BIOMEDICALVIBRATIONALSPECTROSCOPY

Edited By

Peter LaschJanina Kneipp

A JOHN WILEY & SONS, INC., PUBLICATION

InnodataFile Attachment9780470283165.jpg

BIOMEDICAL VIBRATIONALSPECTROSCOPY

BIOMEDICALVIBRATIONALSPECTROSCOPY

Edited By

Peter LaschJanina Kneipp

A JOHN WILEY & SONS, INC., PUBLICATION

Copyright � 2008 by John Wiley & Sons, Inc. All rights reserved

Published by John Wiley & Sons, Inc., Hoboken, New Jersey

Published simultaneously in Canada

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or byany means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permittedunder Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission ofthe Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center,Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department,John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online athttp://www.wiley.com/go/permission.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparingthis book, they make no representations or warranties with respect to the accuracy or completeness of the contentsof this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose.No warranty may be created or extended by sales representatives or written sales materials. The advice andstrategies contained herein may not be suitable for your situation. You should consult with a professionalwhere appropriate. Neither the publisher nor author shall be liable for any loss of profit or any othercommercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact ourCustomer Care Department within the United States at (800) 762-2974, outside the United States at(317) 572-3993 or fax (317) 572-4002.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not beavailable in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.

Library of Congress Cataloging-in-Publication Data:

Biomedical vibrational spectroscopy / edited by Peter Lasch, Janina Kneipp.p. ; cm.

Includes bibliographical references and index.ISBN 978-0-470-22945-3 (cloth)

1. Infrared spectroscopy. 2. Raman spectroscopy. I. Lasch, Peter. II. Kneipp, Janina.[DNLM: 1. Spectrophotometry, Infrared–trends. 2. SpectrumAnalysis, Raman. 3. Diagnostic Imaging–trends.

QC 454.R36 B6151 2008]QP519.9.I48B57 2008535.8’42–dc22 2007046854

Printed in the United States of America10 9 8 7 6 5 4 3 2 1

http://www.copyright.comhttp://www.copyright.comhttp://www.wiley.com/go/permissionhttp://www.wiley.com

v

CONTENTS

Preface xi

Contributors xiii

1 VIBRATIONAL SPECTROSCOPY IN MICROBIOLOGY ANDMEDICAL DIAGNOSTICS 1Dieter Naumann

1.1 Vibrational Spectra in Biomedicine Provide Fingerprint-likeSignatures of Biological Structures 2

1.2 Different Technical Options to Obtain the Spectral Information 31.3 The Need for and Benefit from Data Evaluation 41.4 Perspectives of Biomedical Vibrational Spectroscopy 5

2 BIOMEDICAL VIBRATIONAL SPECTROSCOPY –TECHNICAL ADVANCES 9H. Michael Heise

2.1 Introduction 92.2 Measurement Techniques for Clinical Chemistry 112.3 Measurement Techniques for Pathology 192.4 Measurement Techniques for In Vivo Spectroscopy 262.5 Concluding Remarks 31Acknowledgments 31References 32

3 BIOMEDICAL APPLICATIONS OF INFRARED MICROSPECTROSCOPYAND IMAGING BY VARIOUS MEANS 39David L. Wetzel

3.1 Introduction 393.2 Specimen Sources, Experimental Schemes, and Optical

Substrates 413.3 Applications 423.4 Instrumental Means of Biomedical IMS 593.5 Comment 71Acknowledgments 71Acronyms and Trademarks 72References 72

4 INFRARED SPECTROSCOPY OF BIOFLUIDS IN CLINICALCHEMISTRY AND MEDICAL DIAGNOSTICS 79R. Anthony Shaw, Sarah Low-Ying, Angela Man, Kan-Zhi Liu,

C. Mansfield, Christopher B. Rileg and Mouchanoh Vijarnsorn

4.1 Introduction 794.2 Vibrational Spectroscopy of Biofluids 804.3 Quantification (Regression) and Diagnostic (Classification)

Approaches 814.4 Quantitative Biofluid Analysis 824.5 Diagnostic Biofluid Tests 884.6 Veterinary Applications 924.7 Microfluidics and IR Spectroscopy of Biofluids 954.8 Concluding Remarks 99References 100

5 RAMAN SPECTROSCOPY OF BIOFLUIDS 105Daniel Rohleder and Wolfgang Petrich

5.1 Introduction 1055.2 Background Fluorescence 1065.3 The Putative Drawback of a Low Signal-to-Noise-Ratio 1095.4 Spectroscopy of Blood and Its Derivates 1115.5 In Vitro Raman Spectroscopy of Serum for Laboratory Diagnostics:

A Case Study 1125.6 Raman Spectroscopy of Body Fluids In Vivo 1155.7 Raman Spectroscopy of Other Body Fluids 1175.8 Summary 118Acknowledgments 118References 119

6 VIBRATIONAL MICROSPECTROSCOPY OF CELLS AND TISSUES 121Melissa J. Romeo, Susie Boydston-White, Christian Matth€aus,Milo�s Miljkovi�c, Benjamin Bird, Tatyana Chernenko and Max Diem

6.1 Introduction 1216.2 Infrared Histopathology: IR Microspectroscopic Mapping of Tissues 1226.3 Vibrational Cytology: IR and Raman Spectroscopy of Eukaryotic Cells 1336.4 Concluding Remarks 147Acknowledgments 148References 148

7 RESONANCE RAMAN MICROSPECTROSCOPY AND IMAGING OFHEMOPROTEINS IN SINGLE LEUKOCYTES 153Henk-Jan van Manen, Cynthia Morin, Cees Otto and Dirk Roos

7.1 Hemoproteins 1537.2 Raman Microspectroscopy 154

vi CONTENTS

7.3 Outline of This Chapter 1557.4 Instrumentation and Spectral Data Analysis 1567.5 Resonance Raman Microspectroscopy on Neutrophilic Granulocytes 1597.6 Resonance Raman Microscopy on Neutrophilic Granulocytes 1657.7 Photobleaching and Light-Induced Cell Damage in Resonance

Raman Microspectroscopy 1687.8 Concluding Remarks 172Acknowledgments 172References 172

8 RESONANT RAMAN SCATTERING OF HEME MOLECULES INCELLS AND IN THE SOLID STATE 181Bayden R. Wood and Don McNaughton

8.1 Introduction 1818.2 Electronic Structure of Heme Moieties 1828.3 Resonance Raman Spectroscopy 1848.4 Resonance Raman Spectroscopy of Hemes in Cells and the Solid State 1878.5 Resonance Raman of Heme Derivatives Using Near-Infrared

Excitation in the Solid State 1908.6 Application to Malaria Research 1978.7 Summary 203Acknowledgments 203References 203

9 COHERENT ANTI-STOKES RAMAN SCATTERING(CARS) MICROSCOPY 209Ondrej Burkacky and Andreas Zumbusch

9.1 Introduction 2099.2 Theoretical Considerations 2109.3 CARS Microscopy 2129.4 Suppression of the Nonresonant Background 2139.5 Applications to Biology 2179.6 Outlook 218Acknowledgments 219References 219

10 SURFACE-ENHANCED RAMAN SENSORS FORMETABOLIC ANALYTES 221Olga Lyandres, Matthew R. Glucksberg, Joseph T. Walsh Jr., Nilam C. Shah,

Chanda R. Yonzon, Xiaoyu Zhang and Richard P. Van Duyne

10.1 Background 22110.2 Experimental Setup 225

CONTENTS vi i

10.3 Results and Discussion 22810.4 Conclusion 236Acknowledgments 236References 237

11 SURFACE-ENHANCED RAMAN SCATTERING FOR INVESTIGATIONSOF EUKARYOTIC CELLS 243Janina Kneipp, Harald Kneipp, Katrin Kneipp, Margaret

McLaughlin and Dennis Brown

11.1 Motivation: SERS and Cell Studies 24311.2 Probing Intrinsic Cellular Chemistry 24511.3 SERS-Based Optical Labels for Live Cell Studies 25311.4 Conclusions and Outlook 256Acknowledgments 257References 257

12 COMBINING OPTICAL COHERENCE TOMOGRAPHY AND RAMANSPECTROSCOPY FOR INVESTIGATING DENTAL AND OTHERMINERALIZED TISSUES 263Lin-P0ing Choo-Smith, Alex C.-T. Ko, Mark Hewko, Dan P. Popescu,Jeri Friesen and Michael G. Sowa

12.1 Introduction 26312.2 Optical Coherence Tomography 26612.3 Raman Spectroscopy of Mineralized Tissues 27312.4 Towards Clinical Dental Relevance 28112.5 Conclusions: Our Multi Modal Approach for Evaluating Early

Dental Caries 285Acknowledgments 285References 286

13 SUB-100-NANOMETER INFRARED SPECTROSCOPYAND IMAGING BASED ON A NEAR-FIELD PHOTOTHERMALTECHNIQUE (‘‘PTIR’’) 291Alexandre Dazzi

13.1 Introduction 29113.2 AFMIR: Photothermal-Induced Resonance Experiment 29213.3 Experimental Illustration of the Photothermal Technique 29813.4 Applications: Biological Studies 30313.5 Conclusion and Perspectives 311Acknowledgments 311References 312

viii CONTENTS

14 FROM STUDY DESIGN TO DATA ANALYSIS 315Wolfgang Petrich

14.1 Aspects in the Design of Clinically Relevant Studies in BiomedicalVibrational Spectroscopy 316

14.2 The Role of Noise and Reproducibility in the Raw Spectra 32114.3 Safeguarding the Analysis of Data and Its Interpretation 32314.4 Conclusion 330Acknowledgments 331References 331

15 INTERPRETING SEVERAL TYPES OF MEASUREMENTSIN BIOSCIENCE 333Achim Kohler, Mohamed Hanafi, Dominique Bertrand,

El Mostafa Qannari, Astrid Oust Janbu, Trond Møretrø,

Kristine Naterstad and Harald Martens

15.1 Introduction to the Analysis of Several Data Sets 33315.2 Principal Component Analysis of One Data Table 33715.3 Simultaneous Analysis of Two Data Blocks by Partial Least-Squares

Regression (PLSR) 34215.4 Simultaneous Analysis of Several Data Blocks by Multiblock PCA 34715.5 Alternative Multiblock Methods 352References 354

16 INTERPLAY OF UNIVARIATE AND MULTIVARIATE ANALYSISIN VIBRATIONAL MICROSCOPIC IMAGING OF MINERALIZEDTISSUE AND SKIN 357Guojin Zhang, K. L. Andrew Chan, Carol R. Flach and Richard

Mendelsohn

16.1 Introduction 35716.2 IR Microscopic Characterization of an Unusual Form of Osteoporosis 35916.3 Applications to the Epidermis 36316.4 Concluding Remarks 376Acknowledgments 376References 376

INDEX 379

CONTENTS ix

xi

PREFACE

The interdisciplinary field of biomedical vibrational spectroscopy comprises a growing bodyof methods that support the development of practical applications inmicrobiology, cytology,histology, and clinical chemistry. This is not only due to the advantages inherent tovibrationalspectroscopic methods, but also a result of the spectacular technological progress seen in thelast 15 years. As rapid photonic techniques, infrared (IR) and Raman spectroscopy provideobjective information on molecular structure and composition of the samples underinvestigation. The ease of sample preparation and the speed of the measurement withcollection times in the range of seconds or minutes qualify both methods for the operator-independent, cost-efficient and nondestructive characterization of a sample’s biochemistry.Therefore, they offergreat promise for invivo and ex vivobiomedical diagnosis. Furthermore,the rapid development of both vibrational spectroscopic techniques has benefited consider-ably from the technological progress and scientific breakthroughs, in particular in the fieldsof light sources,multichannel detector technology, nanotechnology, and optics in general. Asinmany other technology-driven fields, these developments have been additionally triggeredby advances in computer science and information technology.

The contributions in this book provide an overview of state-of-the-art experimentalmethods and applications of IR and Raman spectroscopy in biomedicine. The first part ofthis volume contains chapters on established technical concepts and experimental ap-proaches and their applications in biomedical diagnostics and clinical chemistry. In anintroductory contribution, D. Naumann provides his view of the field and discusses thenature of the spectroscopic information, technical options, and the perspectives of vibra-tional spectroscopic methods in microbiology and biomedical diagnostics. The chapter byH. M. Heise reviews technical solutions of IR and Raman spectroscopic applications forclinical chemistry and pathology in vitro, in situ, and in vivo.

In vibrational spectroscopic studies of histological and cytological specimens, thecombination of spectroscopy with microscopy is particularly useful, because it enableslocalized biochemical characterization of cells or tissues. D. L. Wetzel discusses variousapplications of IR microspectroscopy and IR imaging and reviews important instrumentalmeans for their realization, such as ultra-bright synchrotron light sources and focal planearray detectors. R. A. Shaw et al. provide a chapter on the utilization of IR spectroscopy ofbiofluids in clinical chemistry and illustrate how the method can be employed for diseasediagnosis. The potential of Raman spectroscopy for the characterization of body fluidsex vivo and in vivo is demonstrated by D. Rohleder and W. Petrich.

The capabilities of Raman microspectroscopy for studies of cells and tissues was demo-nstratedmore than a decade ago.Meanwhile, owing to the progress in instrumentation and theavailability of high-quality commercial Raman microscopes, Raman spectroscopy-baseddiagnostic tools are being developed. In the chapter byM. J. Romeo et al., both IR andRamanmicrospectroscopy are employed to characterize cells and tissueswith high spatial resolution.

In the second part of the book, attractive new vibrational spectroscopic techniques withhigh potential for biomedical applications are presented. While some of these methods arestill in the phase of maturation, others demonstrate their immediate applicability to dia-gnostic problems or to the elucidation of pathophysiological mechanisms. The possibilities

of exciting Raman scattering in resonance with an electronic transition in the samplemolecule and the resulting signal enhancement are discussed in two chapters for theexample of heme groups in cells: H.-J. van Manen et al. introduce us to spectroscopy andspectral imaging of heme proteins in leukocytes and discuss experimental concepts andlimitations. A contribution by B. R. Wood and D. McNaughton reviews resonant Ramanspectroscopy in red blood cells and heme molecules in the solid state and its application inmalaria research. Improvements in the analytical sensitivity of the inherently inefficientRaman scattering process can also be achieved by coherent anti-Stokes Raman scattering(CARS). As demonstrated by O. Burkacky and A. Zumbusch, CARS has evolved into asensitive microscopic method that provides a great amount of chemical structure informa-tion from cells and other samples.

The favorable properties of localized surface plasmons and the utilization ofnanostructures supporting them are another means of improving both the Ramanscattering cross sections and the lateral resolution. The first can be employed to constructsensors for metabolites as is shown by O. Lyandres et al., who used surface-enhancedRaman scattering (SERS) for the detection of glucose, lactate, and other analytes fromplasma. The group employed multivariate analysis of SERS data for quantitativebiosensing in vivo. SERS microspectroscopic experiments at nanometer-scale lateralprecision in cells are reported by J. Kneipp et al., who studied the endosomal system ofcultured cells by this method.

Another direction of current research is the combination of differentmethods for opticaldiagnosis. In the chapter by L.–P. Choo–Smith et al., a combination of optical coherencetomography and Raman spectroscopy is demonstrated for the detection of caries. A numberof experimental methods have also been proposed to overcome the diffraction limit offar–field IRmicrospectroscopy. A. Dazzi explains in his contribution a photothermalmethodthat directly measures the expansion of a tiny sample due to IR absorption, and he illustratesits applicability for IR imaging of individual virus particles inside bacterial cells.

As the experimental tools for IR and Raman studies become established and new onesare developed, proofs of their usefulness in medical diagnostics are gaining more and moreimportance. Likewise, enormous amounts of spectral data require appropriate concepts andspecific tools for data analysis. In the third part of this book,we therefore discuss fundamentalaspects of study design and present adequate concepts for the analysis of vibrational spectraas multivariate data.W. Petrich has contributed a chapter that exemplifies how clinical studyconcepts can be realized in practice. It is alsodemonstrated howmultivariate spectral analysisis applied for quantification of analytes from body fluids and for disease pattern recognition(classification).A.Kohler,W.Martens, and co-workers present amultiblock analysismethodthat can be employed to analyze and interpret several data sets from one type of biologicalsample. Although multivariate methods proved very valuable for the analysis of vibrationalspectra, the strength of biomedical vibrational spectroscopy is greatly enhanced when theunivariate molecular structure information is incorporated into the mindset for data analysis.The interplay of univariate and multivariate concepts of spectral analysis is demonstrated inthe chapter byG. Zhang et al. These authors present examples of spectral imaging of skin andbone.

We are grateful to all authors who have shared their experience and knowledgein this book.

PETER LASCH

JANINA KNEIPP

Berlin, September 2007

xii PREFACE

xi i i

CONTRIBUTORS

Dominique Bertrand, Unit�e de Sensom�etrie et de Chimiom�etrie, ENITIAA/ INRA, BP82225, 44322 Nantes Cedex 3, France

Benjamin Bird, Department of Chemistry and Chemical Biology, NortheasternUniversity, Boston, Massachusetts 02115, USA

Susie Boydston-White, Department of Chemistry and Chemical Biology, NortheasternUniversity, Boston, Massachusetts 02115, USA

Dennis Brown, Program in Membrane Biology, Harvard Medical School, Boston,Massachusetts 02114, USA

Ondrej Burkacky, Institut für Physikalische Chemie, Ludwig-Maximilians-Universit€atM€unchen, D-80538 M€unchen, Germany

K. L. Andrew Chan, Department of Chemical Engineering, Imperial College London,London, SW7 2AZ, UK

Tatyana Chernenko, Department of Chemistry and Chemical Biology, NortheasternUniversity, Boston, Massachusetts 02115, USA

Lin-P0ing Choo-Smith, National Research Council Canada—Institute forBiodiagnostics, Winnipeg, Manitoba, Canada R3B 1Y6

Alexandre Dazzi, Laboratoire de Chimie Physique, Universit�e Paris—Sud, 91405Orsay Cedex, France

Max Diem, Department of Chemistry and Chemical Biology, Northeastern University,Boston, Massachusetts 02115, USA

Carol R. Flach, Department of Chemistry, Newark College of Arts and Sciences, RutgersUniversity, Newark, New Jersey 07102, USA

Jeri Friesen, National Research Council Canada—Institute for Biodiagnostics, Winni-peg, Manitoba, Canada R3B 1Y6

Matthew R. Glucksberg, Biomedical Engineering Department, NorthwesternUniversity, Evanston, Illinois 60208, USA

Mohamed Hanafi, Unit�e de Sensom�etrie et de Chimiom�etrie, ENITIAA/INRA, BP82225, 44322 Nantes Cedex 3, France

H. Michael Heise, ISAS—Institute for Analytical Sciences at the Technical University ofDortmund, 44139 Dortmund, Germany

Mark Hewko, National Research Council Canada—Institute for Biodiagnostics,Winnipeg, Manitoba, Canada R3B 1Y6

Astrid Oust Janbu, Aquateam AS, Norwegian Water Technology Centre, Postbox 6875Rodeløkka, 0504 Oslo, Norway

Harald Kneipp, Wellman Center for Photomedicine, Harvard Medical School, Boston,Massachusetts 02114, USA

Janina Kneipp, Federal Institute for Materials Research and Testing, Berlin, Germany;and Wellman Center for Photomedicine, Harvard Medical School, Boston, Massa-chusetts 02114, USA

Katrin Kneipp, Wellman Center for Photomedicine, Harvard Medical School, Boston,Massachusetts 02114, USA

Alex C.-T. Ko, National Research Council Canada—Institute for Biodiagnostics,Winnipeg, Manitoba, Canada R3B 1Y6

Achim Kohler, Center for Biospectroscopy and Data Modelling, Matforsk, NorwegianFood Research Institute, 1430 A

�s, Norway; and CIGENE, Department of Mathe-

matical Sciences and Technology, Norwegian University of Life Sciences, 1430 A�s,

Norway

Kan-Zhi Liu, National Research Council of Canada, Institute for Biodiagnostics, Win-nipeg, Manitoba, Canada R3B 1Y6

Sarah Low-Ying, National Research Council of Canada, Institute for Biodiagnostics,Winnipeg, Manitoba, Canada R3B 1Y6

Olga Lyandres, Biomedical Engineering Department, Northwestern University,Evanston, Illinois 60208, USA

Angela Man, National Research Council of Canada, Institute for Biodiagnostics,Winnipeg, Manitoba, Canada R3B 1Y6

Colin D. Mansfield, NRC Institute for Biodiagnostics, Winnipeg, Manitoba, CanadaR3B 1Y6. Present address: L’Institut des Nanotechnologies de Lyon (INL), E�coleCentrale de Lyon, 36 �Ecully, France

HaraldMartens, Center for Biospectroscopy and Data Modelling, Matforsk, NorwegianFood Research Institute, 1430 A

�s, Norway; CIGENE, IKBM/UMB, Norwegian

University of Life Sciences, 1430 A�s, Norway; and Faculty of Life Sciences, University

of Copenhagen, DK 1958, Frederiksberg, Denmark

Christian Matth€aus, Department of Chemistry and Chemical Biology, NortheasternUniversity, Boston, Massachusetts 02115, USA

Margaret McLaughlin, Program in Membrane Biology, Harvard MedicalSchool, Boston, Massachusetts 02114, USA

DonMcNaughton, Centre for Biospectroscopy and School of Chemistry, 3800 Victoria,Australia

Richard Mendelsohn, Department of Chemistry, Newark College of Arts and Sciences,Rutgers University, Newark, New Jersey 07102, USA

Milo�s Miljkovi�c, Department of Chemistry and Chemical Biology, Northeastern Uni-versity, Boston, Massachusetts 02115, USA

Trond Møretrø, Matforsk, Norwegian Food Research Institute, 1430 A�s, Norway

Cynthia Morin, Biophysical Engineering Group, Institute for Biomedical Technology,MESAþ Institute for Nanotechnology, University of Twente, 7500 AE Enschede, TheNetherlands. Present address: Materials Science and Technology of Polymers Group,MESAþ Institute for Nanotechnology, University of Twente, 7500 AE Enschede, TheNetherlands.

Kristine Naterstad, Matforsk, Norwegian Food Research Institute, 1430 A�s, Norway

Dieter Naumann, Robert Koch-Institut, D-13353 Berlin, Germany

xiv CONTRIBUTORS

Cees Otto, Biophysical Engineering Group, Institute for Biomedical Technology,MESAþ Institute for Nanotechnology, University of Twente, 7500 AE Enschede, TheNetherlands

Wolfgang Petrich, Department of Physics and Astronomy, University of Heidelberg,D-69120 Heidelberg, Germany; and Roche Diagnostics GmbH, 68305 Mannheim,Germany

Dan P. Popescu, National Research Council Canada—Institute for Biodiagnostics,Winnipeg, Manitoba, Canada R3B 1Y6

El Mostafa Quannari, Unit�e de Sensom�etrie et de Chimiom�etrie, ENITIAA/INRA, BP82225, 44322 Nantes Cedex 3, France

Christopher B. Rileg, Department of Health Management, Atlantic VeterinaryCollege, University of Prince Edward Island, Charlottetown, PEI, CanadaC1A 4P3

Daniel Rohleder, DIOPTIC GmbH, 69469 Weinheim, Germany

Melissa J. Romeo, Department of Chemistry and Chemical Biology, NortheasternUniversity, Boston, Massachusetts 02115, USA

Dirk Roos, Department of Blood Cell Research, Sanquin Research, and LandsteinerLaboratory, Academic Medical Centre, University of Amsterdam, 1066 CXAmsterdam, The Netherlands

Nilam C. Shah, Department of Chemistry, Northwestern University, Evanston, Illinois60208, USA

R. Anthony Shaw, National Research Council of Canada, Institute for Biodiagnostics,Winnipeg, Manitoba, Canada R3B 1Y6

Michael G. Sowa, National Research Council Canada—Institute for Biodiagnostics,Winnipeg, Manitoba, Canada R3B 1Y6

Henk-Jan van Manen, Biophysical Engineering Group, Institute for BiomedicalTechnology, MESAþ Institute for Nanotechnology, University of Twente, 7500 AEEnschede, The Netherlands. Present address: Akzo Nobel Research and TechnologyCenter, Department of Analytics and Physics, Molecular Spectroscopy Group, Vel-perweg 76, P.O. Box 9300, 6800 SB Arnhem, The Netherlands

Richard P. Van Duyne, Department of Chemistry, Northwestern University, Evanston,Illinois 60208, USA

Mouchanoh Vijarnsorn, Department of Health Management, Atlantic VeterinaryCollege, University of Prince Edward Island, Charlottetown, PEI, Canada C1A 4P3.Present address: Department of Companion Animal Clinical Science, Faculty ofVeterinary Medicine, Kasetsart University, Bangkok, Thailand

Joseph T. Walsh Jr., Biomedical Engineering Department, Northwestern University,Evanston, Illinois 60208, USA

David L. Wetzel, Microbeam Molecular Spectroscopy Laboratory, Kansas StateUniversity, Manhattan, Kansas 66506, USA

Bayden R. Wood, Centre for Biospectroscopy and School of Chemistry, 3800 Victoria,Australia

Chanda R. Yonzon, Department of Chemistry, Northwestern University, Evanston,Illinois 60208, USA

CONTR IBUTORS xv

Guojin Zhang, Department of Chemistry, Newark College of Arts and Sciences, RutgersUniversity, Newark, New Jersey 07102, USA

Xiaoyu Zhang, Department of Chemistry, Northwestern University, Evanston, Illinois60208, USA

Andreas Zumbusch, Universit€at Konstanz, 78457 Konstanz, Germany

xvi CONTRIBUTORS

1

VIBRATIONAL SPECTROSCOPY INMICROBIOLOGY AND MEDICAL

DIAGNOSTICSDieter Naumann

Robert-Koch Institut, Berlin, Germany

Infrared (IR) and Raman spectroscopy are relatively old spectroscopic modalities thatprovide pictures of the molecular vibrations performed by molecules. Since the earlyexperiments of Herschel, who more than 200 years ago discovered heat transportingradiation beyond the range of visible light, it took some 80 years until the first IR spectrumof an organic liquid was obtained. Since then, IR spectroscopy developed into the“workhorse” of vibrational spectroscopy in fundamental science and the industries, whileRaman spectroscopy, discovered only in 1928,was initially restricted to a few laboratories inthe academic area. Infrared and Raman spectroscopy, though fundamentally different inexperimental design and physical background, give complementary information on mo-lecular vibrations and should ideally be used together to attain access to the totality of allvibrational modes of a given molecule.

It has been only for the last two or three decades that both types of vibrationalspectroscopy have been used systematically for the more complex building blocks ofbiological systems or even intact cells, tissues, and biological fluids. These scientificendeavors were facilitated by technological innovations such as the advent of Fouriertransform (FT)-IR spectrometers, powerful low-cost lasers in the near-IR region, sensitivedetector systems, and rapid low-cost computers, which favored new developments such asfocal plane array detectors for true IR imaging systems or surface-enhanced Ramantechniques based on nanostructured materials as optically active elements.

The progress achieved and the practical applications realized until now have definitelydisproved the notion that IR orRaman spectroscopy are “old-fashioned technologies” usefulonly for pure systems and relatively small molecules. It has been convincingly proven that

1

Biomedical Vibrational Spectroscopy, Edited by Peter Lasch and Janina KneippCopyright � 2008 John Wiley & Sons, Inc.

IR and Raman spectra of cells, tissues, or biofluids encode sufficient spectral information todistinguish between different cell types, tissue structures, and biofluids and even to detectchanges in these biological materials induced by pathological processes.

1.1 VIBRATIONAL SPECTRA IN BIOMEDICINE PROVIDEFINGERPRINT-LIKE SIGNATURES OF BIOLOGICAL STRUCTURES

A rationale behind the belief that vibrational spectroscopy may be useful to diagnosediseases or pathologies in individuals is that disease processes must, generally speaking, beaccompanied by changes in the chemistry/biochemistry of cells, tissues, organs, or bodyfluids, and vibrational spectroscopy is indeed ideally suited for sensitive detection of suchchanges as a diagnostic technique. It has furthermore been anticipated that these changesshould be detectable also before morphological and systemic manifestation allow clinicaldiagnosis by conventional methods. Given the fact that sample preparation and measure-ment are very simple and collection times are in the range of seconds or minutes, IR andRaman spectroscopy should be ideal modalities to establish very rapid nonsubjective andcost-effective tools for early diagnosis of disease processes in individuals.

Biomedical IR and Raman spectroscopy probe biological samples in a way that theactive vibrational modes of all constituents present in the mixture are observed in a singleexperiment, resulting in very complex spectra with broad and superimposed spectralfeatures throughout the whole spectral range. Thus, in contrast to fluorescence spectra,obtained from a biological material labeled with some fluorescing dye, common IR andRaman spectra of intact cells, tissues, or body fluids cannot provide information on a singleor even a few specific compounds present. Instead, the spectra provide spectroscopicfingerprints of the total chemical and biochemical composition of the material under study.This situation inevitably results from the fact that the complex superposition of thecharacteristic IR absorption or Raman signals of all constituents in biomaterials (nucleicacids, proteins, carbohydrates, lipids, and other low molecular compounds, etc.) areobserved simultaneously, thereby producing spectral features that encode a vast amountof information potentially useful for biodiagnostic purposes. One peculiarity of vibrationalspectroscopy is that it provides information not only on the composition of complexbiological material but also on structural states of the molecules under study, since certainbands are sensitive, for example, to the secondary structure in proteins, while others reporton the state of order of the membranes or the conformation of the nucleic acid structures. Inthis sense the total information content in vibrational spectra of biological materials isenormous. One can possibly say that there are presently no other techniques available thatcan provide such a huge amount of information in one single experiment. On the other hand,this fact severely limits assignments of experimentally observed bands to single discretestructures and qualifies the techniques mainly as fingerprinting methods, though theassignment of spectral bands has improved significantly in the last two centuries due to,for example, spectral resolution enhancement and “spectral feature extraction” capabilitiesthat allow us to more efficiently visualize and resolve specific, hidden bands from thecomplex spectral signatures.

The nature of information obtained in biomedical vibrational spectroscopy is repre-sented best by the notion of “spectral fingerprints.” Thus, the analysis of these spectralsignatures by evaluating peak intensities, frequencies, or half-widths of a few bands that canby some means be resolved will fail in most cases. Moreover, taking into account thatthousands of spectra have to be analyzed at a given time, the availability of intelligent data

2 VIBRATIONAL SPECTROSCOPY IN MICROBIOLOGY AND MEDICAL DIAGNOSTICS

evaluation concepts is a virtual necessity that should ideally include efficient datapretreatment algorithms such as quality testing, normalization, filtering, and adequatemultivariate statistical techniques to achieve data reduction and finally the classification ofpatterns. With such methods, even hundreds of thousands of spectra – as is the case inspectroscopic imaging – can be analyzed.

Vibrational spectra of cells, tissues, and biofluids are obviously the expression of thesum of cellular chemistry/biochemistry and structure. Therefore they provide an“OMICS”-like view of the total chemical/biochemical status of the samples and give asnapshot on cell division, differentiation, growth and metabolism. In this view, vibrationalspectroscopic techniques provide information on phenotypes andmirror transcriptional andtranslational up- and down-regulation processes and post-translational modifications. In astrict sense, vibrational spectroscopies as applied to biofluids, cells, or tissues are, however,not typical metabolomic techniques. Their advantage is possibly that in some way thetotality of all chemical/biochemical changes including those in the pool of nucleic acids,proteins, or low molecular metabolic compounds are reflected in the spectra, constituting atechnique that cannot easily be assigned to one of the known “OMICS” disciplines in lifescience such as genomics, transcriptomics, proteomics, or metabolomics. But, as do thecommon “OMICS” methods, they deal with complex systems in their entirety and with thesimultaneous analysis ofmany biological individuals or objects rather than a single propertyof a single gene or metabolic product. In many cases the situation might be similar to globalmetabolic fingerprinting, but one has to bear inmind that the basis of changes observed doesnot necessarily have to be purely metabolic. This definition qualifies vibrational spectro-scopies as explorative and rapid analysis techniques par excellence, which can be used todiagnose disease or dysfunction via spectral biomarkers that change as indicators of thepresence of a particular disease or in response to drug intervention, environmental stress, orgenetic modification. When nothing or little is known about an observed phenomenon,vibrational spectroscopy may provide a first hint for further, possibly more specificinvestigations. This is particularly the case when changing systems, whether it is a cellsuspension of synchronized cells or cells treated with some specific drug are measuredtime-dependently. Such experiments can, however, be done with vibrational spectroscopictechniques in a few minutes compared to serial measurements using, for example,fluorescence labels, testing many genes or separating and analyzing proteins or metabolitesfrom complex mixtures. Therefore, the fundamental fingerprinting nature of vibrationalspectra of complex biological samples is a big advantage. It is, however, a disadvantage atthe same time, since comprehensive understanding of these spectra is desirable but notachievable in most cases.

1.2 DIFFERENT TECHNICAL OPTIONS TO OBTAINTHE SPECTRAL INFORMATION

Themost important step forward in biomedical vibrational spectroscopy within the last twodecades is certainly the coupling of spectrometers to light microscopes to obtain spectralinformation from single cells or to achieve spatial resolution in tissue analysis in a way thatis familiar to biologists or pathologists. Since then the technological progress has beenenormous and high-quality IR and Raman microscopes are available on the market, whichcan be used to image tissues and single cells and even analyze subcellular compartments.Raman imaging systems that do not rely on spectral point-by-point mapping are not yet onthe market, thus precluding Raman imaging under clinical constrains. Notwithstanding,

DIFFERENT TECHNICAL OPT IONS TO OBTAIN THE SPECTRAL INFORMATION 3

tissue or subcellular imaging by various different Raman microspectroscopic modalitiesprovides a wealth of biological information not available by other techniques. Today, focalplane array detectors formid-IR imaging allow rapid segmentation of histological structureswithout any tissue staining and to image larger cells. Using focal plane array systems,pioneering applications have been published on IR imaging of various soft and hard tissuesand a vast number of pathologies. Infrared synchrotron radiation sources coupled with IRmicroscopes allowed the analysis of single living cells growing in culture with unprece-dented high signal-to-noise ratio and reproducibility, opening up the possibility to performstrict difference spectroscopic investigations on viable cells – for example, after treatmentwith drugs or other chemicals.

Other technical developments such as fiber-optic probes have dramatically increasedthe possibilities to use Raman spectroscopy as a diagnostic biomedical tool. Fiber-opticapplications useful for in vivo applications have made greatest progress in Ramanspectroscopy, since the production of Raman compatible fiber probes can be based onmaterials already developed for fiber-based telecommunications or fiber-based chemicalsensors. Compared to this situation, optical halide fibers necessary for mid-IR spectroscopyare only available for a few laboratories apart from the detrimental fact that IR radiation hastoo small penetration depths and problems with strong water absorptions to be useful for invivo experiments.

SERS is a very sensitive Raman modality that can detect and characterize extremelysmall amounts of nucleic acids, proteins, or virus particles and can also characterizebiomolecular events in subcellular compartments. The attractiveness of SERS relies ondetection limits close to immunoassay sensitivities with femtomolar detection of, forexample, prostate-specific antigen. Tip-enhanced Raman spectroscopy (TERS), anotherSERS modality, combines SERS spectroscopy with scanning probe technologies andprovides lateral resolutions of around 20 nm and thus provides the possibility to study thesurface chemistry and structure or composition of cell membranes and cell walls.

Many scientists have realized that IR spectroscopy has great potentials as a finger-printing technique, useful for the very rapid diagnosis of disease or dysfunction in humansand animals with high-throughput screening capabilities. At present, however, IR andRaman spectroscopy seem to be best developed in microbiology and clinical chemistry, andfirst dedicated systems for use under practical conditions are already on themarket; also, thedevelopment of vibrational spectroscopy based diagnostics for in vivo glucose screening isnear to practical translation. It has also been recognized that vibrational spectroscopies aresimple and economical techniques to screen for changes in cells or body fluids in response todrug-based intervention, environmental stress, or genetic modifications in organisms. Theresults obtainedwith bone, cartilage, and dental tissues are impressive, and the possibility ofpractical applications developed for clinical or other medical settings seems to be obvious.The FT-IR imaging data obtained on colon, prostate, or brain cancer are also significant andcould be good candidates for translation to routine applications using benchtop IR imagingsystem as the technical platform.

1.3 THE NEED FOR AND BENEFIT FROM DATA EVALUATION

The necessity to use multivariate pattern recognition methodologies when dealing withspectral data of complex biomedical materials has been realized by the spectroscopiccommunity more than 20 years ago. Among the first who recognized this problemwere scientists working with IR spectroscopic data of intact microorganisms. While

4 VIBRATIONAL SPECTROSCOPY IN MICROBIOLOGY AND MEDICAL DIAGNOSTICS

univariate statistical analysis considered only a single property of a given selectionof microbial species (e.g., a single intensity or frequency value at a given wavenumberor peak), multivariate statistical methods allowed the evaluation of several, if not all,properties of the spectra at the same time. Only in this way the interrelations between thesample properties and the spectra could be figured out. This learning process has beenfacilitated at that early time by the need to handle thousands of measurements on hundredsof different microbial species and strains, to evaluate these data systematically for spectralsimilarity, and to exchange data between different laboratories.

Out of the large number of pattern recognition techniques that are presently used for, orhave been adapted to, vibrational spectroscopic data, factor analysis techniques likeprincipal component analysis (PCA) and hierarchical clustering analysis (HCA) or classifi-cation methodologies such as artificial neural nets (ANN), support vector machines (SVM),and linear discriminant analysis (LDA) have experienced broad acceptance. Factor analysisis frequently used to achieve data reduction and the classification of patterns in large datasets, and hierarchical clustering (a so-called unsupervised or data-driven classificationmethod) also attempts to find intrinsic similarity structures within the data sets without theneed for any a priori class assignment, while ANN analysis as a supervised or concept--driven classification technique needs the class assignment of each individual object from thebeginning. Partitioning of thewhole data set into a so-called training and internal validationdata subset is needed to train the system for optimal performance. It took some years by thespectroscopic community to learn that only independent data sets from ideally blindedsamples should be used to objectively test the performance and robustness of the classifierand to evaluate the accuracy of the established models.

Meanwhile, nearly the whole arsenal of multivariate bioinformatic techniques is used,and multivariate statistical analysis of spectroscopic data constitutes an own disciplinewithin the scientific area of biomedical spectroscopy. As for any other scientific discipline,these methods not only can be used to evaluate given data sets, but also allow completelynew problem solutions to be addressed. New applications arose, for example, when it wasrealized that determining the covariance between different large data matrices obtainedfrom the same sample populations with fundamentally different techniques is not only achallenge per se, but also provides insight into the interlink between biological structures.One of these new applications recently published was the use of genetic algorithms incombinationwith partial least-square regression (PLSR) analysis to correlate genes selectedfrom gene expression profiles obtained by microarray technologies to metabolic markersfrom spectral data setsmeasured from the same samples by IR spectroscopy. The analysis ofcovariance patterns in these very complex mixed data sets helped to rapidly recognize andvisualize the interrelationships and trends in a developing and changing biological systemthat is not easily achieved by any other means.

1.4 PERSPECTIVES OF BIOMEDICAL VIBRATIONAL SPECTROSCOPY

Despite all the fascinating potential and technological developments and the vast amount ofexciting research papers in the literature, progress toward factual translation of vibrationalspectroscopic techniques to practical applications is less evident. Moreover, the presentsituation of a multiplicity of different vibrational spectroscopic modalities, which areviewed by the nonspecialists as competing technologies, is possibly confusing.

The use of IR andRaman spectroscopy formicrobial characterization and identificationis presently the best developed and most frequent application of biomedical vibrational

PERSPECT IVES OF B IOMEDICAL V IBRAT IONAL SPECTROSCOPY 5

spectroscopy. It is especially remarkable that both spectroscopies are applied in microbio-logical laboratories not only for research purposes but also for routine analysis, for example,in the food industry for microbiological quality control to guide adequate productionmeasures. This situation has greatly been promoted by dedicated high-throughput IR andRaman instrumentation available now on the market. New avenues of microbiologicalapplications can be expected from the use of IRorRamanmicroscopes, whether it will be for(a) the microspectroscopic analysis of microcolonies to speed up identification of micro-organisms and analyze mixed populations of cells or (b) the identification of single cellsdirectly from environmental samples. The combination of IR focal plane array detectors andmicroarray printing technologies may contribute to make microbiological IR analysis anextremely rapid, cost-effective and unprecedented high-throughput technology for micro-biological analyses. This technology may not only help to scale down the number of cellsneeded for analysis, to investigate mixed cultures, and to perform population analyses, butalso help to detect light-microscopic and spectroscopic features simultaneously, with theprospect of a fully automated IR microscopic system combining detection, enumeration,and identification of microorganisms in one single instrument. One particular aspect ofvibrational spectroscopy in microbiology which constitutes its attractiveness is the possi-bility to achieve subspecies differentiation and the ability to analyze all kind of cells that canbe grown in culture. No other technique is currently available that can trace microbiologicalcontaminations in food microbiology or perform epidemiological investigation in clinicalsettings similarly quickly and easily. It is interesting to note that this potential is currentlyevaluated in several laboratories and that dedicated instrumentations are being designed formicrobial subspecies differentiations in collaboration with industrial partners.

It is the author’s personal belief that best perspectives for practical applications willarise in those fields where the various vibrational spectroscopic modalities are used as“coupled” techniques – for example, in the form of spectroscopy and microscopy, micro-spectroscopy and nanoparticles, spectroscopy and optical fibers, or spectroscopy and opticaltweezers. In the case ofmicrobiology, to give an example, this will not only allow us to scaledown the number of cells needed for analysis to a few or even only a single cell to perform,for example, population analyses in complex habitats, but also allow to detect lightmicroscopic and spectroscopic features of cells simultaneously, which is impossible forother techniques presently in use.

Immense future applications in cell biology, virology, and microbiology mayarise from the use of Raman spectroscopy with optical tweezers. Raman tweezers is arelatively new technology that couples Raman spectroscopy with optical tweezers that arealready routinely used for the noninvasive manipulation of biological particles to achievepreviously impossible sample control. This combination represents a new category ofapplication and may become a modality for flow cytometry to identify cells on the basis ofintrinsic molecular properties instead of the particles’ size, shape, or fluorescence.

Noninvasive methods to image single live cells are resonance Raman scattering (RRS)and coherent anti-Stokes Raman scattering (CARS) microscopy, which provide intrinsicmolecular-vibration-based contrast with a sensitivity that is orders ofmagnitude higher thanconventional Raman microscopy. CARS technology has recently been used to track lipidmetabolism in live cells and may become a significant tool in environmental and medicalmicrobiology.

SERS will most probably gain greatest attention reaching far beyond the relativelysmall community of vibrational spectroscopists, since itmay provide biological informationthat is not available by any other means. SERS used with biocompatible gold nanoparticlesincorporated as sensors by cells holds great promise to sensitively and specifically test

6 VIBRATIONAL SPECTROSCOPY IN MICROBIOLOGY AND MEDICAL DIAGNOSTICS

molecules in selected subcellular compartments in femtoliter-scaled volumes. This Ramanspectroscopic modality could greatly benefit from the fact that defined SERS-activenanoparticles are routinely available and already used along with fluorescence techniquesor electron microscopy in cell biology.

The development of technologies for subwavelength spectroscopy of cells and tissues ispresently a major point of interest, and different approaches are being evaluated by severalgroups. The coupling of atomic force microscopy (AFM) with SERS, the so-calledtip-enhanced Raman spectroscopy (TERS), seems to be very promising. The possibilityto obtain compositional and structural information at a nanoscale level is the most attractiveaspect of this new methodology and could provoke as much attention as AFM did some20 years ago. Also, the coupling of IR lasers with AFM technology, which can probe in aphotothermal deflection near-field experiment the local transient deformation induced by anIR pulsed laser tuned to different absorbing wavelengths, may be developed into amicroscopic technique that yields chemical contrast at lateral resolutions not accessibleby any IR far-field optical technique.

The use of Raman fiber-optic probes may open new avenues for routine in vivo use inclinical settings, since the high specificity of Raman spectroscopy can be combinedwith thepossibility of immediate visualization. For practical applications, such fibers will mostreasonably be used in multimodal fashion with other optical techniques such as lightscattering, optical coherence tomography, or fluorescence spectroscopy, since wide-fieldRaman imaging still needs to be developed. Further technological progress will benecessary, because fiber-optic technologies are not routinely compatible with existingendoscopic technologies and because of fundamental physical limitations. Though notechnical advances are in sight that could allow retrieval of spectra from several centimetersbelow the tissue surface, very efficient in vivo skin analyses based on confocal Ramanspectroscopy are already on the market and in practical use.

Bench-top instrumentation for routine IR imaging of diseased tissue sections isavailable. The vast amount of applications so far published clearly prove that segmenta-tion of histological structures is possible without any staining, and the identification ofcancerous lesions within tissues may be achieved in an objective way using extensivereference data bases. Possibly, the xth publication of data showing that vibrationalspectroscopic imaging can identify pathologies in tissues is not only lacking noveltyhereafter, but even counterproductive. To push biomedical vibrational spectroscopyforward, multicenter clinical trials focusing on selected clinical indications are neededto attract the attention of the clinicians and to establish sensitivity and specificityparameters under practical constraints. At present, however, the following questionsremain: Who could conduct such trials? Which relevant cancer types or clinical samples(fresh patient biopsies or archive material) should be used? Which technological plat-forms should be used?

The use of vibrational spectroscopy together with accepted genomic or metabolomicmethods such as DNA/RNAmicroarrays or mass spectroscopies can be of profit when datasets obtained by fundamentally different experimental techniques from the same selectionof samples are combined to analyze the covariance patterns in these complex data blocks.The combined analysis, for example, of gene expression and biomolecular response data toexternal stress factors in microorganisms would help to close the gap between differentdisciplines, since they can inherently only be done in cooperation between groups that areable to professionally deal with complex technologies. The results of such joint effortswould immediately be recognized by a much broader range of scientists and potential usersof the new vibrational spectroscopic techniques.

PERSPECT IVES OF B IOMEDICAL V IBRAT IONAL SPECTROSCOPY 7

Obviously, no killer application has been found yet that could pave the way forfurther steps forward and that cannot be done with any other type of technology. Althoughvibrational spectroscopy may be superior to competing methods in some cases, no majorapplication could be found to date that can be done in no other way or which is so muchsuperior to replace present technologies in practical use. The scientific community inbiomedical vibrational spectroscopy is perhaps at a turning point where practical applica-tions must arise. It will probably not be easy to invest such a high amount of enthusiasm,money, and time for another 10 or 15 years. Indeed, it will instead become more difficult toattract funding for this scientific field, unless significant progress will be made in thetransfer of basic science to important practical applications accepted by biologists orclinicians. The gap between enthusiasm and optimism on the one side and the necessity tosignificantly contribute to the present practical needs of the medical or biological commu-nity on the other side must be closed. It should also be clear that series of nice publicationswill not be enough to close this gap. What must be paramount are joint efforts that combineexperience, manpower, and budgets of several groups to bring selected applications topractical applications and patents to the industries.

A similarly important point is the necessity to define standards to exchange data andcompare reproducibility levels between the groups and to establish criteria, for example,how sensitivity and specificity values are determined for objective evaluation of spectraldata. Without the definition of standards, protocols, and quality control measures, the valueof large amounts of data will be rapidly lost after completion of the primary research andincrease the probability of reinventing the wheel. This will be critical for the successfuldevelopment and maturation of an emerging technology like vibrational spectroscopy.

8 VIBRATIONAL SPECTROSCOPY IN MICROBIOLOGY AND MEDICAL DIAGNOSTICS

2

BIOMEDICAL VIBRATIONALSPECTROSCOPY – TECHNICAL

ADVANCESH. Michael Heise

ISAS—Institute for Analytical Sciences at the Technical University of Dortmund, Germany

2.1 INTRODUCTION

In recent years, vibrational spectroscopy has been extremely successful and versatile forcondensed and gaseous phase analysis due to a plethora of measurement techniques andmore affordable spectrometers; and still many growing areas can be listed, for whichbiomedical applications are published. The spectral range covers the short-wave near-infrared (NIR) down to the far-infrared. A few instrumental aspects will only be mentionedin the introduction, but relevant references are provided, enabling the reader to familiarizehimself with those areas through the literature cited.

The lowest frequency range has recently attractedmany researchers when the so-calledterahertz radiation, spanning the spectral interval between the microwave and the infrared(IR) region of the electromagnetic spectrum, found new rapidly expanding applications inbiology and biomedicine. In particular, the spectroscopy of compounds such as proteins,enzymes, biological membranes, or whole cells has been carried out using laboratory-scaleterahertz sources.Water absorption dominates spectroscopy and imaging of soft tissues, butthe technology could play a role in diagnosing skin diseases. Despite this, there areadvantages of terahertz methods that make it attractive for pharmaceutical and clinicalapplications. Besides low-frequency bond vibrations, also hydrogen-bonding stretches,torsions, and crystalline phononvibrations can be assigned to this spectral range, interestingenough for crystalline conformation and polymorphism studies; see also the review byPickwell and Wallace.1 Most applications use terahertz radiation generated by short-pulsesolid-state lasers.

9

Biomedical Vibrational Spectroscopy, Edited by Peter Lasch and Janina KneippCopyright � 2008 John Wiley & Sons, Inc.

Other lasers have been mandatory for Raman spectroscopy, and the use of sensitivecharge-coupled device (CCD) detectors has made dispersive Raman spectral acquisi-tion much more rapid. Such Raman spectrometers typically use holographic diffractiongratings and efficient edge or notch filters due to advances in thin-film technology toachieve a high degree of laser rejection. For biological and medical samples for whichproblems with fluorescence exist, Fourier transform (FT)–Raman techniques usingNIR lasers (785 nm diodes, 1064 nm Nd:YAG) have been routinely applied, althoughalso dispersive multichannel instrumentation is in use, even for 1064 nm excitation.Other lasers are often used when enhanced Raman signals are to be observed frommicroscopic objects such as for optically trapped erythrocytes (488.0, 514.5, and568.2 nm for excitation of the heme moiety).2 Wavelengths in the deep ultraviolet (UV)– for example, of 229 nm – are used to enhance aromatic amino acids, while thewavelength of 257 nm leads to a predominant enhancement of Raman bands of nucleicacids.3

For the NIR region, diode lasers operated at room temperature have been exhaustivelyused for gas analysis within the last decades. From the permanently increasing noninvasive13C-breath tests for the investigation of metabolic processes, the urea breath test fordiagnosing a Helicobacter pylori infection, causing in some people peptic ulcers and evencancer in its worst case, is the most prominent, for which diode lasers have been used forisotope-selectivemeasurements. Many recent developments go beyond gastroenterologicalapplications.4 An apparatus recently developed for breath analysis and based on photo-acoustic spectroscopy, using a wavelength-modulated distributed feedback (DFB) diodelaser and taking advantage of the acoustic resonances of the sample cell, allows sensitivemeasurements with detection limits for 13CO2 of a few parts per million (ppm).

5 Alterna-tively, nondispersive IR spectrometric devices have routinely been used for suchdiagnostics.6

Furthermore, several applications of mid-infrared (MIR) quantum cascade lasers(QCL), aimed at monitoring blood glucose, have recently been reported with the claimof allowing a miniaturization of a device to the point where personal use of a wearableinstrument may be realized.7,8 The feasibility of the simultaneous quantification of twodifferent compoundsmeasured at twowavelengths using dualQCL absorption spectroscopyhas been reported by Schaden et al.9 However, miniaturized devices have yet not beenadvanced to the size of portable instrumentation, despite the promises made for QCLtechnology or NIR tunable lasers.10

A further glimpse is caught of important and interesting, but not routinely applied,measurement techniques. In the past, the theoretical basis for using vibrational spectroscopyas a tool for structure analysis has beenwell established.As an example, the conformation ofbiological molecules such as peptides, proteins, nucleic acids, and carbohydrates can bedetailed, much opposed to the view of IR and Raman spectroscopy being low-resolutiontechniques that cannot compete with nuclear magnetic resonance (NMR) orX-ray crystallography. For clarifying this partiality, a recent comprehensive review bySchweitzer-Stenner11 discussed peptide and protein structures elucidated by vibrationalspectroscopy.

In this context, vibrational optical activity (VOA) is another area that must bementioned.12,13 It is composed of two areas, vibrational circular dichroism (VCD),providing the difference in the IR absorbance of a chiral molecule for left versus rightcircularly polarized radiation, and Raman optical activity (ROA), which is thecorresponding difference for Raman scattering. Routinely, VCD spectra are measuredwith Fourier transform–infrared (FT-IR) instruments with commercial spectrometers

10 BIOMEDICAL VIBRATIONAL SPECTROSCOPY — TECHNICAL ADVANCES

available since 1997, which are now used worldwide in research laboratories.14 Later in2003, instrumentation for ROA measurements has also become commercially available.The research group of Nafie and Freedman is interested also in extending VCD and ROAinto newareas such asNIRVCDof overtones and combination bands,15 NIR excitedROA,and surface-enhanced ROA and VCD techniques. Enhancement factors of many orders ofmagnitude have been observed in hot spots with high-surface plasmon fields, enablingeven single molecule detection, thus adding an additional level of chiral sensitivity to thismethod of structural analysis.

In the following, measurement techniques for clinical chemistry analysis will bediscussed in more detail, for which biofluids such as whole blood, serum, dialysates, urine,and others, but also solid samples like gallstones and urinary calculi, must be listed.Reagent-free vibrational spectroscopy can provide quantitative results for the specimencomposition or can furnish the physician with information on the etiopathology of thepatient. Furthermore, pathology assisting and supporting vibrational techniques, either forbiopsies or in vivo diagnosis, are illustrated. Finally, in vivomonitoring of pivotal metabolicparameters and the redox status of important proteins based on near-infrared spectroscopy(NIRS) will be reviewed.

2.2 MEASUREMENT TECHNIQUES FOR CLINICAL CHEMISTRY

2.2.1 Analysis of Liquid Samples

Molecular spectroscopy has brought much progress for medical diagnostics, and particu-larly the marriage of vibrational spectroscopy with clinical chemistry will enable theimplementation into point-of-care analytics for patient monitoring. In the past, this areawasreviewed extensively,16–18 but several novel techniques have been developed since then andthe most interesting applications will be explicated.

Biofluid analysis has several aspects because there is the measurement of liquidaqueous samples involved by using attenuated total reflection (ATR) and transmissionspectroscopy with a goal of such instrumentation being developed for routine analyzers.First applications of IR spectroscopy for substrate analysis in whole blood and bloodplasmawere reported about 20 years ago.19,20 Among the different options, discrete bloodsampling with subsequent sample preparation has been chosen for many glucose assays.Whereas for MIR spectroscopy the ATR technique or transmission measurements havebeen used for the analysis of liquid body fluids, exclusively transmission measurementswere carried out when NIR or even short-wave NIR spectroscopy were exploited.21

However, when simulating the scattering in biological tissue, also diffuse reflectancemeasurements have been carried out with intralipid solutions spiked with glucose.22

Further details on diffuse reflectance measurements for tissue analysis are given in the invivo spectroscopy section.

Some MIR spectral signatures of different biomolecules are displayed in Figure 2.1from transmission measurements of crystalline powders using the KBr pellet technique andspectra obtained by transmission and ATRmeasurements of aqueous solutions, which servefor their quantitative analysis in body fluids. One of theATRmeasurements has been carriedout using a flow-through micro-Circle cell, which contains a pin-like ZnSe crystal withcones at its ends for optimal radiation coupling (inner volume�30 mL).Owing to the severalinner reflections, the transmission equivalent optical sample path length is larger whencompared with the spectral absorbance resulting from two internal reflections in a diamond

MEASUREMENT TECHNIQUES FOR CL IN ICAL CHEMISTRY 11

prism at 45� (see also Figure 2.2). For this fiber-optic probe with a microprism asATR-sensor element, both fibers – that for illumination and the other for waveguiding tothe MCT detector – were of the same square cross section to fill the diamond prism base of1.5mm� 0.75mm completely.

Other accessories such as a horizontal diamond ATR cell with three internal reflections(DurasampleII, SensIR) have been used for continuous fermentation monitoring23 orwhole blood measurements.24 Transmission micro-cells have been fabricated withinner volumes of less than 1mL.25 Best quantification can be achieved by using the MIR

Figure 2.1. Infrared spectra of biologically relevant substances. (A) spectra measured in transmis-

sion using crystalline powders and the KBr pellet technique. (B, C) Aqueous glucose and urea

solution spectra measured in transmission, using a micro-Circle cell with a ZnSe crystal (several

internal reflections) and a diamond microprism with two internal reflections at 45�, respectively.Thewater absorbance fromthe solventhadbeencompensatedbybackgroundmeasurements using

a water-filled cell.

12 BIOMEDICAL VIBRATIONAL SPECTROSCOPY — TECHNICAL ADVANCES