Fimbria-mediated bacterial adhesion to human oral epithelium

Abstract

Oral mucosa biopsies and saliva samples from 12 individuals were processed for transmission (TEM) and scanning (SEM) electron microscopy with and without ruthenium red staining. Additionally performed microbiological estimations indicated in all bacteriological samples a facultative pathogenic flora. SEM and TEM investigation showed a diverse bacterial flora attached to the mucosal surface. Fimbriae comprising the glycocalyx and enabling bacterial attachment to the epithelial cells could be clearly visualised by ruthenium red. The only mode of bacterial attachment to the oral mucosa detected in the present investigation was fimbria-mediated adhesion and co-adhesion. The fimbria-mediated adhesion enables the bacterial persistence in the oral cavity and is the first step in the bacterial colonisation process.

Bacterial adhesion, Ruthenium red staining, Glycocalyx, Fimbria, Adhesin

1 Introduction

Bacterial adhesion to the oral mucosa has been demonstrated by in vitro studies [1] as well as by scanning electron microscopy (SEM) analysis of in vivo mucosal samples [2,3]. The molecular substrates of bacterial adhesion are adhesins located on the bacterial surface. Adhesins exist in fimbrial and non-fimbrial forms [4,5]. The first form of adhesins is composed of proteins and carbohydrates in the form of fimbriae, which are semi-selectively stainable by ruthenium red [1,6–8]. The entirety of fimbriae forms an additional but not obligatory polysaccharide-rich surface layer denoted as glycocalyx [6]. The glycocalyx is peripherally located to the outer membrane in Gram-negative bacteria and to the peptidoglycan in Gram-positive ones. Bacteria can easily be promoted to form fimbriae in vitro by addition of a high concentration of carbohydrates [9,10]. The occurrence of bacterial fimbriae in vitro has been visualised in transmission electron microscopy (TEM) by ruthenium red staining [1,6,11,12]. However, TEM data concerning the mode of bacterial adhesion in the human oral cavity in vivo are lacking.

The present electron microscopic study examines the mode of bacterial attachment to the oral mucosa in vivo as well the bacterial coaggregation in saliva using ruthenium red staining for detection of fibrillar bacterial adhesins [1,6–8]. The purpose of this study was to prove the occurrence of fimbria-mediated adhesion during bacterial colonisation of the human oral mucosa.

2 Materials and methods

2.1 Oral examination and microbiological investigations

As the palatal mucosa of patients with denture stomatitis shows a multitude of attached bacteria [3], such patients were selected in the present work. Denture stomatitis was diagnosed by routinely performed oral inspections in 40 subjects. None of the patients was HIV-positive or has been treated with immunosuppressive drugs, radiotherapy or antibiotics within the last 6 months. The criteria suggested by Newton [13] were used for diagnosis of denture stomatitis. The Candida density was estimated as described previously [14], and only 12 patients in the age-bracket between 28 and 73 years (mean: 56 years) with denture stomatitis type II or III (according to Newton) and moderate Candida density were further included in the present investigation. The study was approved by the local ethics committee, and informed written consent was obtained from all patients. Samples for microbiological investigation were taken with swabs Biotest Transportsystem (Biotest AG, Dreieich, Germany) from about 1 cm² of the denture-bearing mucosa. Sampling took place between 09:00 h and 12:00 h, at least 2 h after the patient's breakfast. Each swab was plated for culturing under aerobic conditions on Columbia sheep blood agar (Biotest AG, Dreieich, Germany), Columbia CNA agar (Becton Dickinson, Franklin Lakes, NJ, USA) and MacConkey agar (Biotest AG, Dreieich, Germany), and incubated at 37°C for 18 h. For anaerobic culturing, the swabs were plated on Schaedler agar (Biotest AG, Dreieich, Germany) and Schaedler KV agar (Becton Dickinson, Franklin Lakes, NJ USA) and incubated at 37°C for 7 days. Microbiological differentiation was performed as follows: viridans streptococci were identified by colony morphology, α-haemolysis, and use of commercially available optochin test. Non-haemolytic streptococci were identified by colony morphology, lack of haemolysis and optochin test. Enterococci were identified by colony morphology and antibiotic susceptibility. Staphylococci were identified by colony morphology, colony colour, and testing of coagulase activity with Staphaurex (Abbott Laboratories, Abbott Park, IL, USA). Hemophilus spp. were identified by colony morphology, Gram-staining, and biochemical identification procedures API NH (Biomerieux SA, Marcy l'Etoile, France). Neisseria spp. were identified by colony morphology, Gram-staining and positive oxidase reaction. Gram-negative bacilli (Escherichia coli, Klebsiella, Enterobacter etc.) as well as Lactobacillus casei were identified by biochemical identification Crystal system (Becton Dickinson, Franklin Lakes, NJ USA).

For semi-quantitative estimation of the bacterial density the following counting procedure was used: Negative bacterial growth was characterised by 0–4 colonies per agar plate; low bacterial density was defined by the growth of at least five colonies per agar plate; moderate bacterial density was defined by the growth over half of the agar plate; high bacterial density was defined by confluent bacterial growth.

2.2 Microscopic investigation

Biopsies (approximately 6×10 mm) were taken from the inflamed denture-bearing mucosa of each patient under regional block anaesthesia. Each biopsy was washed in 50 ml saline by gently swaying for 20 s and subsequently divided into five parts (one part for light microscopy measuring ca. 24 mm², and four parts measuring ca. 9 mm² for electron microscopy). The 24 mm² mucosa samples were fixed in 10% formaldehyde and routinely processed for light microscopy and stained with haematoxylin–eosin. Two of the four samples from each patient proposed for electron microscopy were fixed with 1.2% glutaraldehyde (buffered at pH 6.5 with 0.1 M sodium cacodylate) for 2 h at 4°C. Postfixation of these control samples was performed with 1% osmium tetroxide (buffered at pH 6.5 with 0.1 M sodium cacodylate) for 2 h at 4°C. The other two 9-mm² mucosa samples from each patient were fixed and postfixed by the fixatives mentioned before with addition of 0.05% ruthenium red in each case.

For SEM investigation one control and one ruthenium red-stained specimen/samples of each patient were dehydrated in ascending series of ethyl alcohol, critical-point-dried in a drying equipment CDP 030 (BAL-TEC, Balzers, Liechtenstein) and subsequently sputtered with gold (ca. 10 nm). The specimens were examined in a scanning electron microscope Cambridge Stereoscan S250 (Cambridge Instruments Ltd, Cambridge, UK) operating at 20 kV.

Immediately after the microbiological sampling, the patients were also requested to rinse the mouth with 5 ml saline for 30 s and to spit out the mouth rinse into a sterile glass. The samples were divided into two parts, which were fixed and postfixed for TEM investigation with and without additional use of ruthenium red as described above for the biopsies. After centrifugation at 4000×g for 10 min the sediment was ready for embedding.

All specimens assigned for TEM analysis were routinely embedded in Epon 812. Ultrathin sections were cut on an ultramicrotome Reichert Ultracut S (Optische Werke C. Reichert, Vienna, Austria), contrasted with LKB 2168 Ultrostainer (LKB Produkter AB, Bromma, Sweden) and examined in a transmission electron microscope LEO EM 910 (LEO Elektronenmikroskopie Ltd, Oberkochen, Germany) operating at 80 kV.

3 Results

3.1 Pathological and microbiological examinations

Routinely histopathological examinations were performed to exclude other pathological alteration of the mucosa biopsies than stomatitis. However, in all samples only stomatitis was diagnosed.

Results of the microbiological examination are summarised in Table 1.

1

Results of microbiological investigation: bacterial species and density in patients 1–12

Patient	Age	Viridans streptococci	Neisseria spp.	Haemophilus spp.	Anhaemolysing streptococci	Acinetobacter spp.	Enterobacter spp.	Citrobacter spp.	Klebsiella spp.	Prevotella melaninogenica	L. casei
1	54							+++
2	54	++	+		++
3	36						+++
4	73	++	+
5	35	+++	+	+
6	60	+++		+					+++
7	72						+++
8	62							+++
9	71						+		+++
10	53	+++
11	73	+++				++
12	28									+	++

+=low bacterial density; ++=moderate bacterial density; +++=high bacterial density.

3.2 Electron microscopy

Candida blastospores observed by SEM were sparse in all biopsies. Less than one blastospore per microscope screen at magnification ×500 could be detected. Groups of diverse bacteria colonising the mucosal surface as well as non-colonised areas of mucosa were observed by SEM analysis. Higher resolution indicated smooth bacterial surfaces in the samples contrasted without ruthenium red (Fig. 1). In contrast, the bacterial surface contrasted by ruthenium red appeared granulated (Fig. 2).

1

Specimen stained without ruthenium red: SEM micrograph of mucosal surface with attached cocci-like bacteria. The bacteria display a smooth outside.

2

Specimen additionally stained with ruthenium red: SEM view of mucosal surface with attached cocci-like bacteria. The bacteria show a granular-structured surface.

No Candida cells could be observed in TEM. A large quantity of bacteria, often occurring in groups, was displayed on the mucosal surface. Cocci and cocci-like bacteria (transversal-sectioned rods are not distinguishable from cocci in TEM) fixed in the absence of ruthenium red showed a layer of faint fibrils exterior to the bacterial wall (Fig. 3). In the specimens without ruthenium red staining no close contact between bacteria and eukaryotic membrane could be detected. However, bacterial attachment to the epithelial cells by fimbriae comprising a glycocalyx could be clearly visualised in mucosa samples additionally stained with ruthenium red (Figs. 4–6). Fimbria-mediated bacterial attachment to the mucosal cell surface was the only pattern of bacterial adherence observed by TEM analysis. Another, indirectly way of bacterial attachment to the epithelium was the co-adhesion between two ultra-morphologically different bacterial cells (Figs. 5 and 6). Independently of the number of wall layers, cocci and cocci-like bacteria displayed a electron-dense glycocalyx (capsule), (Figs. 5 and 6). In contrast, the rod-shaped bacteria exhibited a loose glycocalyx displaying only sparse hair-like fimbriae with varying length. Rarely, (diplo)cocci with a glycocalyx composed of separate radially arranged hair-like fimbriae were monitored (Fig. 6). Although adhered bacteria harboured in invaginations of the epithelial membrane could be detected (Fig. 7), there was no evidence of bacterial penetration into epithelial cells.

3

×28 480 Control staining. Cocci-like bacterium at some narrow distance from epithelial cell membrane. The bacterium exhibits a faint additional layer outside the bacterial wall.

4

×22 250 Ruthenium red staining. Two cocci-like bacteria with different large glycocalyces adhered to the epithelial cell surface.

5

×28 480 Ruthenium red staining. A dividing coccus with a monolayer wall adhered to a microvilli-like epithelial protrusion. The glycocalyx of the coccus is interwoven with that of another bacterium, probably adhering to the coccus. The outer layer of the epithelial membrane, invisible without ruthenium red staining, represents probably the glycosylated membrane components functioning as receptors for the adhesins.

6

×36 490 Ruthenium red staining. A dividing coccus or diplococci with a multiple layer wall adhered via glycocalyx to the epithelial microvilli-like protrusion. The glycocalyx is comprised by a multitude of not branched hair-like fimbriae. A rod-shaped bacterium with a loosely arranged glycocalyx and a monolayer wall seems to adhere to both the (diplo)coccus and the epithelial protrusion.

7

×36 490 Ruthenium red staining. Cocci-like bacteria embedded in an invagination of the epithelial cell.

The non-attached bacteria observed in salivary mouth rinse (Fig. 8) displayed a similar ultrastructural pattern of glycocalyx (Figs. 9 and 10) as those attached to the mucosa when contrasted by ruthenium red. The co-adhesion of bacteria in saliva leads to formation of a three-dimensional network of slime (Figs. 9 and 10).

8

×8900 Control staining. Cocci-like and rod bacteria in mouth rinse. Some of them show a faint additional layer outside the bacterial wall.

9

×22 250 Ruthenium red staining. Rods in salivary mouth rinse exhibiting a slime-forming three-dimensional fibrillar network.

10

×9790 Ruthenium red staining. Cocci-like bacteria in salivary mouth rinse exhibiting dense glycocalyx (possibly capsule) and a slime-forming three-dimensional fibrillar network.

4 Discussion

The present TEM investigations, for the first time, demonstrates that bacterial adherence to the oral mucosa in humans takes place via fimbria-mediated adhesion. For detection of bacterial fimbriae specimens were stained by ruthenium red. The ability of the cationic ruthenium red to enhance staining of acidic glycosylated structures at the ultrastructural level is widely used for visualisation of glycocalyces. The action of ruthenium red involves the formation of an electron-dense deposit of OsO₄ at the location of polysaccharides [7].

The supplementary staining by ruthenium red hinders determination of the Gram properties of bacteria by ultrastructural features. The Gram-negative bacteria are delineated in electron micrographs by the outer and inner membrane, in contrast to the Gram-positive bacteria possessing an inner membrane only [6,15]. Appearance of one or more layers within the bacterial envelope can be the consequence of both the fixation method or the environment-related distribution of wall components, and thus cannot be a valid criterion for distinction whether a bacterium is Gram-negative or Gram-positive [15].

The present results provide ultrastructural evidence that bacteria adhere in vivo to the human oral epithelial cell surface by fibrillar adhesins stainable with ruthenium red. The appearance of glycocalyces forming a fibrillar network in salivary mouth rinse as well as the appearance of bacteria adhering to the oral epithelium and to each other via glycocalyx strongly suggests the possibility that the matrix of the growing biofilm consists predominantly of bacterial glycocalyces.

The fimbrial adhesins are an integral part of the bacterial glycocalyx [[1,11,12]. Most bacteria can display a number of adhesins: e.g. for E. coli at least nine adhesins [4], for Bordetella pertussis at least seven [4], and for streptococci at least six adhesins [11] have been described. The adhesion of bacteria to eukaryotic cells is a result of specific reactions between bacterial adhesins and receptors expressed on the membranes of the eukaryotic cells [4,5,11,16–18]. The granulated surface of the bacterial glycocalyces displayed in SEM micrographs might multiply the contacting contact area between bacteria and host, thus amplifying the adhesin–receptor interaction. Attachment to the palatal mucosa appears to be the first step in the bacterial colonisation process. Antibodies to adhesins are associated with protection from colonisation [4,12]. However, the great diversity of bacterial genera found by microbiological investigation of smear as well the inability to differentiate the bacteria by electron microscopy did not allow to distinguish which species adhered directly to the epithelium and which ones via co-adhesion. The host reactions, which are triggered by the attachment of bacteria, have not been exactly defined, yet. Since some of the estimated bacteria, viridans streptococci [19], Haemophilus spp. [20] and Klebsiella spp. [20], are able to stimulate the production of proinflammatory cytokines, a bacterial involvement in development of denture stomatitis might be possible. Even non-pathogenic bacteria are able to elicit differential cytokine response by epithelial cells [21]. With certainty, however, Candida is a pathogenetic factor in formation of denture stomatitis [14,22,23].

In summary, the present study has shown that bacterial fimbria-mediated adhesion to epithelium as well as the fimbria-mediated co-adhesion are the only observed modes of bacterial attachment to the oral mucosa and consequently the primary factors enabling bacterial persistence and colonisation process.

Acknowledgements

The authors thank Mrs. Adelina Vitkov for secretarial assistance.