Strains of the intestinal spirochaete Brachyspira pilosicoli attach to and aggregate erythrocytes



The anaerobic intestinal spirochaete Brachyspira pilosicoli colonizes the large intestine of various species of mammals and birds, where it may induce colitis. Strains of the spirochaete have also been isolated from the bloodstream of immunocompromised human patients and have been seen in liver sections, and a similar systemic spread was recently observed in experimentally infected chickens. Some other spirochaete species that may be present in blood attach to and aggregate erythrocytes, and this is believed to contribute to disease severity. The aim of the current study was to determine whether B. pilosicoli strains have the capacity to attach to and aggregate erythrocytes. Initially, four strains of B. pilosicoli were incubated with erythrocytes from sheep, cows, pigs, dogs, humans, chickens and geese, and were observed by phase-contrast microscopy. Only strain WesB attached, and this was only with erythrocytes from chickens and geese. Subsequently, six other strains of B. pilosicoli were tested just with goose erythrocytes, and five attached to and caused aggregation of the erythrocytes. Scanning and transmission electron microscopy demonstrated that spirochaetes abutted and apparently firmly attached to the erythrocyte membranes. Aggregation of erythrocytes by B. pilosicoli may contribute to disease severity in species that develop a spirochaetaemia.

Significance and Impact of the Study

The intestinal spirochaete Brachyspira pilosicoli has been isolated from the bloodstream of immunocompromised human patients, and spread to the liver has been reported in humans and in experimentally infected chickens. In this study, B. pilosicoli was shown to undergo attachment by one cell end to chicken and goose erythrocytes in vitro and to aggregate them. This activity has the potential to contribute to disease severity in avian and possibly other species that develop a spirochaetaemia and systemic spread. Avian erythrocytes may be useful for studying the mechanisms by which B. pilosicoli attaches to cells.


The anaerobic intestinal spirochaete Brachyspira pilosicoli colonizes the large intestine of many different species of birds and mammals, including human beings (Hampson et al. 2006). Within the large intestine, the highly motile B. pilosicoli characteristically penetrates the mucus layers, attaches by one cell end to the underlying enterocytes and then may induce localized inflammation and mucoid diarrhoea (Trott et al. 1995; Naresh et al. 2009). The spirochaete may also have invasive potential, as there have been a number of reports of it being isolated from the bloodstream of critically ill humans who have underlying immunosuppressive disorders (Fournié-amazouz et al. 1995; Trott et al. 1997; Kanavaki et al. 2002; Bait-Merabet et al. 2008; Prim et al. 2011). Spirochaetes assumed to be B. pilosicoli have also been seen in the liver of immunocompromised human patients with invasive colitis (Kostman et al. 1995). To date, spirochaetaemia with B. pilosicoli has not been reported in other animals or birds, but this may be because blood cultures are rarely taken from species other than human beings and the slow-growing spirochaete has specialized anaerobic growth requirement. It is of interest from a comparative perspective that other spirochaetes including Borrelia and Leptospira species have well-defined spirochaetaemic phases and that Borrelia burgdorferi and relapsing fever Borrelia species adhere to erythrocytes in vitro, with the latter causing erythrocyte aggregation in vivo that contributes to disease severity (Guo et al. 2009). Accordingly, we developed a hypothesis that B. pilosicoli strains also adhere to and aggregate erythrocytes, and we undertook this study with the aim of testing this.

Results and discussion

Kunkle's anaerobic broth was used to maintain viability of the spirochaetes in the assays. When erythrocytes obtained from different species were incubated in this broth, no lysis occurred and they retained a normal appearance when viewed under a phase-contrast microscope. Hence, the broth was considered satisfactory for use in the assays.

All 10 strains of B. pilosicoli that were used grew normally and were actively motile at the time of the assays. As hypothesized, at least some of the strains were shown to attach to and aggregate erythrocytes. In the first part of the study, of the four B. pilosicoli strains (WesB and Karlton from humans and 95/1000 and Cof-10 from pigs) that were incubated with erythrocytes from sheep, horses, cattle, pigs, dogs, chickens, geese and a human being, only strain WesB (originally isolated from an Australian Aboriginal child) adhered to erythrocytes and only to those from chickens and geese. Attachment was first observed after 4 h at an erythrocyte/spirochaete ratio of 1 : 1000, with scores per field of 1. At 6 h, the attachment was seen at both 1 : 100 and 1 : 1000 ratios, with scores of 2 at 1 : 100 and 3 at 1 : 1000. In the latter case, >5 spirochaetes were seen attached to individual erythrocytes (Fig. 1a), sometimes with groups of erythrocytes forming agglutinated clumps or aggregates bridged by the spirochaete cells (Fig. 1b).

Figure 1.

Panel of images showing different microscopic views of the association between B. pilosicoli strain WesB and goose erythrocytes. Panels (a) and (b), phase-contrast micrographs. Panel (a) shows spirochaete cells associated with the surface of a goose erythrocyte (bar = 5 μm), and panel (b) shows associated aggregation of erythrocytes bridged by spirochaetes (bar = 10 μm). Panels (c) and (d), scanning electron micrographs. Panel (c) shows end-on attachment of a single cell of WesB to the surface of a goose erythrocyte, with the tip of the spirochaete apparently invaginated into the erythrocyte cell membrane (bar = 2 μm). Panel (d) shows several WesB cells overlaying and attached to a goose erythrocyte (bar = 5 μm). Panels (e) and (f) show transmission electron micrographs. Panel (e) shows the end of a WesB cell abutting a cell membrane of a goose erythrocyte, but without penetrating it (bar = 200 nm). Panel (f) shows a cell of WesB lying on the surface of a goose erythrocyte (bar = 2 μm).

To determine whether other B. pilosicoli strains besides WesB could attach to erythrocytes, in the second part of the study, six more strains (BR81/80, 29/94 and HJ128/90 isolated from humans and CSP-1, 97.006.7 and 98.00.26 from chickens) were tested with goose erythrocytes. Similar attachments (scores ≥2 at 4 and 6 h) and aggregation of erythrocytes occurred with five of the strains, with only strain 29/94 isolated from human blood failing to attach. These strains were not tested with erythrocytes from other species, as the purpose of this study was only to determine whether the spirochaetes attached and whether it was a common phenomenon. Overall, 3/5 strains from humans (including 2/3 isolated from the blood), 3/3 from chickens and 0/2 from pigs attached to goose erythrocytes (60%). It was not clear why the other four strains did not attach, but it could have been due to the physiological state of the spirochaete cells or association with strain-specific differences in biological properties. B. pilosicoli is a genetically diverse species and does not appear to form groups that are specific to the species from which they are isolated (Neo et al. 2013), and indeed, the strains originally isolated from the blood of human patients have also been shown to be genetically diverse (Trott et al. 1997).

Viewing under the scanning electron microscope revealed end-on attachment to the erythrocyte surface by individual cells of B. pilosicoli strain WesB (Fig. 1c) that closely resembled the end-on attachment that occurs to colonocytes in natural and experimental colonic infections with B. pilosicoli (Duhamel 2001). The attachment was firm, as the preparations were fixed and processed for electron microscopy before viewing. Often, large numbers of attached spirochaetes were also observed lying on their sides over the surface of the cells (Fig. 1d). Using transmission electron microscopy, although end-on attachment of spirochaete cells was seen, no invagination into the cell membrane typical of that seen with colonic enterocytes was identified (Fig. 1e). This may have been attributable to the technical difficulties of sectioning erythrocytes such that a small area of invagination and attachment was located. Spirochaetes were also seen to have attached along part of their length to the outer membrane of erythrocyte, with a close apposition of the spirochaete and erythrocyte membranes (Fig. 1f).

The basis of the unusual polar attachment of B. pilosicoli cells to colonocytes is not understood, although it has been suggested that spirochaete outer membrane proteins could be involved in the process (Trott et al. 2001). In the case of strain WesB, the spirochaete attached only to nucleated erythrocytes from chickens and geese and not to cells from other species. This observed specificity in interaction and attachment by WesB may reflect differences in the outer glycocalyx region of the erythrocyte membranes, as avian erythrocyte membranes have less carbohydrate and different monosaccharide distributions compared with mammalian erythrocyte membranes (Gillis and Anastassiadis 1985). Subsequently, five of the six other spirochaete strains that were assayed with goose erythrocytes attached and caused the erythrocytes to aggregate. These strains were not tested with erythrocytes from other species, so the specificity for avian cells was not assessed. By analogy with other bacterial–eukaryotic cell interactions, the ability to attach to specific cells may reside with molecules such as lectins that are present in the bacterial outer membrane (Ofek et al. 2003); for example, Borrelia burgdorferi expresses a lectin activity that promotes agglutination of erythrocytes and bacterial attachment to glycosaminoglycans (Leong et al. 1995), while the relapsing fever spirochaetes bind to neolacto glycans to cause aggregation of human erythrocytes (Guo et al. 2009). In future work, it would be useful to investigate the nature of the putative protein–carbohydrate interactions between B. pilosicoli strains and erythrocytes from avian and possibly other species. Erythrocytes are easier to obtain and work with than intestinal cell lines such as the Caco-2 cells that have previously been used for this purpose (Naresh et al. 2009) and hence represent a convenient model to study interactions between the spirochaete and cells. To help identify components involved in the interaction, attachment assays could be undertaken following chemical removal of specific carbohydrates from the erythrocyte membranes or following addition of different sugars or lectins to the assay to determine whether these inhibit attachment.

The ability of certain B. pilosicoli strains to attach to and aggregate erythrocytes may become clinically relevant in cases of B. pilosicoli spirochaetaemia. For example, the relapsing fever Borrelia aggregate human erythrocytes, and this is thought to increase tissue invasiveness and haemorrhage and reduce blood flow in microcapillaries (Shamaei-Tousi et al. 1999, 2001). Although to date spirochaetaemia with B. pilosicoli has only been reported in humans, it seems likely that it could occur in other species such as pigs, chickens and poultry species that are quite commonly colonized with B. pilosicoli. In experimental pigs that were injected intravenously with a culture of B. pilosicoli, the spirochaete was isolated from pericardial fluid 5 days later, demonstrating its ability to survive at extra-intestinal sites (Hampson et al. 1998). In a recent experimental infection of chickens with B. pilosicoli, the spirochaete was isolated from the liver of two birds, with moderate to severe hepatic lipidosis being noted. In the spleen, lymphoid hyperplasia with proliferation of ellipsoid macrophages and germinal centres was also recorded, together with increased pyknotic and karyorrhectic debris (Mappley et al. 2013). Interestingly, in natural infections in immunocompromised human beings with invasive colitis, spirochaetes were also observed throughout the liver and cholestatic hepatitis was demonstrated (Kostman et al. 1995). These findings all point to the need for further investigation into spirochaetaemia and systemic spread of B. pilosicoli in humans and other species, together with analysis of erythrocyte aggregation and associated pathological sequelae.

Materials and methods

Brachyspira pilosicoli strains and culture conditions

Ten strains of B. pilosicoli were obtained as frozen stocks from the culture collection held at the Reference Centre for Intestinal Spirochaetes, School of Veterinary and Life Sciences, Murdoch University. The strains included five isolated from human beings (WesB, Karlton, BR81/80, 29/94 and HJ128/90), three from chickens (CSP-1, 97.006.7 and 98.00.26) and two from pigs (95/1000 and Cof-10). The strains were isolated in Australia from faecal samples, except for the last three human strains listed, which were isolated from the bloodstream of French patients (Trott et al. 1997). The strains were thawed and plated onto Trypticase soy agar containing 5% (v/v) defibrinated ovine blood. The plates were incubated for 7–10 days at 38°C in an anaerobic environment (94% N2/6% CO2) generated using anaerobic Gaspak Plus sachets (BBL, Becton, Dickinson and Company, Sydney, NSW, Australia). The purity of the cultures was examined by phase-contrast microscopy, and cells were propagated in modified Kunkle's prereduced anaerobic broth, containing 3·5% (v/v) foetal calf serum and 0·5% (v/v) newborn calf serum (Kunkle et al. 1986).

Source of blood and processing

Heparinized pooled blood specimens from healthy sheep, horses, cattle, pigs, dogs, chickens and geese were purchased from a commercial source (Animal Biological Collection and Delivery, Perth, WA, USA). Human blood was obtained from a healthy volunteer. The blood was held at 4°C and used within 24 h. It was centrifuged at 700 g for 10 min, and the erythrocyte pellet was resuspended in normal saline. Washing was repeated three times to remove remaining anticoagulant, and the final concentration of erythrocytes was adjusted to 1 × 107 per ml in normal saline. The erythrocytes were tested for their tolerance to Kunkle's broth by incubating them at 1 × 107 per ml in the medium for 12 h at 38°C.

Attachment assays

The attachment assays were conducted in sterile 48-well tissue culture plates with lids (Greiner Bio-One, Frickenhausen, Germany). In the first part of the study, B. pilosicoli strains WesB and Karlton isolated from humans and 95/1000 and Cof-10 from pigs were tested with all the erythrocyte types. In the second part of the study, human strains BR81/80, 29/94 and HJ128/90 (isolated from blood), and CSP-1, 97.006.7 and 98.00.26 from chickens were tested with goose erythrocytes. Six replicates were used for each spirochaete strain at each erythrocyte concentration for each incubation period. Each well contained a 1-ml volume comprising 107 actively motile spirochaetes in mid-log phase in Kunkle's broth and 106, 105 or 104 erythrocytes in the case of the first four strains tested. The plates were incubated on a rocking platform at 38°C for 2, 4 and 6 h. The other six strains were tested with goose erythrocytes at an erythrocyte-to-spirochaete ratio of 1 : 1000 and incubated for 6 h.

The contents of the wells were aspirated and an aliquot examined using a phase-contrast microscope at 400× and 1000× magnifications. A semi-quantitative scoring system for the extent of attachment at the different time points was used. An operator blinded to the origin of the samples examined 12 fields and scored these using a system where: 0 represented no attachment; 1 represented attachment of 1–2 spirochaetes to <10% of the erythrocytes; 2 represented attachment of 3–5 spirochaetes to >10–<20% of erythrocytes; and three represented attachment of >5 spirochaetes to >20% of erythrocytes. The presence of any erythrocyte aggregation was recorded.

Where attachment of strain WesB to goose erythrocytes was seen, the cells were processed for electron microscopy. The erythrocytes were transferred to a 2-ml Eppendorf tube, allowed to settle at 4°C for 6 h, and then, the fluid was carefully aspirated without disturbing the pellet. The pellet was resuspended and fixed overnight in 2·5% glutaraldehyde at 4°C. The glutaraldehyde was gently removed without disturbing the erythrocytes at the bottom of tube, and the pellet was washed five times with 0·07 mol l−1 Sorensen's buffer (three parts 0·01 mol l−1 Na2HPO4 and 1 part 0·01 mol l−1 KH2PO4). The cells were postfixed in 1% aqueous osmium tetroxide at 4°C for 1 h and washed three times with 70% ethanol before being dehydrated through an ethanol series. For scanning electron microscopy (SEM), the cells were resuspended in 100% alcohol and spread on an SEM specimen stub. They were sputter-coated with gold to a thickness of 90 nm in a Balzers sputter coater and examined using a Philips XL 20 scanning electron microscope. For transmission electron microscopy, the dehydrated cells were processed for infiltration with propylene oxide (two changes over 20 min), then with a propylene/resin mix (60/40) for 1 h at 4°C and finally with absolute resin on a rotary mixer at 25°C overnight. The cells were embedded with pure resin at 60°C for 24 h, and 90-nm sections were cut with an ultra-microtome and mounted on carbon-coated grids. The grids were stained with freshly prepared uranyl acetate and lead citrate and were examined using a Philips 1 CM-100 transmission electron microscope.


Dr Ram Naresh was in receipt of a postgraduate scholarship from Murdoch University. The authors thank Dr Nyree Phillips and Dr Tom La for technical assistance.

Conflict of Interest

The authors declare no conflict of interest.