Development and validation of an automated, microscopy-based method for enumeration of groups of intestinal bacteria

https://doi.org/10.1016/S0167-7012(99)00049-4Get rights and content

Abstract

An automated microscopy-based method using fluorescently labelled 16S rRNA-targeted oligonucleotide probes directed against the predominant groups of intestinal bacteria was developed and validated. The method makes use of the Leica 600HR image analysis system, a Kodak MegaPlus camera model 1.4 and a servo-controlled Leica DM/RXA ultra-violet microscope. Software for automated image acquisition and analysis was developed and tested. The performance of the method was validated using a set of four fluorescent oligonucleotide probes: a universal probe for the detection of all bacterial species, one probe specific for Bifidobacterium spp., a digenus-probe specific for Bacteroides spp. and Prevotella spp. and a trigenus-probe specific for Ruminococcus spp., Clostridium spp. and Eubacterium spp. A nucleic acid stain, 4′,6-diamidino-2-phenylindole (DAPI), was also included in the validation. In order to quantify the assay-error, one faecal sample was measured 20 times using each separate probe. Thereafter faecal samples of 20 different volunteers were measured following the same procedure in order to quantify the error due to individual-related differences in gut flora composition. It was concluded that the combination of automated microscopy and fluorescent whole-cell hybridisation enables distinction in gut flora-composition between volunteers at a significant level. With this method it is possible to process 48 faecal samples overnight, with coefficients of variation ranging from 0.07 to 0.30.

Introduction

The human colonic microflora is a complex bacterial ecosystem comprising more than 400 bacterial species (Moore and Holdeman, 1974). The composition and functioning of this microflora plays a key role in the protection of the host against several pathogenic conditions like: colonic cancer (Goldin and Gorbach, 1977), gastro-enteritis (Gorbach et al., 1987) and immunological disorders (Eckmann et al., 1995). Furthermore the colonic microflora plays a role in digestion and the production of vitamin K-12 (Savage, 1986). Because of the large influence of the gut microflora on the health of the host, accurate assessment of the composition of this bacterial ecosystem received much attention in both clinical microbiology and food microbiology. Historically, the assessment of the gutflora-composition is performed by means of selective culturing. Recently, new molecular strategies are being developed which are independent of culturability. One of these new methods which is proven to be especially useful for quantification purposes is fluorescent in situ hybridisation (FISH) (Manz et al., 1993) or whole-cell hybridisation. Quantification of positively hybridised bacteria in a faecal preparation is, in most cases, performed by means of visual counting-procedures. Because this latter method is a time-consuming process and heavily depends on the skills and experience of the technician, only moderate levels of accuracy are reached (Langendijk et al., 1995). In order to overcome these problems, it was decided that automation of the counting procedure was needed. In this article we describe: (i) a software-program for fully automated microscopic counting of whole-cell hybridised gut microflora and (ii) a validation of the automated counting procedure. The source code of the software was written using the QUIPPS-interpreter (which forms an integral part of the Quantimet-600HR measure-and-control software) supplied with the Leica Quantimet-600HR image analysis-system. The reason for this choice is that applications built in the QUIPPS-environment can be used to control the Leica DM/RXA ultra-violet microscope while, simultaneously, images acquired with the charge coupled device can be obtained and processed. With respect to the second aim of this study, a sequential validation-protocol was chosen in which faecal samples obtained from 20 different volunteers were processed and measured by the automated method using four different probes. The probes used in this study comprise: one probe specific for all bacteria present in the gut flora (Manz et al., 1993), one probe specific for Bifidobacterium spp. (Langendijk et al., 1995), a digenus-probe specific for Bacteroides spp. and Prevotella spp. (Manz et al., 1996) and a trigenus-probe specific for Ruminococcus spp., Clostridium spp. and Eubacterium spp. (Franks et al., 1998). To account for non-hybridizing bacteria a 4′,6-diamidino-2-phenylindole-stain (DAPI), which stains all DNA containing objects, was also included. The error introduced into the measurement by the process of slide manufacturing solely (i.e. the assay-error) was quantified by measuring one homogenised faecal sample 20 times using each separate probe. In order to account for person-related differences in flora-composition, faecal samples of 20 different volunteers were measured following the same procedure. The error in this dataset is subsequently corrected for the assay-error. Principal results of this study show that the quantification of the predominant groups of intestinal faecal bacteria using automated microscopy and in whole-cell hybridisation is faster and more accurate than visual counting procedures.

Section snippets

Probes

The characteristics of the probes used in this study are listed in Table 1. All probes are covalently linked with fluorescein iso-thiocyanate at the 3′-end. All probes have been validated for whole-cell application (Franks et al., 1998) in faecal material.

Volunteers

Twenty healthy human volunteers, eight male and 12 female, aged 26–53 years, entered this study after they had given written informed consent. None of the volunteers had received antimicrobial chemotherapy during the 2 months preceding the

Results and discussion

From Fig. 1 it can be seen that the input needed for a complete analysis can be entered using one form. As the form is designed in Excel 7.0, the initial parameter values are stored separately from the data produced by the measure- and control-software. This separate storage of data enhances the safety of these data considerably. Furthermore, errors in the input of the initial parameters can be retrieved and corrected easily, whilst a correct set of parameters can be used for repeated

Acknowledgements

This study was supported by a grant from the European Union (FAIR CT 97-3035).

References (12)

  • L. Eckmann et al.

    Intestinal epithelial cells as watchdogs for the natural immune system

    Trends Microbiol.

    (1995)
  • S.L. Gorbach et al.

    Successful treatment of relapsing Clostridium difficile colitis with Lactobacillus GG

    Lancet

    (1987)
  • A.H. Franks et al.

    Variations of bacterial populations in human faeces quantified by fluorescence whole-cell hybridisation with group-specific 16S rRNA-targeted oligonucleotide probes

    Appl. Environ. Microbiol.

    (1998)
  • B.R. Goldin et al.

    Alterations in faecal microflora enzymes related to diet, age, Lactobacillus supplements and dimethyl hydrazine

    Cancer

    (1977)
  • P.S. Langendijk et al.

    Quantitative fluorescence whole-cell hybridisation of Bifidobacterium spp. with genus-specific 16S rRNA-targeted probes and its application in faecal samples

    Appl. Environ. Microbiol.

    (1995)
  • Leica QWIN user Manual vol. 1 and 2, 1997. Leica Imaging Systems Ltd., Clifton Road, Cambridge, CB1 3QH,...
There are more references available in the full text version of this article.

Cited by (0)

View full text