The ability of relapsing fever Borrelia species to survive in chicken serum in vitro and its correlation to interaction with serum components reactive to a polyclonal anti-complement regulator factor H antibody

TL;DR

This study shows that relapsing fever Borrelia bacteria can survive in chicken serum and interact with specific serum components.

Contribution

The study demonstrates that Borrelia species without known avian associations can survive in chicken serum and bind serum components reactive to anti-factor H antibodies.

Findings

All tested Borrelia strains survived in chicken serum in vitro.

Borrelia species bound serum components reactive to anti-factor H antibodies.

Serum survival did not depend on binding to these serum components.

Abstract

Bacteria belonging to the relapsing fever-causing genus Borrelia, are important causes of zoonotic veterinary and human infections. The genus contains both species with a specific avian association and species with a suggested wider host range, including avian hosts. Despite this, existing studies exploring the Borrelia-avian interaction are relatively few. In this short communication, we report on the in vitro survival of eight strains of relapsing fever from seven species in chicken serum and its correlation to bacterial binding of serum components reactive to a polyclonal anti-complement regulator factor H antibody. All tested relapsing fever strains exhibited the capability to survive in chicken serum in vitro. Components of chicken serum reactive to polyclonal anti-factor H antibodies were found to bind to the surface of spirochetes and to whole cell lysates, but this was not mandatory for serum survival.

Image 1

•Relapsing fever Borrelia isolates survive in complement active chicken serum in-vitro.•Borrelia species with no known avian association survive in chicken serum.•Borrelia spp. bind components of chicken serum reactive to polyclonal anti-factor H antibody.

Relapsing fever Borrelia isolates survive in complement active chicken serum in-vitro.

Borrelia species with no known avian association survive in chicken serum.

Borrelia spp. bind components of chicken serum reactive to polyclonal anti-factor H antibody.

Figures

Serum survival after incubation in 50% (v/v) chicken serum (CS) and 50% (v/v) heat-inactivated chicken serum (HICS) with 95% CI indicated.Fig. 1

Whole cell lysate – chicken serum ELISA. Signal detected with polyclonal anti-FH antibody. Boxplot indicating range, with the middle line in the box representing the mean. Statistical significance of differences between individual samples and the negative control was assessed *via* Student’s t-test (∗*P* < 0.05).Fig. 2Fig. 3Binding of chicken serum components reactive to polyclonal anti-FH antibody to the surface of viable spirochetes. Purified human FH served as a positive control. Eluted proteins from human serum components bound to *B. hermsii* HS1 served as a positive binding control.Fig.

Full text

Introduction

1

The family Borreliaceae includes two major zoonotic genera: Borrelia (species causing relapsing fever, RF) and Borreliella (species causing Lyme disease, LD). Species of both genera are transmitted by arthropod vectors and sustained in nature through diverse animal reservoirs (Barbour and Gupta, 2021). LD species commonly infect dogs and horses, with sporadic cases in cats and ruminants (Fritz and Kjemtrup, 2003), while RF species have a broader host range, including livestock, pets, and wildlife (Elelu, 2018). The genus Borrelia also contains a true anthropophilic species, Borrelia recurrentis, for which no animal host has been identified. Borrelia anserina, an RF species, causes avian borreliosis and is strictly host-adapted to birds, possibly due to genomic reduction (Elbir et al., 2017). Other Borrelia species have been shown to infect birds and mammals experimentally, but their avian interactions are understudied (Thomas et al., 2002; Schwan et al., 2007). Interestingly, a study in East Africa showed signs of natural infection with Borrelia duttoni in domesticated chickens, indicating this avian species as host (McCall et al., 2007). Understanding host specificity is essential for clarifying transmission dynamics and improving epidemiological modelling, as it determines reservoir competence and zoonotic risk across ecosystems.

To survive within a host, bacteria and other pathogens must evade various components of the host immune system and specifically evade the complement system (CS). The CS is a key component of the innate immune response and is composed of more than 40 proteins present in serum and on cell surfaces. Upon activation through any of three pathways, these proteins act in a cascade in which complement components are cleaved and assemble into convertases that subsequently activate downstream components. Three pathways may activate the CS: the classical pathway (CP); the lectin pathway (LP); and the alternative pathway (AP). The CP is activated when the component C1q binds to activating structures, for example antibodies bound in immune complexes or acute phase reactants (Roumenina et al., 2005). The LP operates through mechanisms similar to the CP but is activated differently through the binding of mannose-binding lectins and ficolins to specific carbohydrate structures on pathogen surfaces (Garred et al., 2016). The AP is most commonly activated via spontaneous hydrolysis of the CS subunit C3b, but can also be activated by microbial surfaces, lipopolysaccharides and cellular debris (Harboe and Mollnes, 2008). All three pathways converge on a shared terminal pathway, resulting in the formation of the membrane attack complex, a porin-like structure that inserts into target cell membranes and induces cell lysis. To prevent damage from the CS on host cells, the reaction is under the control of several fluid and cell-bound regulators that regulate different points of the cascade. A key negative regulator of the AP is the glycoprotein factor H (FH) that inhibits activation of C3 to C3b and facilitates the proteolysis of formed C3b (Merle et al., 2015). Several pathogens have developed virulence mechanisms for host-specific survival in blood by binding FH to the microbial surface (Józsi, 2017). Interactions between RF Borrelia spp. and the complement system have been studied mainly in human serum, where RF species are known to bind soluble and membrane-bound complement regulators and even plasminogen, which can modulate complement activity. Factor H binding in five RF species has been shown to occur via the proteins FhbA1, FHBP19, BpcA, CbiA, and HcpA ((Röttgerding and Kraiczy, 2020).

However, little is known about these mechanisms in avian hosts. In this study, we assessed whether complement evasion in RF Borrelia spp. correlates with the binding of components in chicken serum reactive to a polyclonal anti-FH antibody.

Materials and methods

2

Bacterial cultivation

2.1

The bacterial strains tested in this study are summarized in Table 1. Tested RF Borrelia strains were cultivated in Barbour-Stoner-Kelly II (BSKII) medium supplemented with 10% (v/v) rabbit serum (v/v), 1.5% gelatin (w/v) and incubated at 37 °C.Table 1. Strains used in the study.Table 1. Bacterial strainOriginPassageReferenceB. hermsii HS1Ornithodoros hermsi, USA, ATCC 35209Passages < 5 after isolation from mouseThompson et al. (1969)B. anserina EsWhite Leghorn Chicken, USAPassage 7DaMassa and Adler (1979)B. hispanica CR1Ornithodoros erraticus, geographical location uncertainPassage < 5 after isolation from guinea pigToledo et al. (2010)B. persica No14Ornithodoros papillipes, UzbekistanPassage < 5 after isolation from mouseElbir et al. (2014)B. turicatae 91E135Ornithodoros turicata, USAPassage < 5 after isolation from mouseSchwan et al. (2005)B. duttonii CR2AUncertainPassage < 5 after isolation from mouseToledo et al. (2010)B. duttonii 1120K3Ornithodoros moubata, CongoPassage < 5 after isolation from mouseToledo et al. (2010)B. recurrentis A1Human blood, EthiopiaPassage < 10Cutler et al. (1994)

Serum testing

2.2

Chicken serum (Agrisera, Umeå, Sweden) was tested for functional complement activity before being used in serum survival according to a modified previously described protocol (Skeeles et al., 1980; Roumenina, 2021). In brief, sheep red blood cells (ShRBC) and rabbit red blood cells (RaRBC) (Hataunalab, Bro, Sweden) were washed three times in in-house prepared gelatin veronal buffer (GVB^0^; 5 mM veronal, 145 mM NaCl, 0.1% gelatin, pH 7.3). Washed ShRBC were sensitized by incubation in rabbit anti-sheep red blood cell stroma serum (S1389, Sigma-Aldrich, St. Louis, USA) diluted 1:100 in GVB^0^ for 30 min at room temperature (RT). Sensitized red blood cells were washed once in GVB++ (5 mM veronal, 145 mM NaCl, 0.1% gelatin, 0.15 mM CaCl_2_, 0.5 mM MgCl_2_, pH 7.3). Serum to be tested was diluted with 1:1 volume GVB++. GVB++ containing sensitized ShRBC corresponding to one-third of the diluted serum volume was added. Washed unsensitized RaRBC was also incubated in diluted serum and included in the protocol below.

The erythrocyte-serum mixtures were incubated at 40 °C for 30 min. After incubation, ice-cold GVBE (GVB^0^ supplemented with 10 mM EDTA) 1:1 volume was added. Samples were centrifuged at 1700× g for 5 min. The supernatants were collected, and the absorbance at 414 nm measured. Hemolytic capability of tested sera was calculated by dividing observed absorbance from serum samples by the absorbance from the control sample containing erythrocytes and Milli-Q water (Merck Millipore, Burlington, USA). Sensitized erythrocytes incubated in heat-inactivated serum (serum heated to 56 °C to abolish complement activity) or only GVB++ were included as controls (Supplementary Fig. S1).

Bacterial survival in chicken serum

2.3

Tested bacteria strains were grown under the conditions described above until judged to be at logarithmic growth phase. Cultured bacteria were counted in a Petroff-Hausser chamber and diluted with fresh BSKII medium to a concentration of 1 × 10^6^ bacteria/ml. Complement active chicken serum was added at an equal volume to the diluted bacterial cultures in triplicate samples. As controls, one sample was diluted with heat-inactivated serum. All assays were done in 50% serum. No serum was added to another sample, and dilution was performed in fresh BSKII medium. The samples were then incubated at 40 °C for 4 h. An incubation temperature of 40 °C was used to match the assumed body temperature of the serum source. The number of living spirochetes was assessed by observing motility in five fields of view in a darkfield microscope at the incubation start, at 1 h and 4 h. By dividing the number of motile bacteria at each time point by the number of motile bacteria at incubation start, a viability percentage was calculated. Experiments were repeated twice for all isolates except for B. persica No14, where cultivation difficulties only allowed for two separate experiments. Despite multiple attempts, the B. persica No14 reached a sufficiently dense culture for downstream experiments on only two occasions. We were unable to determine a definite cause, although certain RF species are known to be challenging to cultivate in vitro.

Antibodies

2.4

Three commercially procured polyclonal anti-human FH antibodies were tested for reactivity against FH in mouse, chicken and human serum via non-reducing western blot (Supplementary Fig. S2). A sheep anti-human FH antibody (Lot# ab8842, Abcam, Cambridge, UK) displayed reactivity against all species and was used in subsequent experiments.

Whole-cell lysate ELISA

2.5

Bacteria from tested strains grown under conditions above to mid-logarithmic phase were used for whole cell lysate ELISA. Mid-logarithmic phase bacterial culture was spun at 6000× g for 20 min at RT and washed twice in PBS 5 mM MgCl_2_. Washed bacteria were resuspended in PBS 1/100 volume of the original culture medium volume. Samples were freeze-thawed three times in −80 °C/RT. Finally, the samples were sonicated and lysed cells spun at 9500× g at 4 °C for 30 min. The supernatant was recovered and protein concentration measured using Bio-Rad Protein Assay (Bio-Rad). Whole cell lysate suspension was diluted in 100 mM bicarbonate buffer (pH 9.5) to a concentration of 10 μg/ml, and 50 μl was added to wells of high-binding ELISA plates (Sarstedt) and allowed to coat overnight at 4 °C. Three wells on each plate were coated in 10% (w/v) BSA (negative reagent control) or 50% (v/v) chicken serum (positive reagent control), all diluted in 100 mM bicarbonate buffer. All wells were blocked in 5% (w/v) skimmed milk diluted in PBS for 1 h at RT. Plates were washed three times in PBS, and 50 μl of 50% (v/v) chicken serum diluted in PBS added, and plates incubated at RT for 60 min. Three wells coated with bacterial lysate had no serum added (negative control), and three wells containing lysate from a strain with known ability to bind FH of human origin (B. hermsii HS1) were incubated with 50% (v/v) human serum (positive control).

Samples were washed and incubated with polyclonal sheep anti-FH (Lot# ab8842, Abcam) 1:500 (v/v) diluted in 1% (w/v) skimmed milk-PBS for 60 min at RT. Samples were washed and incubated with HRP conjugated donkey-anti sheep IgG (Lot# sc2473, Santa Cruz Biotechnology, Texas, USA) 1:5000 (v/v) diluted in 1% (w/v) skimmed milk-PBS and incubated for 60 min at RT. After a final wash, TMB substrate (ThermoFischer, Waltham, USA) was added, and plates were incubated according to the manufacturer’s instructions. After the addition of TMB stop solution, the OD_450_ was determined. Experiments were performed three times in triplicates for all samples except for B. persica No14, where only duplicates were used due to reasons mentioned above. Mean and standard deviation were determined, and significant statistical differences were assessed via Student’s t-test (P-value < 0.05) by comparing wells incubated in chicken serum with the negative control.

Binding assay of serum components to spirochete surface

2.6

As the ELISA-experiment delivered a measurable but weak signal (around or below OD 0.2) and as the observed interaction could be explained by cytosolic factors instead of surface-exposed proteins, this experiment was complemented with a serum binding assay with viable spirochetes as previously described (Pietikäinen et al., 2010). Spirochetes at an amount of 5 × 10^7^ were washed five times in veronal buffered saline (VBS) and incubated in heat-inactivated chicken serum at 50% dilution (v/v) in VBS at 37 °C for 60 min under light agitation. Spirochetes were washed five times in VBS, and the final wash fraction was collected. Proteins bound to the surface of the spirochetes were eluted in 0.1 M glycine-HCl, pH 2, for 15 min at RT. Samples were centrifuged at 6000× g for 4 min at RT, and supernatants were collected. As a binding-control, one sample of 5 × 10^7^ spirochetes of the B. hermsii HS1 strain, with known capability to bind human FH, was incubated in heat-inactivated human serum but otherwise handled under the same conditions as above.

Eluted proteins were subjected to SDS-PAGE under non-reducing conditions (NuPAGE-Novex, Waltham, USA) and transferred to PVDF membranes (Amersham, Cardiff, UK). Membranes were blocked overnight in 5% (w/v) skimmed milk in PBS at 4 °C. Blocked membranes were incubated in polyclonal sheep anti-FH antibody (Lot# 8842, Abcam), dilution 1:1000 in 1% (w/v) skimmed milk-PBS for 1 h at RT. Signal was detected via HRP conjugated donkey-anti sheep IgG secondary antibody (Lot# sc2473, Santa Cruz Biotechnology) at 1:5000 dilution in 1% (w/v) skimmed milk-PBS, using an ECL western blot detection kit according to the manufacturer’s instructions (Amersham) and visualised using a Fujifilm LAS-4000 system (GE Healthcare Lifesciences, Cardiff, UK). Purified human FH (0.3 μg) served as a positive control. Final wash fractions were subjected to SDS-PAGE, and visualisation was performed as described above (Supplementary Fig. S3).

Results

3

The sensitivity of eight RF strains representing seven different species of Borrelia to complement active 50% (v/v) chicken serum was examined in this study (Fig. 1). All of the tested RF isolates displayed a high resistance to chicken serum, with most spirochetes remaining motile after 4 h incubation (90.8–94.0%). None of the tested strains was negatively affected by heat-inactivated serum, as expected.Fig. 1. Serum survival after incubation in 50% (v/v) chicken serum (CS) and 50% (v/v) heat-inactivated chicken serum (HICS) with 95% CI indicated.Fig. 1

Binding of chicken serum proteins that reacted with polyclonal anti-FH antibody was detected both to whole cell lysates and to the surface of viable spirochetes in two of the RF strains (B. duttonii 1120K and B. hermsii HS1) (Fig. 2, Fig. 3). The avian host-derived isolate (B. anserina Es) showed no binding of proteins reactive with the polyclonal anti-FH antibody, either to the bacterial surface or to whole cell lysate.Fig. 2. Whole cell lysate – chicken serum ELISA. Signal detected with polyclonal anti-FH antibody. Boxplot indicating range, with the middle line in the box representing the mean. Statistical significance of differences between individual samples and the negative control was assessed via Student’s t-test (∗P < 0.05).Fig. 2. Fig. 3Binding of chicken serum components reactive to polyclonal anti-FH antibody to the surface of viable spirochetes. Purified human FH served as a positive control. Eluted proteins from human serum components bound to B. hermsii HS1 served as a positive binding control.Fig. 3

Discussion

4

Aside from B. anserina, the capability of RF Borrelia spp. to infect avian species has been sparsely studied. Research conducted in the first half of the 20th century showed that the species B. duttonii, a human pathogen endemic to parts of Africa, was resistant to pigeon and chicken serum and could artificially infect chickens via injection (Cuboni, 1929; Kervran, 1947). More recent studies have found that chickens can serve as hosts for vector ticks, Ornithodoros moubata, and that 11% of the domesticated chickens in a rural setting in Tanzania tested PCR positive for RF Borrelia species in peripheral blood (McCall et al., 2007). Phylogenetic studies of B. hermsii have suggested avian involvement in its geographical spread, and B. hermsii has also been detected in wild birds and shown to infect chickens experimentally (Schwan et al., 2007), with its tick vector O. hermsii, able to feed on birds. Similarly, B. duttonii and B. hermsii strains exhibit serum resistance consistent with avian transmission. Spirochetes closely related to B. turicatae have been found in African penguins (Yabsley et al., 2012), and RF Borrelia spp. have been detected in seabird-associated ticks in Japan and the Mediterranean (Sanz-Aguilar et al., 2020). While direct avian infection with B. hispanica, B. crocidurae, B. persica, or B. recurrentis has not been documented, the vectors of B. hispanica and B. persica have been observed feeding on birds (Palma et al., 2013; Kleinerman et al., 2021).

Resistance against complement-mediated killing by chicken serum was high in all tested RF strains. This was also true for B. recurrentis, a strictly human-adapted pathogen transmitted exclusively by the human body louse, Pediculus humanus corporis. Borrelia recurrentis displayed unexpected serum resistance to chicken serum, despite lacking a known non-human reservoir. Genetic evidence suggests it evolved from B. duttonii through host restriction and genomic reduction (Swali et al., 2025). The observed resistance may reflect retained traits from B. duttonii, which is known to survive in avian hosts. The avian-oriented species B. anserina Es strain was resistant to chicken serum, as expected. Sequencing has shown that the genome of B. anserina is smaller and more condensed than other Borrelia species and does not contain any known proteins that bind FH, which is supported by the findings of this study (Elbir et al., 2017). Factor H is a host negative regulator of the CS, and binding of this protein to the bacterial surface is one of the best-described mechanisms for complement survival among RF species (Röttgerding and Kraiczy, 2020). In the present study, serum proteins reactive to polyclonal anti-FH bound to both whole cell lysate and to the surface of viable bacteria in two strains from two different species, B. duttoni 1120K3 and B. hermsii HS1, each displaying a serum-resistant phenotype. However, serum resistance was also observed in B. recurrentis A1, B. duttoni CR2A, B. persica No14, B. turicatae 91E135, B. hispanica CR1 and B. anserina Es that did not show detectable FH interaction, indicating that additional mechanisms contribute to RF serum survival.

A clear limitation of this study is the in vitro design and the lack of the correct species-specific reagents. The antibody used in this study displayed reactivity towards a protein of the appropriate size of FH from chicken serum. However, it is not possible from this study to prove that RF species bind FH from avian origin due to the antibody being polyclonal and raised against another target species. It is also important to note that the expression of proteins in bacteria differs depending on their growth in vitro or in vivo. Therefore, the failure to detect binding of a protein, such as FH, to the surface of bacteria grown in vitro does not exclude that the phenotype exists in vivo.

Conclusions

5

In this study, we investigated the ability of various relapsing fever Borrelia strains to survive in complement active chicken serum and assessed their interaction with serum proteins that showed reactivity to a polyclonal anti-FH antibody. All tested RF strains showed serum resistance in chicken serum, with possible factor H binding occurring in two strains. This could indicate that binding of chicken FH to RF species occurs, but is not a requirement for serum survival. As this study included species with known, suspected, and unlikely infective capability in avians, the findings might indicate a broadly conserved serum resistance in RF species towards the avian host. Further research is needed to clarify the underlying mechanisms of this phenotype.

Ethical approval

Not applicable.

CRediT authorship contribution statement

Tobias Jakobsson: Investigation, Methodology, Formal analysis, Writing - original draft, Writing - review & editing. Sven Bergström: Conceptualization, Supervision, Funding acquisition, Writing - review & editing. Johan Normark: Conceptualization, Supervision, Writing - review & editing.

Funding

This work was supported by VR-MH (10.13039/501100004359Swedish Research Council) [Grant number 07922].

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.