Session 6, Sunday, October 18th

 10:45 - 12:30 Spectroscopy as an Aid in Clinical Prodedures: Clinical Microbiology

 

 

The Rapid Characterisation of Microbial Pathogens Using Hyperspectral, Whole-Organism Fingerprinting and Chemometrics.
 
Royston Goodacre*1, Éadaoin M. Timmins1, Richard J. Gilbert1, Janet Taylor1, Paul J. Rooney2 and Douglas B. Kell1
 
1Institute of Biological Sciences, University of Wales, Aberystwyth, Ceredigion, SY23 3DD, Wales and 2Bronglais General Hospital, Aberystwyth, Ceredigion, SY23 1ER, Wales.
*Telephone: +44 (0)1970 621947 Telefax: +44 (0)1970 622354 E-mail: rrg@aber.ac.uk

 

Ideal techniques for the rapid characterisation of microbial pathogens would include those which require minimal sample preparation, permit the automatic analysis of many serial samples with negligible reagent costs, allow their rapid characterisation against a stable database, would be easy to use and would be operated under the control of a PC. With recent developments in analytical instrumentation, these requirements are increasingly being fulfilled by pyrolysis mass spectrometry (PyMS) and the vibrational spectroscopic methods of Fourier transform-infrared spectroscopy (FT-IR) and dispersive Raman microscopy.

We have shown that these techniques, but only when combined with chemometric methods, provide very rapid, accurate and generic approaches to the characterisation of micro-organisms. Examples will be given where we have shown that it is possible to (1) discriminate between common infectious agents associated with urinary tract infection [1], (2) assess the resistance or otherwise of Staphylococcus aureus to the antibiotic methicillin [2, 3].

Chemometric prossessing has classically been done with partial least squares and artificial neural networks, but the information in terms of which parts of the spectra are important is not readily available, which is (i) why these methods are perceived as a ‘black box’ approaches to modelling spectra, and (ii) indeed why these 'data crunching' techniques are often out of favour with many spectroscopists. Genetic programming (GP) is an evolutionary technique which uses the concepts of Darwinian selection to generate and optimize a desired computational function or mathematical expression [4, 5].

Examples of our recent developments in GP will be demonstrated; (1) for the identification from their PyMS spectra, of a group of bacteria which have been implicated in periodontitis, endodontic infections and dentoalveolar abscesses [6], and (2) for the detection of the dipicolinic acid biomarker in the FT-IR spectra of Bacillus spores [7].

[1] Goodacre, R., Timmins, É. M., Burton, R., Kaderbhai, N., Woodward, A., Kell, D. B. and Rooney, P. J. (1998). Rapid identification of urinary tract infection bacteria using hyperspectral, whole organism fingerprinting and artificial neural networks. Microbiology 144, 1157-1170.

[2] Goodacre, R., Rooney, P. J. and Kell, D. B. (1998). Rapid analysis of microbial systems using vibrational spectroscopy and supervised learning methods: application to the discrimination between methicillin-resistant and methicillin-susceptible Staphylococcus aureus. SPIE 3257, 220-229.

[3] Goodacre, R., Rooney, P. J. and Kell, D. B. (1998). Discrimination between methicillin-resistant and methicillin-susceptible Staphylococcus aureus using pyrolysis mass spectrometry and artificial neural networks. Journal of Antimicrobial Chemotherapy 41, 27-34.

[4] Koza, J. R. (1992). Genetic Programming: On the Programming of computers by Means of Natural Selection. Cambridge, MA: MIT Press.

[5] Gilbert, R. J., Goodacre, R., Woodward, A. M. and Kell, D. B. (1997). Genetic programming : a novel method for the quantitative analysis of pyrolysis mass spectral data. Analytical Chemistry 69, 4381-4389.

[6] Taylor, J., Goodacre, R., Wade, W. G., Rowland, J. J. and Kell, D. B. (1998). The deconvolution of pyrolysis mass spectra using genetic programming: application to the identification of some Eubacterium species. FEMS Microbiology Letters 160, 237-246.

[7] Goodacre, R., Shann, B., Gilbert, R. J., Timmins, É. M., McGovern, A. C., Alsberg, B. K., Logan, N. A. and Kell, D. B. (1998). The characterisation of Bacillus species from PyMS and FT IR data. In Proc. 1997 ERDEC Scientific Conference on Chemical and Biological Defense Research. Aberdeen Proving Ground.

 

 
 
 
The Challenge of Rapidly Characterizing (Drug Resistant-)Microorganisms: A Confocal Raman Microspectroscopic Approach
 
L.P. Choo-Smith1, K. Maquelin1, H. Ph. Endtz2, H.A. Bruining3, G.J. Puppels1,3
 
1Laboratory for Intensive Care Research and Optical Spectroscopy, Erasmus Univerisity Rotterdam; 2Dept. Medical Microbiology and Infectious Disease, University Hospital Rotterdam "Dijkzigt"; 3Dept. General Surgery, University Hospital Rotterdam "Dijkzigt", Surgery 10M, Dr. Molewaterplein 40, 3015 GD Rotterdam, The Netherlands.
 

Knowledge of the identity and antibiotic susceptibility profile of microorganisms is critical for the management of infection. With the lack of timely laboratory results, inappropriate antibiotics might be prescribed which are ineffective against the micro-organisms responsible for the infection. Not only does the patient's condition deteriorate, the use of inappropriate drugs contributes to the emerging problem of drug resistance in micro-organisms. Thus, there is a need for novel methods which can rapidly (within hours to 1 day of receipt of patient material) provide the clinician with the identity and antibiotic susceptibility profile of micro-organisms. With this knowledge, target antibiotics can be used, with broad-spectrum antibiotics reserved for cases of resistant organisms. Furthermore, the laboratory results will help identify those patients who should be isolated in order to prevent the spread of multi-drug resistant microorganisms within health-care institutions.

In recent years, DNA-based amplification techniques have been developed to rapidly identify microorganisms. These methods are attractive since the tests can be done on a few bacterial cells, do not involve culturing of the microorganisms and results are obtained within 1 day. Using specific DNA probes or amplification primers, the bacteria present in the sample can be identified and in certain cases, the resistance gene determined. However these tests are costly, involve highly trained personnel and still experience technical hurdles such as the problem of cross-contamination. Furthermore, the mere presence of the resistance gene is not always indicative of resistant bacteria. The complexity of resistance mechanism and the need for different primers to cover the various mechanisms have also been complicating factors in the detection of resistance genotypes.

Our approach to the challenge facing the clinical microbiology laboratory is the development of confocal Raman microspectroscopy for the rapid online identification and antibiotic drug susceptibility of microorganisms. Specifically, Raman spectra can be acquired directly from microcolonies (~ 6 hours growth time) still growing on solid culture medium. This approach involves small sample requirements and shorter culturing times (i.e. allow rapid identification and characterization). There is little sample preparation prior to measurement. The petri dish containing the cultured microorganisms can simply be placed under a microscope objective in order to select microcolonies (30-40 m m diameter) for spectral measurement. An added advantage is that the cells are still in contact with the culture medium and are kept viable thus providing a spectroscopic signature of the cells "in-situ" (i.e. growing on the culture medium). A problem with this approach is the presence of the underlying culture medium.

Preliminary studies reveal that Raman spectral measurements combined with multivariate methods have successfully classified (85-95% accuracy) Enterococcus faecium strains resistant or sensitive to vancomycin. The difference spectra reveal variation in spectral features which possibly arise from the drug susceptibility difference. Methods for improving the classification results are under investigation. Confocal Raman microspectroscopy in combination with multivariate methods can offer the microbiology laboratory a new tool for the rapid routine identification and drug susceptibility determination of microorganisms.

 
 
 
FTIR Approach of the Epidemiological Chain of Nosocomial Fungal Infections
 
F. Witthuhn2, G.D. Sockalingum1, D. Aubert2, M. Pinon2, M. Manfait1

 

1Laboratoire de Spectroscopie Biomoléculaire, UFR de Pharmacie, Université de Reims, France.
2 Service de Parasitologie-Mycologie, Hôpital Maison Blanche, Reims, France.

 

Nosocomial infections affect each year 5-10 % of hospitalised patients and would be responsible for more than 10000 deaths/yr in France. Although most of these infections are of bacterial origin, a recrudescence of fungal infections has been observed for the last few years. The evolution of fungal infections are linked to several factors among which: (a) rise in the number of susceptible subjects who are liable to develop fungal infections, (b) repartition of pathogenic species in the Candida genus (higher frequency o f C. glabrata, C. krusei, C. tropicalis isolation), (c) emergence of new pathogenic species, (d) appearance of Candida species resistant to antifungal agents (amphotericin B, fluconazole, itraconazole), (e) increase in invasive aspergillus.

The "screening" of the pathogenic agents by FTIR spectroscopy has been approached since two different strains, characterised by different molecular ultrastructures (protein, glucid, or lipid organization) present different absorption spectra. This approach is rather interesting since conventional phenotyping methods of strain comparison do not allow to confirm the identity or difference when the strains present the same phenotype profile. In these conditions, the epidemiological chain is more complicated to establish.

To validate the epidemic notion, FTIR results on different hospital patient and air samples will discussed.

 
 
 
Can FT-IR Spectroscopy Play a Role in Clinical Microbiology?
 
D. Naumann
 
Robert Koch-Institute, Biophysical Structure Analysis, Nordufer 20, 13353 Berlin
 

The characterization of microorganisms belongs to the most frequent tasks in clinical microbiology laboratories and includes detection, enumeration, differentiation, identification, and susceptibility testing. These analysis are performed using a plethora of very different techniques such as microscopy, serology, biochemical tests, and a variety of sometimes extremely sensitive and specific molecular genetic methodologies such as gene probing, ribotyping, screening for specific sequences by PCR to mention a few. Accepted requirements to be met by a modern technique for microbial characterizations in a clinical environment are (i) high specificity (differentiation down to the species and strain level), (ii) high sensitivity (detection limit should be below 103 cells), (iii) rapidity (results should be available within one working day), (iv) uniform procedures and applicability to all microorganisms, (v) fully automated and computer compatibility, and (vi) low amount of consumables. There is no system available on the market that fulfils all mentioned criteria.

FT-IR spectra of microorganisms are spectral fingerprints that can be used to rapidly differentiate, classify, and identify virtually any kind of pathogenic microorganisms. Besides the advantage of easy operation and direct access to results within minutes after obtaining the pure culture, FT-IR offers a remarkably far-reaching differentiation capacity. Results so far obtained suggest that epidemiological case studies, screening for pathogens and the characterization of clinical isolates for which established or routine tests are not yet available, or which are too time consuming and expensive, belong to the promising applications of the new technique. In the past couple of years an increasing number of FT-IR applications to microbiology appeared in the literature e.g. on the rapid identification and characterization of staphylococci, streptococci, clostridia, enterobacteriaceae, and clinically relevant yeasts. Recently, efforts are undertaken to establish FT-IR also as a means of rapid sensitivity testing of pathogens. FT-IR analysis on cell-drug interactions and testing for vancomycin resistant strains of Enterococcus faecium, multiresistant strains of Staphylococcus aureus or fluconazol resistant Candida albicans have already been reported. By means of an FT-IR spectrometer coupled to a light microscope it is possible to detect, enumerate, and identify pathogens within less than 8 hours (including cultivation) after isolation of the pure culture, or even in the presence of mixed cultures. We anticipate that new FT-IR microscopic technologies will emerge that are able to integrate detection, enumeration, differentiation, and sensitivity testing in one single, uniform and fully automatic instrumentation.

 

References:

1. D. Naumann, D. Helm, and H. Labischinski (1991). Microbiological characterizations by FT-IR spectroscopy. Nature 351, 81-82.

2. D. Helm, H. Labischinski, G. Schallehn, and D. Naumann (1991). Classification and identification of bacteria by Fourier-transfrom infrared spectroscopy. J. Gen. Microbiol. 137, 69-79.

3. É.M. Timmins, S.A. Howell, B.K. Alsberg, W.C. Noble, and R. Goodacre (1998). Rapid differentiation of closely related Candida species and strains by pyrolysis-mass spectrometry and Fourier transfrom-infrared spectroscopy. J. Clin. Microbiol. 36, 367-374.

 
 
Spectroscopic Data Evaluation and Pattern Recognition .
- Strategies and Perspectives -

 

J. Schmitt

 

Dept. Aquatic Microbiology, Spectroscopy Group, IWW and University of Duisburg, Moritzstr. 26, 187546 Mülheim, Germany, e-mail: 100740.1762@compuserve.com

 

Vibrational spectroscopy as a new tool in medicine has witnessed an enormous progress in the past. NIR, FT-IR and FT-Raman spectroscopy is used in manyfold techniques to characterize mammalian cells, tissues, body fluids and for spectral imaging. The common spectral feature of all complex biological samples ist their relatively small spectral variance between the objects, in contrast to the variance of different samples which is often poorly represented for practical reasons to date: a challenging task . To deal with the wealth and the depths of information, which is represented by spectroscopy and imaging, new strategies are required for qualitative and quantitative interpretation. The transfer of these elaborated spectroscopic techniques into routine analysis and in the clinical environment depends strongly on the reliability and accuracy of the evaluation system. Computer-based pattern recognition techniques is the term that encompasses a wide range of mathematical techniques for classifying data. It is useful for both imaging and spectroscopy but plays a different role in the two types of data. For imaging, the main emphasis is on producing automated reproducible techniques which will assist the human analyst in the interpretation of the image, whether to identify structures or to quantify some properties. With spectroscopy, the emphasis is more on discriminating between spectra from different classes of samples, quantitative analysis of spectral components and reducing the large numer of spectral features in order to make the available information more accessible.

The most promising range of methods, which are used to date will be briefly covered and their advantages and disadvantages summarized. The emphasis of the presentation is put on the strategies and the outlook of several pattern recognition techniques and their requirements in respect of their practical application and their value. These guidelines should support the discussion of the development of consistent data libraries and integrated data elucidation systems and their requirements to assist the medical community in research and diagnosis.
 
 
 
 


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