This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
Author's personal copy
The emergence of biospectroscopy in stem cell research 13
proteins, lipids, carbohydrates, and nucleic acids within (Bhargave and Levin, 2005; Levenson et al., 2006). However
individual cells in a single global measurement, achieved the spatial resolution routinely achievable with a synchro-
nondestructively in seconds, and requiring minimal prepara- tron is proving critical in examining cellular changes at the
tion with no staining. boundaries between tissues, and where the changes in cel-
The method relies on knowledge of distinctive spectral lular function are subtle, or confined to small areas. It is
bands that relate to vibrational modes of functional groups becoming common for researchers to combine both FPA data
within the macromolecules. Naumann and co-workers and synchrotron data to achieve a more complete picture.
(Naumann et al., 1991) were the first to identify and start These IR microscope systems are also complemented by
to assign the FTIR spectra of biological cells, with over 40 high-throughput FTIR systems capable of examining large
distinctive bands in mid-infrared frequencies from 4000 to populations of cells grown in a multiwell format on infrared-
600 cm−1 that can be uniquely identified to macromolecular compatible substrates (Benezzeddine-Boussaidi et al.,
components in the FTIR spectrum of biological cells. Many of 2009). Combining such systems with automated sample
these bands overlap in the raw, unprocessed spectrum and handling allows the screening of many cell populations for
are only revealed as distinct and individual bands when the changes in cellular chemistry as revealed by their IR spectra.
data are preprocessed by taking a first or second derivative Bioinformatic approaches to data reduction and analysis
of the original data. The need to apply appropriate mathe- are a cornerstone of this research area. This arises because
matical processing and analysis is the final unifying aspect FTIR spectroscopic data are multivariate in nature containing
that defines biospectroscopy, i.e., the use of sophisticated many bands, with important biological differences often
bioinformatic approaches to objectively define the changes registered as subtle changes in band intensities, band maxima
in biochemistry measured, as well as being used in the position, and broadening, often with a combination of factors
construction of images based on this information. The paper that change in synchrony in response to a biological change or
by Walsh et al. illustrates all these aspects of biospectro- phenomenon. This is illustrated well in the study by Walsh et
scopy and is an instructive starting point for further reading al. (Walsh et al., 2009), where the discrimination of different
for the uninitiated (Diem et al., 2008). cell types is based on clustering in scores plots generated by
The overarching and powerful principal of Naumann's Principal Component Analysis (PCA) (Wold, 1976), the most
early work was that not only is detailed investigation of parts commonly employed multivariate approach used in biospec-
of the FTIR spectra useful for determining the nature of troscopic research. In the study the PCA approach is extended
biochemical changes in cells via the FTIR spectrum, but more to allow objective classification of independent validation sets
importantly that the complete spectral data derived from using Linear Discriminant Analysis (LDA) (McLachlan, 2004).
the cell constituted what Naumann described as a “spectral The multivariate analysis provides the ability to use loadings
fingerprint,” which uniquely defined the type of cell under plots that show the spectral bands that are changing and
investigation. This principle was powerfully demonstrated by explain the clustering observed in the PCA scores plots, which
Naumann and co-workers being able to reproducibly and shows the types of biochemical changes occurring and whether
accurately discriminate different strains within individual the components are increasing or decreasing relative to one
pathogenic bacteria species (Kirschner et al., 2001). Other another. Another classification approach coming into vogue in
studies have shown that human cell types also can be biospectroscopy is Artificial Neural Network (ANN) (Hornik
discriminated in a similar way (Wood et al., 1998) and from et al., 1989; Lasch et al., 2006) classification, which provides a
these studies a variety of cancers and many other human useful comparison to PCA-based analyses, in that it is not
disease states have been discriminated and pathology reliant on linear relationships between the data, but succeed
imaged using FTIR spectroscopy (McNaughton et al., 2008). by the ability to train ANNs to recognize distinctive patterns.
It is this notion of the existence of spectral phenotypes for ANN analysis is also attractive as it is computationally much
different cell types that Walsh et al. use to demonstrate less intensive compared to the linear-based approaches such
longitudinal changes in differentiation states along small and as LDA (Lasch et al., 2006).
large intestinal crypts and also demonstrate relatedness Biospectroscopy can still be considered in its infancy,
between particular cell types. Furthermore, the nature and despite enormous breakthroughs and advances in instrumen-
direction of biochemical changes broadly associated with the tation over the previous few decades, and this is evidenced
differentiation process are identified, underscoring the by most published work still being found within journals
difference between spectroscopic approaches and histologi- strongly associated with physical chemistry. The break-
cal approaches. through for these extremely powerful new techniques into
The work by Walsh et al. employs mapping of areas of mainstream biomedical research areas depends on careful
tissue sections at the single cell level necessitating emplo- integration of the new approaches with the established
ying light from a synchrotron source. This type of approach is measurements, with the aim to correlate and corroborate
usually taken when the highest spatial resolution and signal the new approaches. Walsh et al. (Walsh et al., 2009) rely on
to noise measurements are required. The nature of the position coordinates within equally spaced grids as an
synchrotron source allows an infrared beam to be focused to innovative first approach to establishing the relationship
a spot size that is equivalent to or smaller than the size of a between the cell type and the spectra obtained to achieve a
single mammalian cell, allowing chemical information to be correlation between the FTIR mapping data and histology.
gained, with high sensitivity, at subcellular spatial resolu- Their paper demonstrates the enormous potential for
tion. The trade-off for this gain in sensitivity and spatial biospectroscopic approaches to add new information to the
resolution is that the area of sample that can be studied in an field of stem cell biology by providing biochemical data of a
acceptable time is much smaller than is possible with a state- global nature that can either fingerprint cell types or provide
of-the art Focal Plane Array (FPA) infrared imaging system an indication of overall changes in macromolecular
Author's personal copy
14 P. Heraud, M.J. Tobin
composition that is associated with stem differentiation. fication and identification of enterococci: A comparative pheno-
There can be little doubt that these studies demonstrate the typic, genotypic, and vibrational spectroscopic study. J. Clin.
application of new and powerful complimentary biospectro- Microbiol. 39, 1763–1770.
Krafft, C., Salzer, R., Seitz, S., Ern, C., Schieker, M., 2007. Diffe-
scopic approaches for shedding new light on stem cell biology
rentiation of individual human mesenchymal stem cells probed
and hopefully herald a burgeoning of the new methodologies
by FTIR microscopic imaging. Analyst 132, 647–653.
into stem cell research. Lasch, P., Diem, M., Hänsch, W., Naumann, D., 2006. Artificial neural
networks as supervised techniques for FT-IR microspectroscopic
imaging. J. Chemometr. 20, 209–220.
References Levenson, E., Lerch, P., Martin, M.C., 2006. Infrared imaging: Syn-
chrotrons vs. arrays, resolution vs. speed. Infrared Phys. Technol.
Ami, T., Neri, A., Natalello, P., Mereghetti, S.M., Doglia, M., Zanoni, 49, 45–48.
M., Zuccotti, S., Garagna, C.A., 2008. Redi, Embryonic stem cell McLachlan, G.J., 2004. Discriminant Analysis and Statistical Pattern
differentiation studied by FT-IR spectroscopy. BBA – Mol. Cell Recognition. Wiley-Interscience, USA.
Res. 1783, 98–106. McNaughton, D., Wood, B.R., 2007. Applications of FTIR imaging in
Benezzeddine-Boussaidi, L., Cazorla, G., Melin, A.M., 2009. Valida- cancer research. In: Kneipp, K., Aroca, R., Kneipp, H., Wentrup-
tion for quantification of immunoglobulins by Fourier transform Byrne, E. (Eds.), New Approaches in Biomedical Spectroscopy,
infrared spectrometry. Clin. Chem. Lab. Med. 47, 83–90. ACS Symposium Book Series, 963, pp. 14–29. Washington DC.
Bentley, A.J., Nakamura, T., Hammiche, A., Pollock, H.M., Martin, F.L., McNaughton, D., Bambery, K., Wood, B.R., 2008. Spectral Histo-
Kinoshita, S., Fullwood, N.J., 2007. Characterization of human pathology of the human cervix. In: Diem, M., Chalmers, J.M.,
corneal stem cells by synchrotron infrared micro-spectroscopy. Griffiths, P.R. (Eds.), Vibrational Spectroscopy for Medical
Mol. Vis. 13, 237–242. Diagnosis. Wiley, UK.
Bhargave, R., Levin, I., 2005. Spectrochemical Analyses Using Naumann, D., Helman, D., Labischinski, H., 1991. Microbiological
Multichannel Infrared Detectors. Blackwell, UK. characterizations by FT-IR spectroscopy. Nature 351, 81–82.
Chan, J.W., Lieu, D.K., Huser, T., Li, R.A., 2009. Label-free separation Two heads are better than one. Nat. Methods 1, 183, doi:10.1038/
of human embryonic stem cells and their cardiac derivatives using nmeth1204–183.
Raman spectroscopy. Anal. Chem. 81, 1324–1331. Walsh, M.J., Fellous, T.G., Hammiche, A., Lin, W.R., Fullwood, N.J.,
Diem, M., Chalmers, J.M., Griffiths, P.R., 2008. Vibrational Spectro- Grude, O., Bahrami, F., Nicholson, J.M., Cotte, M., Susini, J.,
scopy for Medical Diagnosis. Wiley, UK. Pollock, H.M., Brittan, M., Martin-Hirsch, P.L., Alison, M.R.,
German, M.J., Pollock, H.M., Zhao, B., Tobin, M.J., Hammiche, A., Martin, F.L., 2008. Fourier transform infrared microspectroscopy
Bentley, A., Cooper, L.J., Martin, F.L., Fullwood, N.J., 2006. Cha- identifies symmetric PO-2 modifications as a marker of the
racterization of putative stem cell populations in the cornea using putative stem cell region of human intestinal crypts. Stem Cells
synchrotron infrared microspectroscopy. Investig. Ophthalmol. Vis. 26, 108–118.
Sci. 47, 2417–2421. Walsh, M.J., Hammiche, A., Fellous, T.G., Nicholson, J.M., Cotte, M.,
Heraud, P., Wood, B.R., Tobin, M., Beardall, J., McNaughton, D., Susini, J., Fullwood, N.J., Martin-Hirsch, P.L., Alison, M.R., Martin,
2005. Mapping of nutrient-induced biochemical changes in living F.L., 2009. Tracking the cell hierarchy in the human intestine using
algal cells using synchrotron infrared microspectroscopy. FEMS biochemical signatures derived by mid-infrared microspectro-
Microbiol. Lett. 249, 219–225. scopy. Stem Cell Res 3, 15–27.
Hornik, K., Stinchcombe, M., White, H., 1989. Multilayer feed- Wold, S., 1976. Pattern recognition by means of disjoint principal
forward networks are universal approximators. Neural Netw. 2, component models. Pattern Recogn. 8, 127–139.
359–366. Wood, B.R., Quinn, M., Tait, B., Ashdown, M., Hislop, T., Romeo, M.,
Kirschner, C., Maquelin, K., Pina, P., Thi, N., Choo-Smith, L.P ., McNaughton, D., 1998. FTIR microspectroscopic study of cell
Sockalingum, G.D., Sandt, C., Ami, D., Orsini, F., Doglia, S.M., types and potential confounding variables in screening for
Allouch, P., Mainfait, M., Puppels, G.J., Naumann, D., 2001. Classi- cervical malignancies. Biospectroscopy 4, 75–91.