In clinical routine detection of bacterial pathogens is a key task. The information on the causative agent of a disease influences the choice of therapeutics and answers questions on hygiene. This is of importance in cases of severe diarrhea, which is not self-limiting, and in case of infected children, who are more compromised by loss of fluids and electrolytes.
Regarding intestinal pathogens, it is mainly based on stool studies and cultivation techniques in combination with morphological and physiological analysis and enzyme immunoassays.
However, also genotyping methods, like real-time PCR and microarray, have made their way into clinical diagnostics as fast detection techniques, but up to now only few have the necessary in-vitro diagnostics (IVD) approval to serve as alone standing diagnostic tool.
Nevertheless, genotyping methods have the potential once to replace phenotypic methods due to their speed in generating a diagnostic answer and attempts to miniaturize assays.
This may result in new point-of-care (POC) diagnostic devices, which can shift the place of generating a diagnose from the laboratory to the physician and, finally, to the patient (Ince 2009).
1.4.1 Phenotypic pathogen identification
Phenotyping methods based on stool enrichment and culture are the gold standard for intestinal pathogen identification in clinical routine. It is a complex and time-consuming task, which requires well-trained personnel. Stool studies for leucocytes, blood and mucus give first information if a bacterial pathogen is likely to be the cause of diarrhea (Hoshiko 1994).
Especially bloody diarrhea is a major diagnostic challenge for the clinician, because of the importance of distinguishing infection from non-specific inflammatory bowel disease and other inflammatory conditions of the colon (Farthing 2002). The definite identification generally requires time-consuming stool enrichment and selective culture. Indications for stool culture are the passage of six or more unformed stools per day, diarrhea for longer than three days, fever ≥38.5°C, dysentery, bloody stool, and multiple cases that suggest an outbreak (DuPont 2009; Chan 2003). Pathogen identification by stool culture usually takes between 12 and 48 hours or even longer depending on the pathogen. The usage of selective enrichment media and growth conditions allows first restriction of the potential etiologic agent. Identification comprises then (I) morphological, (II) biochemical, (III) physiological, and
31 (IV) serological characteristics. The characterization of the morphology includes colony shape, colony dimension, pigmentation, cell shape, Gram reaction, flagellation, and others.
Biochemical properties are defined by the presence of catalyzing enzymes like catalase (Campylobacter spp.), oxidase (Campylobacter spp.), nitrate reductase, or glutamate dehydrogenase (C. difficile). Physiological properties are e.g. the carbohydrate fermentation, indole production, urea splitting and growth under miscellaneous physiological features (temperature, pH, salt, gaseous environment). Serological tests, using immunological methods, for the presence of surface antigens, e.g. somatic (O-antigen), flagellar (H-antigen), or capsular (K-antigen) antigens, give valuable information for the identification of a pathogen and allow epidemiological conclusions. For salmonellae, more than 2,000 serovars based on H- and O-antigens were described. E. coli is classified into ~250 serovars by O-, H-, and K-antigens.
This phenotypic search for and characterization of a pathogen requires a proper algorithm to reduce the number of required tests and by this the cost of a diagnostic result. Moreover, phenotypic pathogen identification is accompanied by many difficulties. Some pathogens, such as Vibrio species, require special media and are more difficult to culture and routine stool cultures will for example not distinguish between EHEC and non-pathogenic E. coli from the normal intestinal flora. Slow growers, such as enteropathogenic mycobacteria, make the diagnostic process lengthy. Bacteria such as Campylobacter spp. require a rapid transport to the laboratory to ensure vitality and some bacteria, such as Yersinia enterocolitica, require a several days lasting cold enrichment. Problems in disease diagnosis will also occur in case of failed cultivation of the potential pathogen. Although the histological diagnosis of the infectious process plays a valuable role, acute infectious-type colitis is often indistinguishable from other inflammatory conditions of the gut such as ischaemia or chronic idiopathic inflammatory bowel disease (Lamps 2007). Furthermore, many cases of gastroenteritis caused by Vibrio spp., a pathogen that is mainly related to seafood, are under-recognized because culture of this species is not routinely done in clinical laboratories (Chan 2003).
The culture-based diagnostic process is facilitated by automated systems, such as VITEK 2 (BioMérieux, France), Phoenix (Becton Dickinson, USA), and Dynal eAIMS (Invitrogen), which allow fully automated cultivation, pathogen identification based on colorimetric tests, antibiotic susceptibility testing and analysis of resistance mechanisms, but they still suffer from the problems mentioned above.
1.4.2 Genotypic pathogen identification
Genotyping methods are a promising alternative to the current culture-based diagnostic methods for pathogen detection. The long time, which is required for identification of the causative agent by phenotyping, generally forces the physician to start an empirical therapy, since acute diarrhea strongly affects the life quality and can become life threatening.
Therefore, the results of stool culture are often of no therapeutic consequence anymore (Koletzko 2009). This was also described in a study, where 88.5% of gastroenteritis outpatients had recovered by the time that culture positive results were available, and thus a change of antimicrobial therapy was not required anymore (Chan 2003). For different bacterial pathogens varying antibiotics are recommended and non-adequate therapy can increase the risk of developing the haemolytic-uremic syndrome (DuPont 2009). Additionally, the mentioned study found a higher level of Ciprofloxacin resistance in Campylobacter spp.
in the 130 stool culture positive samples, underpinning the need for an evidence-based therapy. This is supported by other studies and a permanent increase of resistance among enteric pathogens can be observed (Garcia 2009; Senok 2007; NARMS 2009; Threlfall 2006). Evidence-based therapy is, therefore, of main importance in terms of a cautious usage of antibiotics. Prescription of unnecessary or unspecific antibiotics promotes the spread of resistance determinants (Donskey 2006). Moreover, an early identification of a bacterial pathogen allows changes in clinical management. Cost-intensive separation of a patient, which is for example required in case of a potential norovirus infection, is not
32 necessary anymore, if a bacterium is identified as causative agent. This can reduce hospitalization time and the cost of medical care. In one study, genotyping methods have also proven to increase the detection rates and by this to close a diagnostic gap (Ajjampur 2008).
To overcome the limitations of classical identification methods, that is low speed and accuracy, poor reproducibility, and intensive labour of trained personnel, new techniques that identify bacteria based on their genetic information without prior cultivation have been developed, such as FISH (fluorescence in-situ hybridization), DHPLC (denaturing high performance liquid chromatography), microarrays, real-time PCR and high-throughput sequencing. In principle, all methods that were used for the investigation of the intestinal flora (chap. 1.2) could be adopted for pathogen detection. An all-embracing overview cannot be given in this work, due to the amount of published methods for gastroenteritis-related pathogens. From the clinical point of view, however, only some methods have practical applicability, which is mainly a question of the methods cost, multiplexing capacity, handling efficiency, and speed. The increased sensitivity of nucleic acid-based tests requires improved contamination prevention and quality control in the clinical laboratory.
The adaption of amplification strategies based on PCR (Ke 1999), such as real-time PCR, RT-PCR (Zheng 2008; Zeng 2008), nested PCR (Chen 2000), PCR-ELISA (Sails 2001), and multiplex-PCR (Farfan 2010; Espineira 2010) was the first step towards intestinal pathogen genotyping in clinical laboratories. Assays have not only been described for clinical specimens but also for food and water, which can be the source of an intestinal infection. An isothermal amplification strategy, nucleic acid sequence-based amplification (NASBA), has been described for detection of viruses from faecal specimens (Lamhoujeb 2009) and Mycobacterium avium subsp. paratuberculosis from water and milk (Rodriguez-Lazaro 2004). Real-time PCR, which monitors the amplification reaction in real time, is to date the most relevant method in clinical laboratories. It has already made its way into medical laboratories as a fast identification method with reliable quantification ability, using specific primers for selective amplification of expected pathogens. In LightCycler technology it is combined with melting curve analysis of the PCR products (Lyon 2009). The main advantage of real-time PCR is its capability of quantifying the pathogen and its high sensitivity down to 10 or fewer copies of target (Liu-Stratton 2004). Assays have been developed for C. jejuni (Skanseng 2006), C. coli (Keramas 2004), EHEC (Hsu 2005; Fu 2005), Listeria (Huijsdens 2003), C. difficile (Penders 2005), and some pathogenic protozoa (Blessmann 2002; Limor 2002; Verweij 2004; Ng 2005). It is mainly restricted by its limited multiplexing capacity, which is due to two principle reasons: First, the interference between multiple primer pairs or probes and, second, the number of different fluorophores that can be used for simultaneous non-overlapping detection in a single tube, which is determined by the number of available channels (Petrik 2006). New devices, however, improved the multiplexing capacity by offering many PCR reactions in parallel using low sample volumes (LightCycler 1,536 Real-Time PCR System, ABI PRISM 7900HT Fast Real-Real-Time PCR System 384-well, Fluidigm BioMark System 9,216-plex). Assays for the detection of Campylobacter, Salmonella, L. monocytogenes, Listeria, E. coli O157 (Roche Applied Science 2010), and C. difficile (Prodesse ProGastro Cd/ GenProbe; Xpert C. difficile/ Cepheid) are commercially available.
LUMINEX’s xMAP technology is a method combining flow cytometry with colour-coded microspheres that allow up to 100 parallel bioassays by coating the beads with different capturing molecules (www.luminexcorp.com). The advantage of this bead-based assay is the favourable reaction kinetics in solution. Based on the LUMINEX technology detection assays were described for the most common food-borne pathogens (Dunbar 2003; Fitzgerald 2007).
However, IVD approved assays are not yet available for these pathogens.
DHPLC was used for the identification of genitourinary tract pathogens (Domann 2003) or Candida infections of the blood or GI tract (Goldenberg 2005). However, a parallel identification of more than 11 species based on the retention time only, was not yet shown.
When analysing complex faecal samples, Goldenberg et al. were not able to assign
33 individual peaks to species without subsequent sequencing due to day-to-day variations of retention times (Goldenberg 2007).
Mass spectrometry has also been shown to detect bacterial pathogens reliably (Gielen 2007;
Mandrell 2005). Very fast detection times of few minutes could be realized already by the MALDI Biotyper (Bruker Daltonics), which identifies bacteria on the spectra of 16S rRNA genes. Nevertheless, the method still relies on over-night culturing of the pathogen, which makes it a good supplementation of culture-based diagnostics without replacing it. The main disadvantages are the high cost of the instrumentation and the impossible integration in small point-of-care devices.
Sequencing methods have a high potential for fundamental research on the composition of the intestinal microbiota (chap. 1.2). Although several studies indicated that sequencing of the ribosomal genes could provide an effective diagnostic tool (Kolbert 1999), this method has not yet established in clinical diagnostic due to its user-intensive highly technical nature.
Although high-throughput sequencing has overcome limitations of Sanger sequencing, e.g.
low multiplexing capacity, it is still not suitable for most clinical applications due to high cost and required time, which may be overcome in future. Clinical applications of sequencing for pathogen identification are usually based on previous cultivation and/or 16S amplification or amplification of a structural gene (Hou 2008; Gleesen 2008; Justesen 2010; Luna 2007). It was assumed that the most probable application will be a combined one with established methods of pathogen identification and cultivation (Luna 2007). However, sequencing technology is underway to serve as full diagnostic tool.
Various nucleic acid hybridization-mediated methods have been developed to detect intestinal pathogens. A peptide nucleic acid-FISH (PNA-FISH) assay was published for Salmonella spp. identification from different samples (Almeida 2010). Although the assay showed high specificity, it was still lengthy, because an over-night enrichment step was required. A line-probe assay, which uses immobilized probes on a paper test strip, is available for the detection of mycobacteria (Innogenetics, Belgium) but not yet for intestinal pathogens.
Microarrays have huge potential for pathogen genotyping in clinical diagnostics, due to their ability to generate much genetic information in one experiment, in short time, and with relatively low technical effort. Several microarrays for cancer marker analysis and pathogen detection have been marketed yet. A concept for on-chip sequencing with overlapping probes, which cover the entire diagnostic sequence, was published for pathogen detection (Wilson 2002). Microarray technology and its application are described in detail in chapter 1.5.