DNA microarray-based identification of bacterial and fungal pathogens in bloodstream infections

Abstract

The accurate and rapid identification of pathogens in blood is a major challenge in clinical pathogen diagnostics because of the high mortality of sepsis. Here we report the development of DNA microarray for the identification of pathogens causing bloodstream infections. Species-specific and bacteria- and fungi-broad-ranged probes were designed to identify 50 bacteria and 7 fungi. The specificities and sensitivities of the selected probes were successfully validated by applying reference strains. To assess the performance of the DNA microarray in a clinical setting, blind tests were performed using 112 blood culture specimens that showed preliminary presence of pathogenic microorganisms by culture-based method, resulting in the correct identification of pathogens in 104 samples showing the sensitivity of 93%. In addition, closely-related species could be discriminated by the distinct hybridization patterns. This DNA microarray-based pathogen diagnosis takes approximately 10 h starting from a positive blood culture, considerably reducing time required to sufficiently identify pathogens by subsequent agar-culture and biochemical tests which requires altogether at least 1–3 days. Also, the amount of sample required for the identification of pathogens is much less than that required for biochemical assays. Thus, the DNA microarray reported here should be useful for the effective identification of microbial pathogens in blood cultures from septicemic patients.

Introduction

Despite continued advances in therapeutics, severe sepsis still has an unacceptably high mortality of 20–50% [1]. To achieve the best prognosis for a bloodstream infection, treatment must begin within 6 h after the onset of symptoms to prevent the infection from progressing to severe sepsis [2]. Thus, early and accurate pathogen diagnosis is the most important factor for successful administration of the appropriate antimicrobial agents [3].

Conventional systems for diagnosing sepsis generally involve the cultivation of blood in broth culture bottles that are continuously monitored for the metabolites (e.g. CO₂) produced during the growth of microbes. The conventional blood culture requires currently 1–5 days for the preliminary detection of the presence of most organisms or longer for fastidious organisms [4]. Another limitation of conventional blood culture techniques is the low sensitivity due to several factors, including inadequate blood volumes, the presence of antimicrobial agent in the patient sera, and fastidious growth characteristics of certain species that are not satisfied in standard blood culture systems [4], [5].

Once the blood culture shows preliminarily positive result, routine microbiological tests involving Gram staining is carried out, which is followed by various biochemical tests. The biochemical tests that can monitor the expression of metabolic activities and morphological features have been developed and widely employed for the identification of microbial pathogens at the genus or species level [6], [7], [8], [9]. However, these systems have several potential problems. First, the low specificity of pathogen identification requires additional tests to finally determine the species [6], [8]. Second, several tests must be carried out simultaneously in order to enhance diagnostic accuracy and to discriminate among the species because one kind of biochemical assay allows detection of only a limited number of species which makes them less cost and time efficient [8]. Third, biochemical assays take 1–3 days to identify pathogens because most of them require subcultures of blood culture samples showing positive signals. Subcultivation of fastidious microorganisms is of particular problem as it prolongs the assay time to obtain the results. Fourth, phenotypic tests can yield non-reproducible results in repeated analyses [6], [8]. These limitations mentioned above can cause ambiguous identification of pathogens, making doctors administer inadequate and broad-spectrum antimicrobial drugs for the treatment, which can consequently lead to the emergence of antimicrobial-resistant pathogens and waste of medical care cost. Therefore, it is important to develop a rapid, sensitive, and specific method for identifying bacterial and fungal pathogens in order to promote more timely, appropriate and accurate antimicrobial therapy [9].

Molecular diagnostic methods, such as multiplex PCR and real-time PCR, have considerably enhanced the speed and sensitivity of detection and identification of microbial pathogens in blood samples [3], [10], [11], [12]. Although some advanced PCR assays, such as SeptiFast^® assay, allow detection of pathogens in half a day without blood cultivation, only a limited number of pathogens can be detected. For example, the most popular SeptiFast^® assay cannot diagnose some important pathogens, such as Neisseria meningitides [13]. Despite of high sensitivity of PCR assay in general, the sensitivity tends to be inversely related to the specificity due to some nonspecific products, making it difficult to correctly identify true pathogens.

DNA microarrays, in which many DNA probes are immobilized at high-density, have been used for high-throughput identification of pathogens in positive blood culture samples [14], [15], [16]. The sequence-specific hybridization step in the DNA microarray-based assay enables the unambiguous detection of the pathogens by removing the artifacts resulting from the PCR amplification only. For the high-throughput identification of pathogens, the probes have been designed from various target genes such as ribosomal DNA, virulence factors, antibiotic resistance-relevant genes, and genus- and species-specific genes [14], [15], [16], [17], [18]. These specific probes allowed unambiguous detection of the corresponding pathogenic microorganisms. Also, the hybridization signal patterns, rather than the specific hybridization to the corresponding probes, in the DNA microarray have been shown to be effective in discriminating the closely-related species [15], [16]. Although several DNA microarrays have been developed for the detection of various blood-borne pathogens with improved sensitivity, specificity, reliability and high-throughput capability, they are currently limited to identifying up to 25 microbial pathogens [16]. Thus, it is necessary to develop a DNA microarray that is capable of detecting more diverse pathogens of emerging clinical importance.

In this study, a highly sensitive and specific DNA microarray-based diagnostic method is developed for the identification of 50 bacterial and 7 fungal pathogens that are clinically important. Blind tests using this DNA microarray allowed the successful diagnosis of microbial pathogens in 104 out of 112 clinical blood culture samples.