Session 1, Saturday, October17th
9:30 - 11:30     Improving the Quality of Diagnosis and Prognosis (I)

Sensitivity of Infrared Spectral Features Toward Differentiation, Maturation, Cell Cycle Dependence and State of Health of Human Cells

Luis Chiriboga1,2, Susie Boydston-White1 and Max Diem

1Department of Chemistry and Biochemistry, City University of New York, Hunter College, 695 Park Avenue, New York, NY 10021, USA
2Molecular Diagnostics Laboratory, Department of Pathology, Bellevue Hospital, New York University, 27th Street and 1st Avenue, New York, NY 10016

The changes in the infrared spectral patterns observed between healthy and diseased tissues can be classified into two different classes which we refer to as gross spectral changes, and disease induced changes. The first class of changes includes, for example, variations of glycogen content in tissue sections or single cells that are nonspecific for disease. Similarly, changes in the structural protein content of tissue, increased vascularity, or the changes accompanying cell maturation and differentiation can be observed infrared spectroscopically, but these changes are not necessarily correlated with the occurrence of disease. However, these changes are ideally suited to create maps of tissues or distributions of cells that can augment the information obtainable from photomicrographs of stained tissues used in pathology. In fact, the advantage of infrared (false color) mapping is that such a map contains more information, per pixel element, than stained tissue, and can be constructed and interpreted totally objectively by computer methods.
Among the disease induced spectral changes, we and others have found that certain spectral regions due to nuclear DNA appear to be enhanced in samples with diagnosed cancerous disease. Since the DNA spectral features are superimposed on those due to nuclear and cytoplasmic RNA, these changes are very subtle. In an attempt to quantify and interpret the changes in the DNA/RNA spectral contributions in healthy and abnormal cells and tissue, we have found that similar changes in the DNA/RNA spectral features are observed for cells that are at different stages in the cell division cycle. This cell cycle dependent spectra were collected for cultured myeloid leukemia (ML-1) cells that were isolated into pure fractions according to the stage in the cell cycle. This observation leads to a possible conclusion that all hitherto observed spectral changes between normal and abnormal cells and tissues are due to different distributions of cells at given stages of their development. On the other hand, the possibility exists that a small number of abnormal cells with large DNA spectral contributions dominates the observed spectra of abnormal samples.
A recent analysis of spectral data from single cells confirms that spectral patterns from a sample of exfoliated cells can be interpreted in terms of the cell maturation, and in terms of distributions of cells according to their stage in the cell cycle. We, therefore, believe that the understanding of the infrared spectra of cells in terms of cell differentiation, maturation and life cycle is crucial for the correct interpretation of infrared spectra of healthy and abnormal cells and tissues.


IR-Spectroscopy and IR-Microscopy of Breast Tumor Cell Lines and Human Breast Tumor Tissues

Heinz Fabian, Arnfried Schwartz, Ralf Wessel and Iduna Fichtner

Max-Delbrück-Center for Molecular Medicine, Berlin, Germany

Peter Lasch and Dieter Naumann

Robert Koch-Institute, Berlin, Germany

Due to recent advances in instrumentation and data processing techniques, infrared techniques are not only invaluable tools for the structural characterization of the building blocks of living organisms, but now also allow successful studies of complex biomedical materials. In order to evaluate the diagnostic potential of FTIR spectroscopy, we have started to analyse samples that can be isolated in a highly reproducible manner, and which can be characterized by other, independent techniques. Such systems are tumor cell lines. Comparative studies of human breast tumor cell line suspensions revealed that IR methods have the potential for differentiating between related cell lines and are sensitive to biochemical changes associated with the cell death. The collection of IR spectra through microscope optics permits this analysis to be made from few cells, which raises the possibility of analysing tissues by IR microscopy and searching for small cluster of abnormal cells. To assess the potential of the IR technique for the diagnosis of breast tumor tissue, thin sections of tissue were mapped by IR microspectroscopy. The spectra of the maps were analysed using functional group mapping or pattern recognition techniques. The output values of the different approaches were then reassembled into IR images of the tissues. To allow a definite correlation between tissue features and spectroscopic properties, alternate sections of tissue were used for infrared spectroscopic measurements and for light microscopic staining. Chemical mapping based on single band intensities turned out to be an easy way to identify major breast tissue constituents, such as connective tissue, malignant epithelium and fat. Cluster analysis of the infrared spectra, combined with immunohistochemical staining of the tissue sections, provided deeper and more detailed insights. Based on these data, the potential of IR microspectroscopy for the diagnosis of breast cancer tissue sections is discussed


Examination of Cervical Smears by FTIR Spectroscopy: A Study of the Dilution Effect and the Archiving of Smears

Janie Dubois1, Ashraf A. Ismail1 Jocelyne Arseneau2, and Manon Auger3

1McGill IR Group, Macdonald Campus of McGill University, 21,111 Lakeshore Road, Ste-Anne-de-Bellevue, Quebec, Canada, H9X 3V9.
2Department of Oncology, McGill University, Montreal, Quebec, Canada.
3Department of Pathology, McGill University, Montreal, Quebec, Canada.

The potential applicability of FTIR spectroscopy for the examination of cervical cells has now been under investigation for nearly a decade. The impetus of these studies is twofold. On a fundamental level, infrared spectroscopy can provide information about the molecular changes that occur as cells are transformed from a normal to an abnormal state. A secondary goal of these studies is the development of FTIR spectroscopy as a diagnostic tool that can replace or complement the current methods for screening which predominately rely on visual examination of exfoliated cells to diagnose cervical pre-cancerous lesions. We have investigated the sample handling and archiving aspects of cervical smears as part of an ongoing project aimed at the development of a dedicated FTIR-based cell analyzer. In North America, Pap smears have to be kept for 5 years, for follow-up purposes, in cases where no abnormality is dedicated. It is also common practice to keep all abnormal samples indefinitely. Therefore, it is necessary to prepare cervical smears for infrared analysis in a form allowing preservation of the samples. We have established a procedure involving minimal sample preparation and the use of custom-made polyethylene slides as a disposable IR support and have evaluated the stability of the samples during 5 years of storage. Minimal changes in the spectra of the samples have been observed over this period. The main problems encountered arose from excessive handling of the samples, resulting in detachment of the cells from the support. This problem was seen in <10% of the samples examined. Various corrective measures are under investigation.
The archived samples were also used to evaluate the feasibility of detecting abnormal cells in the presence of a large number of visually normal cells (and in some cases the presence of bacteria resulting from infection). It was hypothesized that the presence of one or two abnormal cells among hundreds of normal cells, while sufficient for the classification of a sample as abnormal by visual microscopy, would not have a significant influence on the infrared spectrum of the whole sample due to the dilution effect. FTIR spectra of cervical cell smears recorded from cross-sectional areas ranging from 5000 ´ 5000 mm to 20 ´ 20 mm were compared. Some samples showed definitive differences between the spectra of the whole sample and those of specific areas. On the other hand, some samples diagnosed as abnormal by cytopathologists exhibited remarkably similar spectra from all portions of the sample, whether or not each particular portion contained cells morphologically classified as abnormal. The explanation for these findings is not clear yet because exfoliated cervical cells do not come from one specific area of the cervix and therefore should, in principle, not be equally affected by pre-cancerous lesions.


Infrared Micro-Spectroscopy of Cervical Smears and Activated Lymphocytes

D. McNaughton, B. Wood and M. Romeo

Chemistry Department, Wellington Rd., Monash University, Clayton, Victoria Australia.

Over the last few years we have built a database of infrared spectra of cervical smears from patients attending the dysplasia clinic of the Royal Womenís Hospital (Melbourne). We have also investigated the spectra of a number of components that may appear as confounding variables in the spectra of cervical smears. These include bacteria common to the female genital tract, semen, endocervical mucin and blood components. We find lymphocytes, particularly activated lymphocytes, which will often be present due to non specific disease or inflammation, to be of most concern for analysis and diagnosis. The presence of these confounding variables indicates that an approach using multivariate statistical techniques or artificial neural networks is the most appropriate for analysis and subsequent diagnosis. To date we have subjected some of our data, where biopsy results are available for comparison, to analysis using the multivariate statistical techniques of SIMCA (Soft Independent Modelling of Class Analysis) and K-nearest neighbours (KNN) in addition to developing an analysis based on artificial neural networks. The methodology we have developed will be described and results of the preliminary analyses will be discussed and compared.
Lymphocytes activated with the agent phytohaemagluttinate (PHA) show spectral differences when compared with non activated lymphocytes and we have carried out a study monitoring lymphocyte activation over time. These studies show distinct spectral changes over the first few hours after activation and with the use of principal component analysis (PCA) of derivative spectra these changes can be attributed to RNA synthesis. These results, which show that infrared may be useful in the initial screening of donors in organ transplantation together with the results of a study monitoring the activation of mixed lymphocytes will be discussed.


How Can Infrared Spectroscopy Contribute to Diagnosis and Prognosis of Chronic Lymphocytic Leukemia?

Christian P. Schultz

Institute for Biodiagnostics, National Research Council of Canada

Chronic lymphocytic leukemia (CLL) is a disorder of morphologically mature but immunologically less mature lymphocytes and causes immunosuppression, failure of the bone marrow, and infiltration of malignant (cancerous) cells into organs. Usually the symptoms and the course of the disease will develop gradually and occur in the elderly with 90% of the cases in people over 50 years old. Many cases are accidentally detected by routine blood tests in people with no symptoms. The cause of CLL is unknown and until today no curative therapy has been established. Treatment decisions are made by the physician on the basis of a variety of factors: a) rate of increased lymphocyte count, b) pattern of lymphocyte spread in the bone marrow, c) size of spleen and lymph nodes and d) rate development of anemia and low platelet counts.
Infrared spectroscopy can successfully distinguish normal mature lymphocytes from apparently mature CLL cells, based on IR band analysis and pattern recognition techniques. The same techniques reveal that the doubling time of CLL cells is imprinted in the IR spectra and may therefore be used to predict the course of a patientís disease, the treatment and possible outcome. The infrared spectra of CLL cells also appear to contain many other differentating features, suggesting that the biochemical composition of these cells varies more than is indicated by conventional methods. In vitro drug treatment of CLL cells usually provides information on the cellís drug sensitivity and may reflect the in vivo cell response to a particular drug - very important for the selection of the most effective drug. Infrared spectra of the untreated CLL cells can be correlated with these values of in vitro drug resistance (e.g. MTT assay) and then used for the construction of a large training set containing spectral patterns for sensitive and resistant cells. Applying this model to IR spectra of undetermined and untreated cell samples then allows the prediction of drug sensitivity on the original CLL cells without performing the actual MTT test. Instead of correlating IR spectra with drug resistance values, the drug testing itself can be performed and monitored by infrared spectroscopy. This has the benefit of accessing cell changes within cancer cells directly which may lead to a better understanding of the drug cell interaction and the actual mechanism. For example, IR spectroscopy was succesfully used to follow the membrane changing action of Bryostatin 1, a potentially new drug (in clinical trials) that transforms drug resistant CLL cells back into sensitive ones.

FT-IR Workshop Series, Berlin, Robert Koch-Institute Home

Any opinions, findings and conclusions or recommendations expressed in this publication are those of the workshop organizers and do not necessarily reflect the views of the Robert Koch-Institute. © 2018 Peter Lasch