Complement activation and cytokine response by BioProtein, a bacterial single cell protein

Authors

  • L. I. B. Sikkeland,

    Corresponding author
    1. Centre for Occupational and Environmental Medicine, Rikshospitalet-Radiumhospitalet Medical Center, Oslo, Norway,
    2. Department of Respiratory Medicine, Faculty Division Rikshospitalet, University of Oslo, Norway,
    Search for more papers by this author
  • E. B. Thorgersen,

    1. Institute of Immunology, Rikshospitalet-Radiumhospitalet Medical Center, and Faculty Division Rikshospitalet, University of Oslo, Norway, and
    Search for more papers by this author
  • T. Haug,

    1. Centre for Occupational and Environmental Medicine, Rikshospitalet-Radiumhospitalet Medical Center, Oslo, Norway,
    Search for more papers by this author
  • T. E. Mollnes

    1. Institute of Immunology, Rikshospitalet-Radiumhospitalet Medical Center, and Faculty Division Rikshospitalet, University of Oslo, Norway, and
    2. Department of Laboratory Medicine, Nordland Hospital, Bodø and University of Tromsø, Norway
    Search for more papers by this author

Liv Ingunn Bjoner Sikkeland, Centre for Occupational and Environmental Medicine, Rikshospitalet-Radiumhospitalet Medical Center, N-0027 Oslo, Norway.
E-mail: liv.sikkeland@rikshospitalet.no

Summary

The bacterial single cell protein (BSCP), BioProtein, is dried bacterial mass derived from fermentation of the gram negative bacteria Methylococcus capsulatus, used for animal and fish feed. Workers in this industry suffer frequently from pulmonary and systemic symptoms which may be induced by an inflammatory reaction. The aim of the present study was to examine the effect of BSCP on inflammation in vitro as evaluated by complement activation and cytokine production. Human serum was incubated with BSCP and complement activation products specific for all pathways were detected by enzyme-linked immunosorbent assay (ELISA). Human whole blood anti-coagulated with lepirudin was incubated with BSCP and a panel of 27 biological mediators was measured using multiplex technology. BSCP induced a dose-dependent complement activation as revealed by a pronounced increase in alternative and terminal pathway activation (fivefold and 20-fold, respectively) at doses from 1 µg BSCP/ml serum and a similar, but less extensive (two- to fourfold) increase in activation of the lectin and classical pathways at doses from 100 and 1000 µg BSCP/ml serum, respectively. Similarly, BSCP induced a dose-dependent production of a number of cytokines, chemokines and growth factors in human whole blood. At doses as low as 0·05–0·5 µg BSCP/ml blood a substantial increase was seen for tumour necrosis factor (TNF)-α, interleukin (IL)-1-β, IL-6, interferon (IFN)-γ, IL-8, monocyte chemoattractant protein (MCP)-1, macrophage inflammatory protein (MIP)-1α, MIP-1β, IL-4, IL-9, IL-17, IL-1Ra, granulocyte–colony-stimulating factor (G-CSF) and vascular endothelial growth factor (VEGF). Thus, BSCP induced a substantial activation of all three initial complement pathways as well as a pronounced cytokine response in vitro, indicating a potent inflammatory property of this agent.

Introduction

The bacterial single cell protein (BSCP), BioProtein, is dried bacterial mass derived from fermentation of mainly the Gram-negative bacteria Methylococcus capsulatus. The BSCP is used as a protein additive in animal and fish feed. This protein pellet contains bacteria cell wall fragments such as endotoxins (approximately 1400 ng/mg). Workers in the plant producing BSCP reported attacks of fever, fatigue, chest tightness, skin dryness and inflammation in the eyes after high exposure. A decrease in lung function and an increase in leucocytes and interleukin (IL)-6 in peripheral blood has been demonstrated in these workers during a shift with high BSCP exposure (> 1000 ng/m3 endotoxins) [1]. A mouse-feeding study of BSCP induced a specific mucosa and systemic immune response, characterized by immunoglobulin A production in saliva and blood, respectively [2]. A similar rat-feeding study gave a humoral response involved in induction of BSCP-specific IgM in both sexes and IgG in females [3].

Release of cytokines and activation of the complement system are important events in the development of the inflammatory reaction and are involved in both the innate and the adaptive immune responses. Factors in the immune response that are triggered after exposure to BSCP are not well characterized. In the present study, we examined the complement-activating potential of BSCP in human serum and the cytokine-producing capacity of BSCP in human whole blood. The data indicate that BSCP is a potent inflammatory stimulator by activating complement and releasing a broad spectrum of cytokines. A possible role for the inflammatory potential of BSCP in the adverse effects seen in humans exposed to BSCP is suggested.

Materials and methods

BSCP

The BSCP was provided by Norferm (Tjeldbergodden, Norway). The protein consists of approximately 90% M. capsulatus, 10% Alcalligenes and Bacillus species. This product contains whole bacteria with all the chemical compounds found normally in the living and reproducing cell. The protein product is a spray-dried biomass containing 70% protein, 12% carbohydrates, 10% fat, 7% minerals and 1% fibre. Levels of lipopolysaccharide (LPS) in BSCP are ∼14 000 EU/mg (1400 ng/mg) (LAL-test; Norferm, Odense, Denmark).

Serum experiments

Pooled normal human serum from several (n = 10) donors was incubated with BSCP to final concentrations ranging from 0·1 to 1000 µg/ml serum for 30 min at 37°C. Human serum albumin (HSA) was used as control. Complement activation was stopped at the end of incubation by adding ethylenediamine tetraacetic acid (EDTA) to a final concentration of 20 mM. Baseline samples (T0) were processed immediately, whereas the other samples were placed on ice after incubation. All samples were centrifuged for 15 min at 3000 g (4°C). The supernatants were stored at − 70°C until analysed.

Complement activation products

Serum incubated with BSCP or controls (HSA) were analysed for the following complement activation products: C1–inh–C1rs complexes (classical pathway), C4bc (classical and lectin pathway), C3bBbP (alternative pathway) and terminal complement complex (TCC) (terminal pathway).

C1–inh–C1rs complexes were measured using the KOK-12 monoclonal antibody specific for a neoepitope in C1-inhibitor when it is in complex with the protease [4]. Briefly, microtitre plates were coated with the KOK-12 antibody, reacted with plasma and control samples, and the complex was detected using a cocktail of anti-C1r and anti-C1s antibodies. The standard was normal human serum activated with heat-aggregated IgG according to the following protocol: human IgG (gammaglobulin; Kabi, Uppsala, Sweden) was diluted to 10 mg/ml in phosphate-buffered saline, pH 7·2, incubated in a waterbath at 63°C for 15 min and then cooled on crushed ice. The heat-aggregated IgG was added to normal human serum (1 mg/ml serum), incubated at 37°C for 30 min and centrifuged at 6000 g. The supernatant was tested for complement activation products and found to be comparable with a serum fully activated by zymosan. This standard was defined to contain 1000 arbitrary units (AU)/ml. C4bc (i.e. the sum of C4b, iC4b and C4c) was measured mainly as described previously [5]. The monoclonal antibody C4-1 reacting with a neoepitope exposed in C4b, iC4b and C4c was used as capture antibody and a polyclonal anti-C4 as detection antibody. The standard was normal human serum activated with heat-aggregated IgG, as described above, and defined to contain 1000 AU/ml. The monoclonal antibodies to C1-inhibitor and C4bc were a kind gift from Professor C. E. Hack (Amsterdam, the Netherlands). C3bBbP (i.e. the alternative pathway convertase) was measured as described previously [6]. The monoclonal anti-properdin antibody 2 was used as capture antibody and a polyclonal anti-C3c as detection antibody. The standard was normal human serum activated with zymosan and defined to contain 1000 AU/ml. TCC (terminal SC5b-9 complement complex) was measured as described previously [7] and later modified [8]. The monoclonal antibody aE11 reacting with a neoepitope exposed in C9 after incorporation in the C5b-9 complex was used as capture antibody and a biotinylated monoclonal anti-C6 as detection antibody. The standard was normal human serum activated with zymosan and defined to contain 1000 AU/ml.

The concentrations of the complement activation products are given in AU/ml for the following reasons: some of the neoepitope-specific monoclonal antibodies recognize activation epitopes present in several of the activation products, as is the case for the antibodies to C4bc (iC4b and C4c) and C3bc (C3b, iC3b and C3c), implying that it is the amount of the activation epitope which is quantified, irrespective of the molecular weight of the fragments detected. In other cases activation product complexes may be heterogeneous with respect to composition and molecular weight (e.g. C1rs–C1–inh complexes and TCC). For these reasons we have defined our standards according to the amount of activation products present in a fully activated serum, defined to contain 1000 AU/ml for all activation products.

Whole blood experiments

Human whole blood from three different donors was collected. The blood was anti-coagulated with lepirudin 50 µg/ml, as described previously [6], placed on ice and split into tubes immediately for incubation. BSCP or HSA was added to the final concentrations, ranging from 0·05 to 50 µg/ml whole blood. The samples were then incubated for 4 h at 37°C. Baseline samples and test samples were prepared as described for the serum samples above.

Cytokines

Plasma samples from the whole blood experiments were analysed using a multiplex cytokine assay (Bio-Plex Human Cytokine 27-Plex Panel; Bio-Rad Laboratories Inc., Hercules, CA, USA) containing the following cytokines, chemokines and growth factors: IL-1β, IL-1 receptor antagonist (IL-1Ra), IL-2, IL-4, IL-5, IL-6, IL-7, IL-8 (CXCL8), IL-9, IL-10, IL-12 p70, IL-13, IL-15, IL-17, eotaxin (CCL11), basic fibroblast growth factor (bFGF), granulocyte–colony stimulating factor (G-CSF), granulocyte–macrophage colony stimulating factor (GM-CSF), interferon (IFN)-γ, chemokine (C-X-C motif) ligand 10 [interferon-inducible protein (IP)-10 or CXCL10], monocyte chemotactic protein [(MCP)-1 or CCL2], macrophage inflammatory protein [(MIP)-1α or CCL3], MIP-1β or CCL4, platelet-derived growth factor-BB (PDGF-BB), regulated upon activation T cell expressed and secreted (RANTES or CCL5), tumour necrosis factor (TNF)-α and vascular endothelial growth factor (VEGF). The analysis was preformed according to the manufacturer's instructions.

Ethics

The study was approved by the Regional Committee for Medical Research Ethics, Southern Norway, Oslo, Norway.

Results

Effect of BSCP on complement activation in serum

BSCP induced a dose-dependent activation of all complement pathways at doses up to 1000 µg BSCP/ml serum (Fig. 1). Activation of the alternative (C3bBbP) and terminal (TCC) pathways was most pronounced (fivefold and 20-fold increases, respectively, compared to control) and seen already at 1 µg BSCP/ml serum. Activation of the classical pathway (C1–inh–C1rs) was very modest (twofold increase) and seen first at 1000 µg BSCP/ml serum, whereas C4bc (reflecting both the classical and lectin pathways) increased moderately (fourfold) at 100 µg BSCP/ml serum. Thus, the increase in C4bc most probably reflects activation of the lectin pathway.

Figure 1.

Complement activation after incubation of bacterial single cell protein (BSCP) for 1 h in human serum. C1–inh–C1rs complexes (upper left panel) reflects classical pathway activation, C4bc (upper right panel) reflects both classical and lectin pathway activation, C3bBbP (lower left panel) reflects alternative pathway activation and terminal complement complex (TCC) (lower right panel) reflects final common terminal pathway activation. Median and range of three separate experiments are depicted.

Cytokine induction by BSCP in whole blood

Of the 27 different cytokines, chemokines and growth factors, which were measured after BSCP incubation in human whole blood, 14 increased substantially and dose-dependently. There was a 1000-fold increase for many of the proinflammatory cytokines (Figs 2–4).

Figure 2.

Proinflammatory cytokines induced by bacterial single cell protein (BSCP) after incubation of lepirudin-treated human whole blood for 4 h. A dose-dependent increase in tumour necrosis factor (TNF)-α (upper left panel), interleukin (IL)-1β (upper right panel), IL-6 (lower left panel) and interferon (IFN)-γ (lower right panel) was observed. Median and range of three separate experiments are depicted.

Figure 3.

Chemokines induced by bacterial single cell protein (BSCP) after incubation of lepirudin-treated human whole blood for 4 h. A dose-dependent increase in interleukin (IL)-8 (upper left panel), monocyte chemoattractant protein (MCP)-1 (upper right panel), macrophage inflammatory protein (MIP)-1α (lower left panel) and MIP-1β (lower right panel) was observed. Median and range of three separate experiments are depicted.

Figure 4.

Cytokines and growth factors induced by bacterial single cell protein (BSCP) after incubation of lepirudin-treated human whole blood for 4 h. A dose-dependent increase in interleukin (IL)-4 (upper left panel), IL-9 (upper right panel), IL-17 (middle left panel), IL-1Ra (middle right panel), granulocyte–colony-stimulating factor (G-CSF) (lower left panel) and vascular endothelial growth factor (VEGF) (lower right panel) was observed. Median and range of three separate experiments are depicted.

A dose-dependent and pronounced increase in the proinflammatory cytokines TNF-α, IL-1β, IL-6 and IFN-γ was seen from 0·5 µg to 50 µg BSCP/ml blood (Fig. 2). A similar dose-dependent and pronounced increase in the chemokines IL-8, MCP-1, MIP-1α and MIP-1β was seen from 0·5 µg BSCP/ml blood, but at 50 µg BSCP/ml blood the production of these chemokines decreased (Fig. 3). At doses above 50 µg BSCP/ml blood all mediators tested tended to decrease, suggesting a toxic effect on the cells (data not shown).

Production of the Th2 cytokines IL-4 and IL-9 (Fig. 4, upper panels) and of IL-17 and IL-1Ra (Fig. 4, middle panel) increased moderately and dose-dependently at doses as low as 0·05 µg BSCP/ml blood, whereas G-CSF and VEGF (Fig. 4, lower panels) increased dose-dependently from 0·5 µg BSCP/ml.

Production of the cytokines IP-10, PDGF-BB, eotaxin and RANTES was low and did not increase when increasing the BSCP concentration, whereas IL-2, IL-5, IL-7, IL-10, IL-12p70, IL-13, IL-15, bFGF and GM-CSF were not detected (data not shown).

Discussion

In the present study we have shown for the first time that BSCP activates complement and induces synthesis of a number of cytokines. The in vitro human whole blood model used in this study has the advantage of keeping the inflammatory systems intact and mutually able to interact, as the highly specific thrombin inhibitor lepirudin is used as anti-coagulant which, in contrast to most anti-coagulants, does not interfere with other biological processes [6]. Thus, as this model simulates closely the in vivo situation, the data obtained indicate that BSCP may contribute to the suffering among the workers exposed to BSCP by inducing a complement- and cytokine-mediated inflammatory reaction. However, it cannot be excluded that the mechanism of BSCP in the airways would be different from the blood.

BSCP contains constituents from dried bacteria, mainly from the Gram-negative bacteria M. capsulatus. In the present study BSCP activated several of the complement activation pathways, consistent with the potential of Gram-negative bacteria to activate complement. The classical pathway is activated through antibodies in serum complexed with antigen, or by C-reactive protein or C1q bound directly to the bacterial surfaces [9,10]. The lectin pathway is activated through recognition of bacterial structures by mannose-binding lectin or ficolins [11,12]. Lastly, the alternative complement pathway is activated directly by discriminating non-self-structures on the bacteria from self-structures [13] or by amplifying the classical and lectin pathways [14]. Our data indicate that BSCP activates all three pathways of complement directly. First, the alternative pathway convertase C3bBbP was generated at substantially lower doses of BSCP than C1–inh–C1rs complexes and C4bc, supporting a direct alternative pathway activation. Secondly, C1–inh–C1rs complexes were formed, although in moderate amounts, indicating classical pathway activation. This activation is due probably to antibodies present in normal human serum. Chronic exposure to BSCP may, however, lead to an elicited antibody response which may increase this activation. Thirdly, C4bc was formed at a lower BSCP dose than C1–inh–C1rs complexes and in relatively larger amounts than could be explained easily by the modest classical pathway activation, also indicating a possible lectin pathway activation. At present, no assays for specific detection of lectin pathway activation products in the fluid-phase are available. Finally, in addition to being directly activated, the alternative pathway may have been amplified by the classical and lectin pathways at concentrations of BSCP where these pathways were activated.

LPS is regarded traditionally as a potent activator of the alternative pathway. Recent data, however, indicate that purified LPS in solution requires large amounts (µg/ml) to activate complement and that an LPS-deficient strain of Neisseria menigitidis activated complement as efficiently as the LPS-sufficient strain [15]. In the present study, BSCP activated the alternative pathway at an LPS concentration (10 µg/ml BSCP, which equals ∼14 ng/ml LPS) which is far below the complement-activating dose. Thus, the present data indicate that there are other biological active bacterial substances than LPS present in the BSCP that activate complement.

Several cytokines were synthesized heavily during BSCP incubation, induced at a concentration of 0·5 µg/ml BSCP, corresponding to approximately 0·7 ng LPS/ml, or even lower. In contrast to the relatively high concentration of LPS required for complement activation, cytokines may be induced at concentrations of 10–100 pg LPS/ml [16,17]. LPS may thus be a candidate trigger of the cytokine response induced by BSCP, also in vivo. It has been shown recently that a single dose of intravenous LPS to humans gave a broad and rich profile of gene expression changes in blood, both with cytokines associated with LPS-induced inflammation and several genes that have not been associated previously with LPS [18].

A common activation pattern for bacteria is initiated by activating microbial pattern recognition receptor on cell surfaces. LPS is known to activate macrophages/monocytes through the TLR4/MD2/CD14 pathway, which induces secretion of several inflammatory cytokines including the proinflammatory cytokines TNF-α, IL-1β, IL-6 and IFN-γ, and the chemotactic proteins IL-8, MCP-1, MIP-1α and MIP-1β. All these mediators were increased after BSCP stimulation in the present study, further supporting LPS as a candidate trigger of synthesis. Besides the TLR4/MD2/CD14 pathway, it has been discovered recently that bacteria or LPS can induce inflammation and neutrophil recruitment through the IL-23/IL-17/G-CSF pathway [19–21]. IL-23, produced by monocytes, macrophages or dendritic cells, has been shown to stimulate IL-17 production in T helper 17 (Th17) cells. IL-17 then stimulates granulopoiesis through G-CSF [22]. Activated Th17 cells also produce TNF-α and IL-6. Notably, IL-17 and G-CSF, as well as TNF-α and IL-6, were increased after BSCP stimulation. It is therefore possible that BSCP is activated by both the TLR4/MD2/CD14 and the IL-23/IL-17/G-CSF pathways.

The Th2 cytokines, IL-4 and IL-9, were increased after BSCP stimulation despite the short incubation period of 4 h. Whereas IL-4 was increased only marginally, IL-9 was increased markedly to 30-fold from baseline. Activated Th2 cells produce IL-4 and IL-9, and it is shown that IL-4 and IL-9 are capable of inhibiting in vitro human blood monocytes activated by LPS [23]. The source of synthesis and the biological role of IL-9 induced by BSCP remain uncertain.

IL-1Ra is capable of inhibiting IL-1 both in vitro and in vivo, thus representing a natural powerful mechanism to control IL-1-dependent responses. It has been shown in humans that following injection of LPS or TNF-α, plasma IL-1Ra levels increase rapidly, suggesting that TNF-α may be an intermediate in LPS-induced IL1-Ra production [24]. Taken together, BSCP induces an inflammatory reaction represented by the proinflammatory mediators TNF-α and IL-1β, associated with the subsequent physiological counteraction by the anti-inflammatory IL-1Ra, simulating closely the in vivo situation.

VEGF, a central cytokine/growth factor for endothelial cells, was induced by BSCP. The lung is one of the organs with highest expression of VEGF. VEGF is critical for the development of the lung, and serves as a maintenance factor during adult life [25]. However, there is increasing evidence that VEGF is important in the pathobiology of lung diseases. Increased expression of VEGF is seen among patients with asthma and pulmonary hypertension. Our data indicate that BSCP may stimulate VEGF production in the lungs of individuals exposed to BSCP.

Workers in the BSCP industry were exposed to 6–8900 ng LPS/m3 during a working day, depending on their working task [1]. A possible entrance of trapped BSCP particles to the lung interstitium and subsequently to blood should be considered. In an animal model, Goto and Rylander have demonstrated that LPS can penetrate the lung barrier and be detected in the arterial and venous blood afterwards [26], providing support to this hypothesis, although exposure to LPS in this model was higher (400 µg/m3). Notably, the workers at the BSCP factory had ongoing inflammation both in lung (unpublished results) and blood [1], indicating that LPS may translocate from the inflammatory lung directly into the blood. Alternatively, it is also possible that the systemic inflammation seen among the workers is related to secondary inflammatory signals generated in the lung as a response to BSCP.

In conclusion, BSCP induced a powerful inflammatory response, as revealed by activation of complement and induction of proinflammatory cytokines, chemokines and growth factors. This response is important for the efficient control of growth and dissemination of invading pathogens, but may harm the host when induced improperly, as may be the case in workers exposed to BSCP.

Acknowledgements

Financial support was kindly provided by the Research Council of Rikshospitalet, the Research Council of Norway, the Working Environmental Fund, Confederation of Norwegian Enterprise, the Family Blix Foundation, Sigvald Bergesen d.y. and wife Nanki's Foundation, and NIH grant no R01-EB-003968–01. Finally, we thank Norferm for providing the test material.

Ancillary