend J 1665 C=O stretch, NH 2 bend Table 3 5. T he band identity, wavenumber, and band assignment for L proline L Proline Band Identity Wave number (cm 1 ) Band Assignment A 626 CC bend B 722 NH 2 rock, NH bend C 853 CN stretch D 912 CC stretch ( normalization peak) E 930 CC stretch, CO bend F 989 CC stretch, CH 2 rock G 1047 CN stretch H 1075 CC stretch I 1096 CH bend, CN H 2 asym. bend J 1185 NH 3 rock CH bend K 1230 CN bend, NH 3 rock L 1278 CH bend, NH rock M 1384 CO bend
144 Table 3 6 T he band identity, wavenumber, and band assignment for glycine L proline Glycine L proline Band Identity Wavenumber (c m 1 ) Band Assignment A 815 CC stretch, CH 2 rock B 867 CC stretch, CH 2 stretch C 924 CC stretch, CN bend ( normalization peak) D 989 CC stretch, CH 2 rock E 1054 CN stretch F 1185 NH 3 rock, CH bend G 1195 NH 3 rock CN 2 bend H 1206 CC stretch, CO stretch I 1278 CH bend, NH rock J 1302 CH 2 wag K 1323 CH 2 wag, CH bend L 1413 CO bend M 1457 CH 2 bend N 1484 CH 2 bend, NH 3 sym. bend O 1665 C=O stretch, NH 2 bend Table 3 7. T he wavenumber, factor, and band assignment for all traditiona l Raman studied amino acids All Materials Raman Wavenumber (cm 1 ) Factor Band Assignment 855 2 ( 0.1704) CH bend 867 1 (0.1154), 3 (0.1201) CC stretch, CH 2 stretch 897 3 ( 0.1356) CC stretch 937 1 (0.1651) CC stretch 971 2 ( 0.2252), 3 (0.1252) CC wag 1328 3 ( 0.2470) COO sym stretch 1485 1 (0.1371) CH bend, CH wag 1655 2 (0.1237) Amide I (C=O stretch) 1666 1 (0.1270) Amide I (C=C stretch) Table 3 8. T his is the wavenumber, factor, and band assignment for traditional Raman studied peptides Gly Pro, Ala Raman Wavenumber (cm 1 ) Factor Band Assignment 538 2 ( 0.1293), 3 (0.1265) Amide IV (NCC bend, C=O stretch) 855 1 (0.1574), 2 ( 0.1874), 3 (0.1889) CH out of plane bend 867 3 ( 0.1276) CC stretch, CH 2 stretch 897 1 ( 0.1181) CC stretch 9 12 1 (0.1790), 2 (0.2138) CC stretch 1323 3 (0.2281) CH 2 wag, CH bend 1328 1 ( 0.2186) COO sym stretch 1455 3 (0.1111) CH 3 asym bend
145 Table 3 9. The wavenumber, factor, and band assignment for traditional Raman studied dipeptides GlyGly, GlyPro, AlaGly Raman Wavenumber (cm 1 ) Factor Band Assignment 817 1 (0.1039), 3 (0.1520) CC stretch 867 2 ( 0.2104) CC stretch, CH 2 stretch 937 1 (0.1456) CC stretch 971 1 ( 0.1222), 2 (0.1919) CC wag 1278 3 ( 0.1068) CH bend, NH rock 1447 3 (0.1053) CH 2 bend 1 485 1 (0.1328), 3 ( 0.1065) CH bend, CH wag 1666 1 (0.1229) Amide I (C=C stretch) Table 3 10. T he wavenumber, factor, and band assignment for all DLIPS studied amino acids All Materials DLIPS Wavenumber (cm 1 ) Factor Band Assignment 538 2 ( 0.2594) Amide IV (NCC bending, C=O stretch) 836 1 ( 0.0697), 3 ( 0.1523) CO 2 out of plane bend 963 3 (0.0995) CH bend 1323 2 ( 0.1443), 3 (0.1261) CH 2 wag, CH bend 1340 2 (0.1199), 3 ( 0.1171) Amide III (CH 2 wag) 1367 2 (0.1106) CH 2 twist 1422 2 (0.1231) CH 3 asym bend Table 3 11 The wavenumber, factor, and band assignment for DLIPS studied peptides Gly, Pro, Ala DLIPS Wavenumber (cm 1 ) Factor Band Assignment 538 2 ( 0.2594) Amide IV (NCC bending, C=O stretch) 836 1 ( 0.0814) CO 2 out of plane bend 10 01 3 (0.0963) CC stretch 1323 2 ( 0.1443), 3 (0.1126) CH 2 wag, CH bend 1340 2 (0.1199) Amide III (CH 2 wag) 1367 2 (0.1106) CH 2 twist 1422 2 (0.1231) CH 3 asym bend
146 Table 3 12. T he wavenumber, factor, and band assignment for all DLIPS studied d ipeptides GlyGly, GlyPro, AlaGly DLIPS Wavenumber (cm 1 ) Factor Band Assignment 554 1 ( 0.0821) CN twist, NH bend 662 1 ( 0.0834) Amide IV (NCC bend, C=O stretch) 867 2 (0.1753), 3 (0.0957) CC stretch, CH 2 stretch 893 3 (0.0954) CC stretch 926 2 ( 0.0924) CH 2 rock 963 2 (0.1381) CH bend 975 2 ( 0.1445) CC stretch 1070 3 (0.1401) CC stretch 1246 2 ( 0.1116) CH 2 twist 1666 3 ( 0.1409) Amide I (C=C stretch) Table 3 13. T he predicted class of all materials (i.e. peptides and dipeptides) studied u sing traditional Raman within the training set Raman Training Set Glycine Proline Alanine GlyGly GlyPro AlaGly Glycine (19 spectra) 100% 0% 0% 0% 0% 0% Proline (18 spectra) 0% 100% 0% 0% 0% 0% Alanine (18 spectra) 0% 0% 100% 0% 0% 0% GlyGly (20 spectra) 0% 0% 0% 100% 0% 0% GlyPro (19 spectra) 0% 0% 0% 0% 100% 0% AlaGly (15 spectra) 0% 0% 0% 0% 0% 100% Table 3 1 4 T he predicted class of all materials (i.e. peptides and dipeptides) studied using tradi tional Raman within the validation set Rama n Validation Set Glycine Proline Alanine GlyGly GlyPro AlaGly Glycine (20 spectra) 100% 0% 0% 0% 0% 0% Proline (20 spectra) 0% 95% 5% 0% 0% 0% Alanine (20 spectra) 0% 0% 100% 0% 0% 0% GlyGly (20 spectra) 0% 15% 0% 85% 0% 0% GlyPro (20 spectra) 0% 65% 0% 0% 25% 20% AlaGly (20 spectra) 0% 65% 35% 0% 0% 0% Table 3 15 The predicted class of the peptides studied using traditional Ra man within the training set Raman Training Set Glycine Proline Alanine Glycine (19 spectra) 100% 0% 0% Prolin e (18 spectra) 0% 100% 0% Alanine (18 spectra) 0% 0% 100%
147 Table 3 16 T he predicted class of the peptides studied using tradi tional Raman within the validation set Raman Validation Set Glycine Proline Alanine Glycine (20 spectra) 100% 0% 0% Prolin e (20 spectra) 0% 95% 5% Alanine (20 spectra) 0% 0% 100% Table 3 17 T he predicted class of the dipeptides studied using traditional Raman within the training set Raman Training Set GlyGly GlyPro AlaGly GlyGly (20 spectra) 100% 0% 0% GlyPro (19 s pectra) 0% 100% 0% AlaGly (15 spectra) 0% 0% 100% Table 3 18 T he predicted class of the dipeptides studied using tradi tional Raman within the validation set Raman Validation Set GlyGly GlyPro AlaGly GlyGly (20 spectra) 100% 0% 0% GlyPro (20 spec tra) 0% 95% 5% AlaGly (20 spectra) 0% 0% 100% Table 3 19 T he predicted class of all materials (i.e. peptides and dipeptides) studied using t he DLIPS method within the training set DLIPS Training Set Glycine Proline Alanine GlyGly GlyPro AlaGly Glycine (15 spectra) 100% 0% 0% 0% 0% 0% Proline (20 spectra) 0% 100% 0% 0% 0% 0% Alanine (16 spectra) 0% 0% 100% 0% 0% 0% GlyGly (19 spectra) 0% 0% 0% 100% 0% 0% GlyPro (20 spectra) 0% 0% 0% 0% 100% 0% AlaGly (14 spectra) 0% 0% 0% 0% 0% 100% Tabl e 3 20 T h e predicted class of all materials (i.e. peptides and dipeptides) studied using t he DLIPS method within the validation set DLIPS Validation Set Glycine Proline Alanine GlyGly GlyPro AlaGly Glycine (16 spectra) 100% 0% 0% 0% 0% 0% Prolin e (20 spectra) 0% 100% 0% 0% 0% 0% Alanine (20 spectra) 0% 0% 100% 0% 0% 0% GlyGly (20 spectra) 0% 0% 0% 100% 0% 0% GlyPro (20 spectra) 0% 0% 0% 0% 100% 0% AlaGly (15 spectra) 0% 0% 0% 0% 0% 100%
148 T able 3 21 The predicted class of the peptides studie d using the DLIPS method within the training set DLIPS Training Set Glycine Proline Alanine Glycine (15 spectra) 100% 0% 0% Proline (20 spectra) 0% 100% 0% Alanine (16 spectra) 0% 0% 100% Table 3 22 T he predicted class of the peptides studied us ing t he DLIPS method within the validation set DLIPS Validation Set Glycine Proline Alanine Glycine (16 spectra) 100% 0% 0% Proline (20 spectra) 0% 100% 0% Alanine (20 spectra) 0% 0% 100% Table 3 2 3 T he predicted class of the dipeptides studied u sing t he DLIPS method within the training set DLIPS Training Set GlyGly GlyPro AlaGly GlyGly (19 spectra) 100% 0% 0% GlyPro (20 spectra) 0% 100% 0% AlaGly (14 spectra) 0% 0% 100% Table 3 24 T he predicted class of the dipeptides studied using t he DLIPS method within the validation set DLIPS Validation Set GlyGly GlyPro AlaGly GlyGly (20 spectra) 100% 0% 0% GlyPro (20 spectra) 0% 100% 0% AlaGly (15 spectra) 0% 0% 100% Table 3 2 5 Th e band identity, wavenumber, and band assignment for colla gen Collagen Band Identity Wavenumber (cm 1 ) Band Assignment A 392 CN bend B 735 NH 2 rock (normalization peak) C 1310 CH 2 twist, CH bend D 1392 CN stretch, NH 3 bend
149 Table 3 2 6 T he band identity, wavenumber, and band assignment for fibrinogen Fibrinogen Band Identity Wavenumber (cm 1 ) Band Assignment A 392 CN bend B 735 NH 2 rock (normalization peak) C 1016 CC stretch D 1070 CC stretch E 1139 CH bend, CN stretch F 1310 CH 2 twist, CH bend G 1392 CN stretch, NH 3 bend H 1454 CH bend, CH 2 stretching Table 3 2 7 T he band identity, wavenumber, and band assignment for IgA IgA Band Identity Wavenumber (cm 1 ) Band Assignment A 392 CN bend B 735 NH 2 rock (normalization peak) C 1227 CC stretching D 1310 CH 2 twist, CH bend E 1392 CN str etch, NH 3 bend Table 3 2 8 T he band identity, wavenumber, and band assignment for IgG IgG Band Identity Wavenumber (cm 1 ) Band Assignment A 392 CN bend B 735 NH 2 rock (normalization peak) C 1016 CC stretch D 1392 CN stretch, NH 3 bend E 1454 CH be nd, CH 2 stretching F 1676 Amide I, C=O stretch Table 3 29 T he band identity, wavenumber, and band assignment for IgM IgM Band Identity Wavenumber (cm 1 ) Band Assignment A 422 CC bend B 524 CN stretch C 806 CH bend, CN stretch (normalization peak) D 920 CC stretch E 1035 CH bend, CN stretch F 1070 CC stretch G 1156 CN stretch H 1260 Amide III I 1310 CH 2 twist, CH bend J 1480 Amide II K 1594 C=O stretch, CO 2 bend
150 Table 3 3 0 T he wavenumber, factor, and band assignment for all traditional R aman technique studied materials All Materials Raman Wavenumber (cm 1 ) Factor Band Assignment 360 1 (0.1002) CC stretch 411 1 (0.1161) NH 2 wag 489 1 (0.0885) CO 2 rock 735 1 ( 0.1017), 3 (0.1431) NH 2 rock 1008 2 (0.1067) CO stretch, CC stretch 106 2 3 ( 0.1148) CH 3 rock 1130 2 (0.1104), 3 ( 0.2016) CC stretch, CH 2 rock 1302 2 (0.1141), 3 ( 0.1868) CH 2 wag 1447 2 ( 0.1087), 3 ( 0.1261) CH 2 bend Table 3 31 T he wavenumber, factor, and band assignment for only the Igs traditional Raman technique studied materials Ig s only Raman Wavenumber ( cm 1 ) Factor Band Assignment 360 1 (0.0991) CC stretch 411 1 (0.1155) NH 2 wag 489 1 (0.0882) CO 2 rock 735 1 ( 0.0898), 3 (0.1431) NH 2 rock 1008 2 (0.1164) CO stretch, CC stretch 1062 3 ( 0.1148) CH 3 roc k 1130 3 ( 0.2016) CC stretch, CH 2 rock 1240 2 (0.0837) Amide III(CH 2 wag, CN stretch) 1302 3 ( 0.1868) CH 2 wag 1447 2 (0.0857), 3 ( 0.1261) CH 2 bend Table 3 32 T he wavenumber, factor, and band assignment for all the DLIPS method studied materials A ll Materials DLIPS Wavenumber (cm 1 ) Factor Band Assignment 360 1 (0.1009) CC stretch 392 1 (0.1089) CN bend 444 1 (0.1038) NCC bend 489 1 (0.0949) CO 2 rock 735 3 ( 0.0949) NH 2 rock 806 3 ( 0.0966) CH bend, CN stretch 1008 2 (0.1062) CO stretch, CC stretch 1099 3 (0.0845) CH 2 wag, CN stretch 1130 3 (0.1807) CC stretch, CH 2 rock 1240 2 (0.0899) Amide III 1302 3 (0.1590) CH 2 wag 1447 3 (0.1143) CH 2 bend
151 Table 3 3 3 T he wavenumber, factor, and band assignment for only the Igs DLIPS method stud ied materials Ig s only DLIPS Wavenumber (cm 1 ) Factor Band Assignment 392 1 (0.1093) CN bend 444 1 (0.1034) NCC bend 489 1 (0.0948) CO 2 rock 806 3 ( 0.0948) CH bend, CN stretch 1008 2 (0.1106) CO stretch, CC stretch 1062 3 (0.0900) CH 3 rock 1099 3 (0.1083) CH 2 wag, CN stretch 1130 3 (0.1879) CC stretch, CH 2 rock 1240 2 (0.0929) Amide III 1302 3 (0.1534) CH 2 wag 1447 3 (0.1020) CH 2 bend Table 3 3 4 T he predicted class of all materials studied using traditional Raman within the training set Ra man Training set Collagen Fibrinogen IgA IgG IgM Collagen (14 spectra) 100% 0% 0% 0% 0% Fibrinogen (14 spectra) 0% 100% 0% 0% 0% IgA (14 spectra) 0% 0% 100% 0% 0% IgG (14 spectra) 0% 0% 0% 100% 0% IgM (14 spectra) 0% 0% 0% 0% 100% Table 3 3 5 T he predicted class of all materials studied using traditional Raman within the validation set Raman Validation set Collagen Fibrinogen IgA IgG IgM Collagen (14 spectra) 100% 0% 0% 0% 0% Fibrinogen (18 spectra) 0% 100% 0% 0% 0% IgA (17 spectra) 0% 0% 88.20% 11.70% 0% IgG (15 spectra) 0% 33.30% 0% 66.70% 0% IgM (16 spectra) 0% 0% 0% 0% 100% Table 3 3 6 T he predicted class of only the Igs studied using traditional Raman within the training set Raman Training set IgA IgG IgM IgA (14 spectra ) 100% 0% 0% IgG (14 spectra) 0% 100% 0% IgM (14 spectra) 0% 0% 100%
152 Table 3 3 7 T he predicted class of only the Igs studied using traditional Raman within the validation set Raman Validation set IgA IgG IgM IgA (17 spectra) 82.35% 17.65% 0% IgG (15 spectra) 0% 100% 0% IgM (16 spectra) 0% 0% 100% Table 3 3 8 T he predicted class of all materials studied using the DLIPS method within the training set DLIPS Training set Collagen Fibrinogen IgA IgG IgM Collagen (19 spectra) 100% 0% 0% 0% 0% Fibrinogen (18 spectra) 22.22% 77.78% 0% 0% 0% IgA (19 spectra) 0% 0% 100% 0% 0% IgG (19 spectra) 0% 5% 0% 95% 0% IgM (18 spectra) 0% 0% 0% 0% 100% Table 3 39 T he predicted class of all materials studied using the DLIPS method within the validation set DLIPS Validation set Collagen Fibrinogen IgA IgG IgM Collagen (19 spectra) 89.50% 10.50% 0% 0% 0% Fibrinogen (19 spectra) 15.80% 79.00% 0% 5.20% 0% IgA (19 spectra) 0% 0% 100% 0% 0% IgG (19 spectra) 0% 0% 21.05% 78.95% 0% IgM (18 spectra) 0% 0% 0% 0% 100% Table 3 40 T he predicted class of only the Igs studied using the DLIPS method within the training set DLIPS Training set IgA IgG IgM IgA (19 spectra) 100% 0% 0% IgG (19 spectra) 0% 100% 0% IgM (18 spectra) 0% 0% 100% Table 3 4 1 T he predicted class of only the Igs studied using the DLIPS method within the validation set DLIPS Validation set IgA IgG IgM IgA (19 spectra) 100% 0% 0% IgG (19 spectra) 21% 79% 0% IgM (18 spectra) 0% 0% 100%
153 Table 3 42 T he value of the x a xis intercept of type IV collagen, fibrinogen, IgA, and IgG Collagenous Material X axis Intercept Wavelength (nm) Type IV Collagen 390 Fibrinogen 500 IgA 420 IgG 450
154 CHAPTER 4 SUMMARY OF RESEARCH AND PROPOSAL OF FUTURE WORK 4.1. Summary of Research The focus of this ongoing research is to analyze biological material s using the differential laser induced perturbation spectroscopy (DLIPS) technique. The DLIPS sensing scheme incorporates three complementary techniques to improve upon previous fluoresce nce based biosensing strategies: laser induced fluorescence emission or Raman scattering ultraviolet (UV) laser perturbation of the biological material, and difference spectroscopy. The intimate coupling of the perturbing UV light source to the biological matrix molecular structures of abnormal tissue (e.g. dysplastic precancerous cells) will respond differently than normal tissue structure. The proposed technique of DLIPS examines differ ences in fluorescence response or Raman response hence the resulting differential response will potentially avoid the major limitation of the current biosensing schemes, namely the significant variations in the absolute optical response, as generally observed in patient to patient populations. In summary, the DLIPS techniqu e takes advantage of the specificity of laser tissue coupling in combination with difference spectroscopy to offer an essential step forward in biodetection and/or bioi maging. Throughout this work, the DLIPS sensing approach has provided significant enhanc ements in sensitivity and specificity which ultimately are necessary to function in real world (i.e. compl ex backgrounds) environments. Traditional fluorescence and Raman based biosensing schemes are intimately linked to the local cellular structure. The strong fluorescence peak centered on 460 nm for the fluorescence spectra recorded for the mice skin, can most reasonably be assigned to the tissue fluorophore nicotinamide adenine dinucleotide ( NADH ) While this is an important tissue fluorophore, and is a primary target of fluorescence tiss ue assessment, it serves as a marker of
155 increased cellular metabolism, which is a hallmark of dysplasia, but not necessarily unique to it. In conjunction with increased cellular metabolism, one of the earliest markers of pre cancerous progression is cellul ar infiltration and the release of growth factors and cytokines. This unchecked cellular signaling, results in the proliferation of fibroblasts, increased collagen synthesis, and suppression of collagenase production, with the overall effect of restructuri ng the local extra cellular matrix ( ECM ) These effects suggest that techniques which target collagen remapping might provide direct diagnostic coupling of the spectra to the pathology. Hence in principle the development of neoplastic disease and cellular change are potentially d ete ctable in the optical response. However, the variability of the fluorescence and Raman response among patient populations in real world environments, as well as the often broadband nature of biological fluorescence (i.e. spectrally broad and lacking in spectrally sharp features) has severely limited their applicability to date Significant patient to patient variation in fluorescent and optical properties has been shown to be associated with race, age, sex, air temperature, and even deformation of the ti ssue when applying the probe. Patient and sampling variability includes: fluctuations in absolute emission intensity, emission peak shifts, and changes in the scattering and absorption properties of th e tissue, among other effects. A truly new methodology that provides improved sensitivity and/or specificity is highly desirable and could be used for in situ and in vivo cancer screening, detection of biological pathogens for biodefense, as well as food (e.g. pathogen detection) and building safety. The benef its of improved screening and earlier diagnosis are the promoting of earlier interdiction, and therefore primary caregivers need new instruments and methods to continue pushing forward the time lines for disease detection and diagnosis. Of particular utili ty are strategies that reduce the
156 need for highly trained personnel and healthcare infrastructure to support screening and diagnosis, instead offering sample to answer capabilities at the point of care (POC). In the DLIPS scheme, a combination of optical p robing (e.g. fluorescence or Raman), UV photochemical perturbation, and repeat optical probing to realize a powerful new spectral dimension based on difference spectroscopy that will be strongly coupled to the local biomolecular matrix thereby providing a differential response with expectations of enhanced sensitivity and specificity. In this technique, the pertur bation pulses from the deep UV e xcimer laser (193 nm, 6.4 eV) are strongly absorbed by biological tissue and while no direct ablation is realized, a single photon of 193 nm radiation exceed s nearly all bond energies in the biological matrix; hence permanent photochemistry is induced despite being below the critical photon flux to affect material removal. The permanent alteration of the underlying ma terial structure will result in changes within the fluorescence spectrum, specifically with respect to photoreactive biomolecules. The perturbation pulses are used to cleave molecular bonds and the irradiation of biological matrices at 193 nm can cause pho toionization, including strand breakage, locally denatured sites, interstrand cross linking, reactions via photo dimers, and other products. In summary of the DLIPS technique, this method may mitigate unwanted contributions from unperturbed bio logical material fluorophores, broadband fluorescence, and importantly, variations in fluorescence emission bands which are unique to the patient, but not necessarily to the targeted pathology. The complexity of the local fluorescence environment provides the opportunity for the perturbing UV radiation to affect a unique change to the resulting fluorescence or Raman response. Also, the specificity of coupling to important photosensitive
157 biomarkers of early pathological changes has promise to mitigate the su bstantial noise due to inter patient variations. In summary, the work reported here has demonstrated that the biological matrix (collagen and amino acids) have been altered by low intensity ultraviolet radiation (primarily 193 nm) such that the resulting f luorescence and Raman scatter properties are perturbed. Probing the biomatrix with fluorescence monitoring both before and after perturbation with the deep UV light source affords the opportunity to look for a differential response (i.e. difference spec tr oscopy), as demonstrated in an amino acids, a collagenous materials, and an animal model study. The spectral regions where the fluorescence signal was particularly enhanced or suppressed due to e xcimer perturbation, for each of the above studies, were assi gned to a particular biological fluorophores found in the literature. For each amino acid studied, a distinctive spectral shape has immerged. Before 430 nm, another effect occurs in the DLIPS spectra of glycine glycine, L alanine glycine and glycine L prol ine The slope of glycine L proline decreases from 380 nm 420 nm, while the slopes of glycine glycine and L alanine glycine increases. Also, in comparing all absolute Raman preperturbation spectra, glycine glycine and L alanine glycine had a comm on band t hat occurred at 1260 cm 1 with the band assigned to a CC stretch and a CO bend. G lycine glycine and L alanine glycine, also showed to have a significant DLIPS value associated with an Amide III. Therefore, the increasing slope of glycine glycine and L alan ine glycine DLIPS spectra from 380 nm thru 420 nm may be attributed to the CC stretch and CO bend that occurs at 1260 cm 1 or the Amide III which had a significant DLIPS value associated with it, both of which did not occur in glycine L proline. The decrea sing slope of glycine L proline from 380 nm thru 420 nm could also be attributed to a CH 2 bend, a CH 2 wag and/or a NH 3 symmetric bend, all of which
158 were present in the absolute Raman preperturbation spectrum and have a significant DLIPS value in glycine L proline alone. In the amino acids study, when comparing the DLIPS spectra of the dipeptides with their peptide counterparts, it can be seen that the slopes after 430 nm, of L alanine glycine, glycine glycine, and glycine L proline are less than L alanine, glycine, and L proline. Again, in comparing these two groups, the absolute Raman preperturbation spectra of glycine glycine, L alanine glycine, and glycine L proline showed a common bond that occurs at 1665 cm 1 and is assigned to a C=O stretch and a NH 2 b end that was not present in L alanine, L proline, or glycine. In addition, the glycine L proline and L alanine glycine absolute Raman preperturbation spectra shared a common band at 924 cm 1 which is assigned to a CC stretch and a CN bend. Therefore, the reason for the dipeptides having smaller positive slopes after 430 nm than the peptides may be because of the common bond that occurs at 1665 cm 1 and is assigned to a C=O stretch and a NH 2 in the dipeptides. In the studied collagenous and related biologic al materials, the DLIPS spectrum of IgM, which occurs at a much higher intensity and with a completely different spectral shape than the other studied collagenous material, may be attributed to the specific band wavenumbers and assignments that were only f ound in IgM. Those are a CC bend that occurs at 422 cm 1 a CN stretch that occurs at 524 cm 1 and a CH bend and a CN stretch that occurs at 806 cm 1 and a CC stretch that occurs at 920 cm 1 a CH bend and a CN stretch that occurs at 1035 cm 1 a CN stre tch that occurs at 1156 cm 1 an Amide III that occurs at 1260 cm 1 an Amide II that occurs at 1480 cm 1 and a C=O stretch and CO 2 asymmetric stretch that occurs at 1594 cm 1 However, it should also be noted that IgM films used in this study was the onl y collagenous material studied
159 which was diluted to a 1:10 ratio to deionized water (1:1 ratio to deionized water was used for all others presented in this study). As mentioned above, the DLIPS spectra of the studied collagenous material seem to follow a s imilar pattern occurring at different intensities, except for IgM which occurs at a much higher intensity with a completely different spectral shape. The cause of this specific pattern may be related to specific wavenumber and band assignments of molecules found only in type IV collagen, fibrinogen, IgA and IgG. Those bands are a CN bend that occurs at 394 cm 1, a NH2 rock that occurs at 735 cm 1, and a CN stretch and NH3 bend that occurs at 1392 cm 1. Interestingly, also in the collagenous materials study the x axis intercept of the DLIPS spectra of type IV collagen, fibrinogen, IgA and IgG, (i.e. were the intensity changed from negative to positive), was unique. Therefore, the x axis intercepts of the DLIPS spectra could be used as a means of differentia tion. The cause of the different x axis intercepts may be attributed to a CH bend and CN stretch that occurs at 1139 cm 1 which was only seen in fibrinogen, a CC stretch that occurs at 1227 cm 1 which was only seen in IgA, or to an Amide I and C=O stretc h at 1676 cm 1 which was only observed in IgG. In the animal model study, a promising preliminary evaluation of DLIPS for the in vivo detection of disease was presented. DLIPS has shown to offer a new source of unique information and serves as a convenie nt probe for ECM structure that was previously inaccessible when exciting natural fluorescence at 355 nm, as it has been noted that at this excitation wavelength it is not possible to resolve collagen fluorescence from that of NADH. Of particular note were the findings that the DLIPS spectra emphasize different spectral regions and the endogenous pathology detection potential for DLIPS showed superior performance. The DLIPS technique coupled specifically to morphological changes in the ECM, which is unique compared
160 to las er induced fluorescence. It was demonstrate d that this technique is primarily sensitive to changes induced by the pathology model. In summary, these findings show preliminary support for applying this technique as a stand alone medical diagn ostic tool or as a complementary technique to traditional fluorescence or Raman spectroscopy for the early detection of biological material changes in vivo However, in pursuit of fully understanding the DLIPS spectra response, its assignments, and its dia gnostic merit, further investigation using DLIPS will be necessary. Association of all parts of a biological materials absolute preperturbation Raman spectrum, which shows the molecular structure, with all parts of its corresponding DLIPS spectrum would gi ve a complete picture of the biological materials photochemical response as well as creating a method of comparing and contrasting similar biological material DLIPS spectra 4.2. Proposal of Future Work The long term vision for this work is to develop the DLIPS technology into a clinical system that impacts disease morbidity and mortality by facilitating earlier, non invasive diagnosis and staging of diseases. Specifically, the future work proposed here is to develop and test the DLIPS system as well as co ntinue to push the technology to other spectral domains, further investigate the fluorophores responsible for the spectral changes under DLIPS detection, therefore, increasing the amount of information obtained about the interrogated biological material. T o accomplish these goals, the proposed future work includes the refinement of this technique in animal and collagen based material studies, through the introduction of in situ Raman spectroscopy, a fiber probe, and a tunable laser i nto the DLIPS scheme. Th rough the introduction of in situ Raman spectroscopy into the DLIPS scheme, a specimen could be analyzed using Raman spectroscopy in the exact location as the specimen is
161 perturbed. This would mean that the pre and postperturbation Raman spectra represent the effect on exact same area (i.e. some ensemble of molecules) before and after it is perturbed. In the current Raman DLIPS scheme, the specimen is moved in between the pre and postperturbation spectra which create diff iculties in spot realignment. The cu rrent laser material interface will be transformed to allow for minimally invasive sensing of biological material. Especially for future animal studies, a fiber probe must be custom designed and manufactured for the DLIPS scheme and applications. The fiber probe must be capable of detection and excitation, as well as a probe composition of fibers that are capable of transm itting deep UV radiation down t o a value of 193 nm. Typically, standard optical probes are composed of silica fibers and radiation below 300 nm (deep UV) degrades there transmission. As transmission degrades, the light absorption in the fiber increases and data loses validity. Also, the excitation energy should be directed to the region in front of the detection fiber to insure that the reg ion of overlap between the excitation and emission fibers is optimized, resulting in the complete detection of all emissions. To maximize the UV throughput, the probe length should be as short as the assembly allows. The fiber housing materials and adhesiv es should also be considered in designing the probe. Ocean Optics sells an Xtreme Solarization resistance optical fiber (XSR) which has silica core fibers that are doped with fluorine to mitigate the solarizing effects of UV radiation and provides a custom izing option in probe fibers. Therefore, a custom fiber probe can be manufactured to these specifications in a timely manner. A tunable laser is a laser in which the output wavelength can be changed. Adding a tunable laser to the DLIPS system will result i n extending the biosensing capabilities of the DLIPS technique. The emission intensities can now be a function of excitation wavelength a function of emission wavelength or in an emission map, a function of excitation and emission
162 wavelengths. Because the excitation/emission wavelength pair is a combination of the molecular structure and the overall molecular environment (e.g. biological matrix), the DLIPS fluorescence spectra, when compared over a range of excitation wavelengths may provide an additional m ethod for discrimination among biological materials or increase the abilities of the current DLIPS technique. In addition, for the DLIPS with a Raman probe, careful analysis of the peak position should also be performed. For example, the 897 cm 1 peak of g lycine in Figure 3 26 shows a structure su ggesting a true peak shift. Such peak shifts may reveal additional new information about the perturbations, and may have diagnostic information themselves. Finally, in addition to using PCA analysis for classificat ion and for identification of bond cleavage, as done in this work, PCA could be explored for quantitative analysis of targeted species. For example, can IgA be isolated at different concentration levels in a collagen matrix.
163 Appendix Table A 1. The chemical index of all materials used Man ufacturer Product Name Product part number Sigma Aldrich Glycine G8898 Sigma Aldrich Proline P0380 Sigma Aldrich Alanine A7627 Sigma Aldrich Glycine Glycine G0199 Sigma Aldrich Glycine Proline G3002 Sigma Aldrich Alanine Glycine A0878 Fisher Scienti fic Water, deionized ultra filtered 7732 18 5 Sigma Aldrich Collagen C7521 Sigma Aldrich Fibrinogen F3879 Sigma Aldrich Immunoglobulin type A I1010 Sigma Aldrich Immunoglobulin type G I4506 Sigma Aldrich Immunoglobulin type M I8260
164 LIST OF REFEREN CES Alfano, R., Tata, D., Cordero, J., Tomashefsky, P., Longo, F., Alfano, M., "Laser induced fluorescence spectroscopy from native cancerous and normal tissue," Journal of Quantum Electronics 20 1507 1511 (1984). in nanosecond Er:YAG laser ablation of water, liver, and 81 329 338 (2005). induced thermal degradation and ablation of 68 615 625 (1999). Arnold, J., Dwek, 106 103 110 (2006). amino acid 108 3186 3197 (1986). N internal rotations and vibrational spectral analysis of non Biomolecular Spectroscopy 72 523 527 (2008). Bailey, A., Paul, R., Knott, L., "Mechanisms of maturation and ageing of collagen," Mechanisms of Aging and Development 106 1 56 (1998). nical importance of increased serum levels of immunoglobulin A and G in combination with normal or diminished immunoglobulin 16 195 198 (1967). ress in Biophysics and Molecular Biology 74 141 173 (2000). Bhartia, R., Hug, W., Salas, E., Reid, R., Sijapati, K., Tsapin, A., Abbey, W., Nealson, K., Lane, A., Conrad, P., "Classification of organic and biological material with deep ultraviolet excita tion," Journal of Applied Spectroscopy 62 1070 1077 (2008). 137 596 601 (2011). Birdwell, J., Vals 44 3246 3253 (2010). Bottiroli, G., Croce, A., Locatelli, D., Marchesini, R., Pignoli, E., Tomatis, S., Cu zzoni, C., DiPalma, S., Dalfante, M., Spinelli, P., "Natural fluorescence of normal and neoplastic human 16 48 60 (1995).
165 Brookner, C., Utzinger, U., Follen, M., Richards Kortum, R., Cox, D., Atkinson, E., "Effects of biographical variables on cervical fluorescence emission spectra," Journal of Biomedical Optics 8 479 483 (2003). Brownstein, M., Hoffman, R., Levenson, R., Milner, T., Dowell, M., Williams, P., White, G., Gaigalas, A., Hwang, J., "Biophotonic Tools in Cell and Tissue Diagnostics," Journal of Research of the National Institute of Standards and Technology 112 139 152 (2007). spectrosc 42 355 362 (2011). of quantitative conformer analyses by nuclear magnetic resonance and Raman optical activity for model dipept 112 8633 8640 (2008). Budesinsky, M., Sebestik, J., Bednarova, L., Baumruk, V., Safarik, M., Bour, P., Gly motif in the D Ala L Pro Gly D Ala model peptide explored by a statisti Journal of Organic Chemistry 73 1481 1489 (2008). Calabro, K., Curtis, A., Galarneau, J., Krucker, T., Bigio, I., "Gender variations in the optical properties of skin in murine animal mo dels," Journal of Biomedical Optics 16 011008 (2011). Calvert, J., Pitts, Jr., J., Photochemistry. New York: John Wiley and Sons, Inc., 1966. Carletti, P., Roldan, M., Francioso, O., Nardi, S., Sanchez of humic li ke substances with conventional and surface Journal of Molecular Structure 982 169 175 (2010). Chakravarti, D., Pelling, J., Cavalieri, E., Rogan, E., "Relating aromatic hydrocarbon induced DNA adducts and c H ras mutat ions in mouse skin papillomas: The role of apurinic sites," Proceedings of the National Academy of Science 92 10422 10426 (1995). Chang, R., General Chemistry. New York: McGraw Hill Companies Inc., 2003. Chatterjee, B., Kang, D., Herceg, E., Trenary, M., 119 ,10930 10941 (2003). resolved spectrofluorometry to measur 26 67 77 (1997). photothermal effect on skin tissue Ray Science and Technology 14 207 216 (2006).
166 examination of the electronic excited state of glycglycing and other dipeptides: Observation of a carboxylate Chemical Society 118 9705 9715 (1996). Raman spectroscopy used to identify and grade human skin 68 75 79 (2005). Chwirot, B., Kowalska, M., Sypniewska, N., Michniewicz, Z., G radziel, M., "Spectrally resolved fluorescence imaging of human colonic adenomas," Journal of Photochemistry and Photobiology B: Biology 50 174 183 (1999). Coghlan, L., Utzinger, U., Drezek, R., Heintzelmann, D., Zuluaga, A., Brookner, C., Richards Kortum R., Gimenez Conti, I., Follen, M., "Optimal fluorescence excitation wavelengths for detection of squamous intra epithelial neoplasia: results from an animal model," Optics Express 7 436 446 (2000). Coghlan, L., Utzinger, U., Richards Kortum, R., Brookne r, C., Zuluaga, A., Gimenez Conti, I., Follen, M., "Fluorescence Spectroscopy of Epithelial Tissue Throughout the Dysplasia Carcinoma Sequence in an Animal Model: Spectroscopic Changes Precede Morphologic Changes," Lasers in Surgery and Medicine 29 1 10 ( 2001). 39 8142 8148 (2005). on of Raman spectra of amino acids using 77 978 983 (2010). Diagaradjane, P., Yaseen, M., Yu, J., Wong, M., Anvari, B., "Autofluorescence Characterization for the Earl y Diagnosis of Neoplastic Changes in DMBA/TPA Induced Mouse Skin Carcinogenesis," Lasers in Surgery and Medicine 37 382 395 (2005). edited Raman spectroscopy of prot eins: A general strategy to probe individual peptide bonds with 123 7919 7929 (2001). transfer: a report fo European Journal of Chemical Physics and Physical Chemistry 10 1389 1398 (2009). Drezek, R., Sokolov, K., Utzinger, U., Boiko, I., Malpica, A., Follen, M., Richards Kortum, R., "Understanding the contributions of NADH and collagen to cervical tissue fluorescence spectra: Modeling, measurements, and implications," Journal of Biomedical Optics 6 385 396 (2001).
167 usion in 191 466 468 (1976). Eichhorn, R., "Early diagnosis of melanotic melanoma based on laser induced melanin fluorescence," Journal of Biomedical Optics 14 33 34 (2009). Eker, C., Rydell, R., Svanberg, K., Andersson Engels, S. "Multivariate analysis of laryngeal fluorescence spectra recorded in vivo," Lasers in Surgery and Medicine 28 259 266 (2001). driven protonation changes of internal aspartic acids of bacteriorhodopsin: An investigation by static and time resolved infrared difference spectroscopy using [4 13 Biochemistry 24 400 408 (1985). 221 628 631 (1981). Funovics, M., Weissleder, R., Mahmood, U., "Catheter based in Vivo Imaging of Enzyme Activity and Gene Expression: Feasibility Study in Mice1," Radiology 231 659 666 (2004). 38 1133 1147 (2007). Goletz, C., Wagner, M., Grubel, A., Schmidt, W., Korf, N., Wermer, P., fluorescence excitation emission 85 650 656 (2011). energy transfer measurements using fluorescence mic 74 2702 2713 (1998). Gorner, H., "Photochemistry of DNA and related biomolecules: quantum yields and consequences of photoionization," Journal of Photochemistry and Photobiology 26 117 139 (1994). Gortzak Uzan, L., Jimenez, W., Nofech Mozes, S., Ismiil, N., Khalifa, M., Dube, V., Rosen, B., Murphy, J., Laframboise, S., Covens, A., "Sentinel lymph node biopsy vs. pelvic lymphadenectomy in early stage cervical cancer: Is it time to change the gold standard?," Gynecologic Oncol ogy 116 28 32 (2010). of Raman Spectroscopy 28 125 129 (1997). dipepti 112 2712 2718 (2008).
168 48 261 268 (1976). induced breakdown spectroscopy (LIBS), Part I: Revi ew of basic diagnostic and plasma particle interactions: Still challenging issues within the analytical 64 336A 366A (2010). Haque, M., Harris, M., Meghji, S., Barrett, A., "Immunolocalization of cytokines and grow th factors in oral submucous fibrosis," Cytokine 10 713 719 (1998). Haringsma, J., Tytgat, G., "Fluorescence and autofluorescence," Bailliere's Clinical Gastroenterology 13 1 10 (1999). He, X., Xi, B., Wei, Z., Jiang, Y., Yang, Y., An, D., Cao, J., Liu, emission matrix spectroscopy with regional integration analysis for characterizing composition Materials 190 293 299 (2011). Hernandez short peptides in hydrated media. IV. Amino acids with hydrophobic side chains: L alanine, L valine, and L 113 3169 3178 (2009). to Trans isomerization of gly cy 121 3762 3766 (1999). induced sep aration and transport of living 12 054016 (2007). infrared Raman spectroscopy for optical diagnosis of lung 107 047 1052 (2003). Hudson, B., 268 26033 26036 (1992). Vibrational Spectroscopy 8 279 291 (1995). Jares 21 1387 1395 (2003). cta Part A 61 1585 1594 (2005). Jesneck, J., Nolte, L., Baker, J., Floyd, C., Lo, J., "Optimized approach to decision fusion of heterogeneous data for breast cancer diagnosis," Journal of Medical Physics 33 2945 2954 (2006).
169 Jhan, J., Chang, W., Chen, H photon imaging and Raman spectroscopy for characterizing structure constituent correlation of 16 16431 16442 (2008). Journee Progress in Retinal and Eye Research 21 303 317 (2002). infrared multichannel Raman spectroscopy toward real time 33 498 502 (2002). Kaneta, T., Maeda, H., Miyazaki, M., Miyake, R., Izaki, H., Sakoda, Y., Kinoshita, S., Imasaka, modified capillary electrophoresis laser tographic Science 46 712 716 (2008). Tetrahedron Letters 51 4670 4672 (2010). c acid) 18 881 890 (2011). 27 302 307 (1991). K leinsmith, L., Principles of Cancer Biology San Francisco: Pearson Education, Inc., 2006. and Laser Surgery 22 411 417 (2004). Kollias, N., Gillies, R., Moran M., Kochevar, I., Anderson, R., "Endogenous Skin Fluorescence Includes Bands that may Serve as Quantitative Markers of Aging and Photoaging," The Journal of Investigative Dermatology 111 776 780 (1998). Kollias, N., Zonios, G., Stamatas, G., "Fluorescen ce spectroscopy of skin," Vibrational Spectroscopy 28 17 23 (2002). guided 113 2198 2202 (2006). Koppel, 16 1315 1329 (1976). Comparative evaluation of differential laser induced perturbation spectroscopy as a technique to discriminate emerging 17 67 79 (2012).
170 Co 94 7514 7518 (1990). Lagant, P., Loucheux 22 1267 1283 (1983). Lagant, P., Loucheux Glycyl L 22 1285 1300 (1983). Lagant, P., Vergoten, G., Fleury, G., Loucheux vibratio 139 137 148 (1984). Lakowicz, J., Principles of Fluorescence Spectroscopy (Springer, New York, 2006). disease, 24 283 287 (1991). Lerouge, F., Ferrer Ugalde, A., Vinas, C., Teixidor, F., Sillanpaa, R., Abreu, A., Xochitiotzi, E., di and tetra 40 7541 7550 (2011). Raman study of 4 aminophenyl ammonium 30 451 458 (1997). Li, X., Eckard, J., Shah, R., Malluck, C., Frenkel, K., "Interleukin 1alpha Up Regulation In Vivo by a Potent Carcinogen 7,12 Dimethylbenz(a)anthracene (DMBA) and Control of DMBA induced Inflammatory Responses," Cancer Research 62 417 423 (2002). Liu, Q., Grant, G., Vo 16 927 940 (2010). Macaskill, A., Chernonosov, A., Koval, V., Lukyanets, E., Fedorova, O., Smith, W., Faulds, K., enhanced resonance Raman scattering of phthalocyanine 35 1 6 (2007). Mahadevan, A., Mitchell, M., Silva, E., Thomsen, S., Richards Kortum, R., "Study of t he fluorescence properties of normal and neoplastic human cervical tissue," Lasers in Surgery and Medicine 13 647 655 (1993). peptide aggregation by NMR and time reso Chemistry 103 2262 2269 (1999).
171 Martin, S., Wood, A., McRobbie, M., Mazilu, M., McDonald, M., Samuel, I., Herrington, C., "Fluorescence spectroscopy of an in vitro model of human cervical precancer identifies neoplastic phenotype," International Journal of Cancer 120 1964 1970 (2007). infrared solid state emitters based on isophorone: Synthesis, crystal structure and spectro scopic 23 862 873 (2011). d Spectroscopy 64 871 887 (2010). Meyer, A., Meyer, B., Morrison, J., Pepler, W., "Calcium, Collagen, Elastin and hexosamine levels in the arteries of whites and Bantu," South African Medical Journal 39 1017 1020 (1965). 39 851 862 (2006). Miomandre, F., LePicier, E., Munteanu, S., Galangau, O., Audibert, J., Meallet Renault, R., e and bodipy dyes using total internal reflection fluorescence microscopy coupled to 3 690 696 (2011). Mitchell, M., Cantor, S., Ramanujam, N., Tortolero Luna, G., Richards Kortum, R., "Fluorescence Spec troscopy for Diagnosis of Squamous Intraepithelial Lesions of the Cervix," Obstetrics & Gynecology 93 462 470 (1999). Moragas, A., Garcia Bonafe, M., Sans, M., Toran, N., Huguet, P., Martin Plata, C., "Image analysis of dermal collagen changes during skin aging," Analytical and Quantitative Cytology and Histology 20 493 499 (1998). Moreira, L., Silvira Jr., L., Santos, F., Lyon, J., Rocka, R., Zangaro, R., Villaverde, A., Pacheco, Spectroscopy: An International Journal 22 1 19 (2008). Spectroscopy Reviews 46 492 542 (2005). Gly cine under high pressures: a Raman 339 23 30 (2003). Nelson, M., Futscher, B., Kinsella, T., Wymer, J., Bowden, G., "Detection of mutant Ha ras genes in chemically initiated mouse skin epidermis before the development of benign tumors," Proceedings of the National Academy of Science 92 6398 6402 (1992). Niemz, M., Laser Tissue Interactions. Germany: Springer Verlag Berlin Heidelberg, 1996.
172 nment of Spectroscopy 56 146 153 (2011). Palmer, G., Zhu, C., Breslin, T., Xu, F., Gilchrist, K., Ramanujam, N., "Comparison of Multiexcitation Fluorescence and Di ffuse Reflectance Spectroscopy for the Diagnosis of Breast Cancer," Biomedical Engineering Transactions 50 1233 1242 (2003). cly glycine Pi Pi acid charge transfer transiti on moment orientations near resonance Raman single crystal 118 9716 9726 (1996). Pathak, I., Davis, N., Hsiang, Y., Quenville, N., Palcic, B., "Detection of squamous neoplasia by fluorescence imaging c omparing porfimer sodium fluorescence to tissue autofluorescence in the hamster cheek pouch model," American Journal of Surgery 170 423 426 (1995). Pedrotti, F., Pedrotti, L., Introduction to Optics. New Jersey: Prentice Hall Inc., 1987. Peiris, B., Budm 60 1385 1392 (2009). Peng, Q., Juzeniene, A., Chen, J., Svaasand, L., Warloe, T., Giercksky 71 1 28 (2008). protein aggregates in the cerebrospinal fluid of Alzheimer's patients by fluorescence correlation spectroscopy," Nature Medicine 4 832 834 (1998). Podstawka, E., Ozaki, Y., Pr enhanced Raman spectroscopy Spectroscopy 58 570 581 (2004). Pradhan, A., Pal, P., Durocher, G., Villeneuve, L., Balassy, A., Babai F., Gaboury, L., Blanchard, L., "Steady state and time resolved fluorescence properties of metastatic and non metastatic malignant cells from different species," Journal of Photochemistry and Photobiology B 31 101 112 (1995). Prasad, P., Introduction to Biophotonics. New Jersey: John Wiley and Sons, Inc., 2003. Quintanilla, M., Brown, K., Ramsden, M., Balmain, A., "Carcinogen specific mutation and amplification of Ha ras during mouse skin carcinogenesis," Nature 322 78 80 (1986). Ramanujam, N., "Fluores cence Spectroscopy of Neoplastic and Non neoplastic Tissues," Neoplasia 2 89 117 (2000).
173 Ramanujam, N., Follen Mitchell, M., Mahadevan Jansen, A., Thomsen, S., Staerkel, G., Malpica, A., Wright, T., Atkinson, N., Richards Kortum, R., "Cervical precancer d etection using multivariate statistical algorithm based on laser induced fluorescence spectra at multiple excitation wavelengths," Journal of Photochemistry and Photobiology 64 720 735 (1996). Rao J., Dragulescu Andrasi Current Opinion in Biotechnology 18 17 25 (2007). Ravery, V., Dominique, S., Panhard, X., Toublanc, M., Boccon Gibod, L., Boccon Gibod, L., "The 20 Core Prostate Biopsy Protocol -A New Gold Standard?," The Journal of Urology 179 504 507 (2008). R echberger, T., Postawski, K., Jakowicki, J., Gunja Smith, Z., Woessner, J., "Role of fascial collagen in stress urinary incontinence," American Journal of Obstetrics and Gynecology 179 1511 1514 (1998). ng protein mobility and activity in living 3 145 147 (2001). Richards Kortum, R., Mitchell, M., Ramanujam, N., Mahadevan, A., Thomsen, S., "In vivo fluorescence spectroscopy: potential for non invasive, automated diagnosis of c ervical intraepothelial neoplasia and use as a surrogate endpoint biomarker," Journal of Cellular Biochemistry 19 111 119 (1994). Roldan, M., Corrado, G., Francioso, O., Snachez herbicide paraquat analyzed by surface enhanced Raman scattering and fluorescence 699 87 95 (2011). laser fiber for neurosurgery: laser 112 434 443 (2010). urnal 88 4274 4288 (2005). Sieler, G., Schweitzer protonation states of dipeptides probed by polarized Raman UV Resonance Raman, and FTIR ry Part B 103 372 384 (1999). induced breakdown spectroscopy for biomedical 26 673 687 (2011). Smith, S., Buchanan, K., Hahn, D., Excimer laser induced perturbat ion spectroscopy: an approach to biosensing (SPIE, 2010).
174 Smith, S., Shanyfelt, L., Buchanan, K., Hahn, D., "Differential laser induced perturbation spectroscopy using a deep ultraviolet excimer laser," Optical Letters 36 2116 2118 (2011). Song, S., Xiong C., Zhou, M., Lu, W., Huang, Q., Ku, G., Zhaeo, J., Flores Jr., L., Ni, Y., Li, C., animal PET of tumor damage induced by photothermal ablation with 64Cu bis DOTA 52 792 799 (2011). Taubert, A., Napoli, A., assembly of reactive amphiphilic block copolymers as 8 598 603 (2004). Thumanu, K., Tanthanuch, W., Ye, D., Sangmalee, A., Lorthongpanich, C., Parnpai, R., Heraud, P ., "Spectroscopic signature of mouse embryonic stem cell derived hepatocytes using synchrotorn Fourier transform infrared microspectroscopy," Journal of Biomedical Optics 16 5 57 (2011). Tseng, S., Bargo, P., Durkin, A., Kollias, N., "Chromophore concentr ations, absorption and scattering properties of human skin in vivo," Optics Express 17 14599 14617 (2009). using solid state laser 213 and Excimer laser 193 nm: A randomized, contralateral, comparative, 49 1415 1420 (2008). Veld, D., Witjes, M., Sterenborg, H., Roodenburg, J., "The status of in vivo autofluorescence spectroscopy and imaging f or oral oncology," Oral oncology 41 117 131 (2005). resolution with light Schlieren 31 1812 1814 (2006). laser 93 4481 4500 (2007). Vogel, A., Noack, J., Huttman, G., Paltau 81 1015 1047 (2005). Chemical Reviews 103 577 644 (2003). Wang, C., Tsai, T., Chen, H., Chang, S., Chen, C., Chiang, C., "Autofluorescence spectroscopy for in vivo diagnosis of DMBA induced hamster buccal pouch pre cancers and cancers," Journal of Oral Pathology and Medicine 32 18 24 (2003). frared spectra and molar absorption coefficients of the 20 alpha amino acids in aqueous solutions in the spectral range from 1800 to 500 cm 1 Acta Part A 64 987 1001 (2005).
175 HILIC NMR: Towards the Science and Technology 45 3880 3886 (2011). enhanced Raman spectrum of Gly Gly adsorbed on 977 56 61 (2010). participation of basement membrane zone antibodies in the formation of the lup us band in 43 420 427 (2004). sics A 64 331 339 (1997). optical activity of L 99 835 843 (1995). Yu, G., Wu, M., Luo, emission Management 31 1729 1736 (2011). Zenklusen, J., Stockman, S., Fischer, S., Conti, C., Gimenez Cont i, I., "Transforming growth factor beta 1 expression in Syrian hamster cheek pouch carcinogenesis," Molecular Carcinogenesis 9 10 16 (1994). small yellow croaker (Pseudociaena polyactis) otoliths: an in situ Raman microspectroscopy and 390 777 782 (2008). investiga tion of solid samples using a low repetition rate pulsed Nd:YAG laser as the excitation 61 38 48 (2007). Zhu, C., Burnside, E., Sisney, G., Salkowski, L., Harter, J., Yu, B., Ramanujam, N., "Fluorescence Spectrosc opy: An Adjunct Diagnostic Tool to Image Guided Core Needle Biopsy of the Breast," Biomedical Engineering Transactions 56 2518 2528 (2009). Zhu, G., Zhu, Spectrochimica Acta Part A 78 1187 1195 (2011).
176 BIOGRAPHICAL SKETCH Sarah Elizabeth Smith was born in Martin Kentucky to James C. Smith and Jerri E. Smith She graduated f rom Morehead State University in Morehead Kentucky in May of 2007 with a Bachelor of Science in p hysics. Upon completion, she attended West Virginia University and graduat ed with a Master of Science in a e rospace e ngineering with thesis in 2009. The work pr esented here is in culmination of the research carried out during her PhD studies at the University of Florida.