- Email: [email protected]
Antifungal activity of the local complement system in cerebral aspergillosis
Microbes and Infection 7 (2005) 1285–1295 www.elsevier.com/locate/micinf
Original article
Antifungal activity of the local complement system in cerebral aspergillosis Günter Rambach a, Magdalena Hagleitner a, Iradj Mohsenipour b, Cornelia Lass-Flörl a, Hans Maier c, Reinhard Würzner a, Manfred P. Dierich a, Cornelia Speth a,* a
Department of Hygiene, Microbiology and Social Medicine, Innsbruck Medical University, and Ludwig-Boltzmann-Institute for AIDS Research, Fritz-Pregl-Str. 3, 6020 Innsbruck, Austria b Department of Neurosurgery, Innsbruck Medical University, Anichstr. 35, 6020 Innsbruck, Austria c Department of Pathological Anatomy, Innsbruck Medical University, Müllerstr. 44, 6020 Innsbruck, Austria Received 13 December 2004; accepted 22 April 2005 Available online 13 June 2005
Abstract Dissemination of aspergillosis into the central nervous system is associated with nearly 100% mortality. To study the reasons for the antifungal immune failure we analyzed the efficacy of cerebral complement to combat the fungus Aspergillus. Incubation of Aspergillus in non-inflammatory cerebrospinal fluid (CSF) revealed that complement levels were sufficient to obtain a deposition on the surface, but opsonization was much weaker than in serum. Consequently complement deposition from normal CSF on fungal surface stimulated a very low phagocytic activity of microglia, granulocytes, monocytes and macrophages compared to stimulation by conidia opsonized in serum. Similarly, opsonization of Aspergillus by CSF was not sufficient to induce an oxidative burst in infiltrating granulocytes, whereas conidia opsonized in serum induced a clear respiratory signal. Thus, granulocytes were capable of considerably reducing the viability of serum-opsonized Aspergillus conidia, but not of conidia opsonized in CSF. The limited efficacy of antifungal attack by cerebral complement can be partly compensated by enhanced synthesis, leading to elevated complement concentrations in CSF derived from a patient with cerebral aspergillosis. This inflammatory CSF was able to induce (i) a higher complement deposition on the Aspergillus surface than non-inflammatory CSF, (ii) an accumulation of complement activation products and (iii) an increase in phagocytic and killing activity of infiltrating granulocytes. However, levels and efficacy of the serum-derived complement were not reached. These data indicate that low local complement synthesis and activation may represent a central reason for the insufficient antifungal defense in the brain and the high mortality rate of cerebral aspergillosis. © 2005 Elsevier SAS. All rights reserved. Keywords: Cerebral aspergillosis; Innate immunity; Complement; Aspergillus; Phagocytes
1. Introduction Invasive aspergillosis is an opportunistic infection that contributes considerably to the morbidity and mortality among immunocompromized patients, including patients with acute leukemia, transplant recipients, patients with autoimmune diseases treated with immunosuppressive regimens, and AIDS patients [1]. A dangerous complication of invasive aspergillosis is the involvement of the brain, which occurs in 10–20% of all cases and has a mortality rate of nearly 100% [1]. Inva* Corresponding author. Tel.: +43 512 507 3405; fax: +43 512 507 2870. E-mail address: [email protected] (C. Speth). 1286-4579/$ - see front matter © 2005 Elsevier SAS. All rights reserved. doi:10.1016/j.micinf.2005.04.014
sion of the central nervous system (CNS) mostly occurs either via blood from a pulmonary focus, or directly by infection after head trauma or surgical procedures [2]. The clinical presentation includes lesions, brain abscess formation, granulomatous reactions, cerebral infarction, hemorrhage, meningoencephalitis and meningitis [1]. Prominent symptoms of the patients are fever, altered mental status, headache, hemiplegia and seizures [3]. Treatment regimens with antifungal agents are often of limited efficacy because of their poor penetration into the brain, underlining the relevance of an efficient local immune response. The reasons for the insufficient antifungal response in the brain are not yet known. In general, the brain is an immuno-
1286
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
privileged organ and thus protected from harmful inflammatory processes. This immunoprivilege is guaranteed by several factors, including (i) the blood–brain barrier that restricts the access of peripheral immune effectors like lymphocytes and antibodies; (ii) the deactivation of infiltrating granulocytes and monocytes/macrophages by resident astrocytes [4]; and the low complement levels in the cerebrospinal fluid (CSF) [5]. Complement is a central element of the innate immune system and enables a fast and effective antimicrobial defense. In general, activation of the complement cascade does not only induce lysis of invading pathogens but also opsonizes them to become a target for phagocytosis by complementreceptor bearing cells. In addition, the pro-inflammatory anaphylatoxins C3a and C5a generated by complement activation attract phagocytes to the site of infection and can activate different cell types via specific intracellular signal transduction pathways (reviewed in: [6]). Whereas the interaction of complement with Aspergillus spp. in the brain has not yet been studied, some aspects are known from the peripheral blood. Aspergillus conidia and hyphae are potent activators of the serum complement cascade by classical and alternative pathway leading to deposition of C3 on the fungal surface [7,8]. Complement deficiency is correlated with enhanced susceptibility to a disseminated infection by A. fumigatus [9]. The degree of complement deposition on Aspergillus correlates indirectly with pathogenicity, with highly virulent species binding less C3 on their surface than non-pathogenic species [10]. The CNS produces its own complement with astrocytes, neurons, microglia and oligodendrocytes being all able to synthesize complement factors [11,12]. However, the constitutive synthesis in the brain is low, making it unclear whether the complement levels in CSF are sufficient for a competent antifungal defense. We, therefore, studied the ability of cerebral complement to fulfill functions like opsonization, induction of phagocytosis and oxidative burst in phagocytic cells. Our experiments showed that the complement concentration in non-inflammatory CSF was sufficient for opsonization of fungal hyphae, but the intensity was much weaker than the opsonization in serum. Aspergillus conidia opsonized by human serum triggered an oxidative burst in granulocytes whereas the burst signal induced by CSF-opsonized conidia corresponded to that of non-opsonized conidia. Similarly, phagocytosis and lysis of fungal conidia opsonized by CSF was low in case of granulocytes, monocytes, macrophages and microglia, whereas a high percentage of these cell types ingested and killed serum-opsonized conidia. The concentration of complement in CSF was highly upregulated during the pathogenesis of cerebral aspergillosis, giving raise to increased opsonization of the fungal surface, higher phagocytosis and better killing by granulocytes. However, this upregulation might still be insufficient or occur too late to inhibit effectively the fungal infection and spreading of hyphae in the brain tissue.
2. Materials and methods 2.1. Aspergillus isolate and culture Aspergillus fumigatus isolate 14 was obtained from a hospitalized patient with cerebral aspergillosis; the patient suffered from acute myeloic leukemia and neutropenia as underlying disease [13]. The fungus was grown for at least 5 days on Sabouraud (Oxoid, Wesel, Germany) agar plates at 28 °C until sporulation was clearly visible. The conidia were swept off from sporulating colonies with phosphate-buffered saline (PBS) containing 0.05% Tween-20 (Sigma, St. Louis, USA) and kept at 4 °C. 2.2. Deposition of complement on fungal germination tubes Both human serum and CSF were used for studies of complement deposition on fungal hyphae. Human serum was obtained from five to six healthy individuals, pooled and frozen at –80 °C for further use. CSF pools were obtained from 15 individuals who were investigated for neurological noninflammatory diseases and also stored at –80 °C. For opsonization experiments 105 conidia were grown on prelubricated microscope slides at 28 °C until formation of germination tubes was seen. The adherent fungi were washed with PBS and fixated with acetone and methanol at –20 °C. Fixated fungi were incubated with serum or CSF at various concentrations, or, as a negative control, with GVB buffer (veronal buffer saline containing 0.1% gelatin, 0.15 mM CaCl2, 0.1 mM MgCl2; bioMérieux, Marcy l’Etoile, France). Deposition of complement factors was detected by standard indirect immunofluorescence procedure after 1.5 h for C1q, C4 and C3 and after 3.5 h for C5 and C7. Briefly, the slides were washed with PBS to remove serum or CSF, followed by blocking of unspecific binding with PBS/1% BSA (Sigma). The specific primary antibody was added for 1 h at 37 °C. After extensive washing the fluorescence-labeled secondary antibody was incubated for 30 min and visualized in a Zeiss Axioplan microscope. The following antibodies were used for the immunofluorescence analysis: polyclonal a-C1q, a-C4, a-C3d and a-C5 antibodies (Dako, Glostrup, Denmark); a monoclonal a-C7 [14]; a secondary goat-a-rabbit or goat-a-mouse Ig, labeled with the dye Alexa 488 (Molecular Probes, Leiden, The Netherlands). 2.3. Quantification of complement factors in CSF by ELISA Complement quantification of C1q, C3, C5 and TCC was performed by sandwich ELISA using the following antibodies: the monoclonal coating a-C3d antibody BB5 [15] and a rabbit polyclonal a-C3c (Dako, Glostrup, Denmark) for detection; for C1q a monoclonal coating antibody (Quidel, San
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
Diego, USA) and a rabbit polyclonal a-C1q (Dako); for C5 a monoclonal coating antibody (clone N19-8) [14] and a rabbit polyclonal detection antibody (Dako); for TCC a monoclonal coating antibody (clone WU 7-2) and a biotinylated polyclonal goat a-C6 [14]. Microplates were coated overnight at 4 °C with the corresponding coating antibody in 0.1 M NaHCO3, pH 9.6. Unspecific binding was inhibited by saturation with 1% BSA (Sigma) in PBS. CSF samples were added to the coated wells and incubated for 1 h at room temperature. After washing with PBS the bound complement was detected with the corresponding second antibody, followed by a peroxidaselabeled antibody (Dako). For TCC-ELISA an avidinhorseradish peroxidase conjugate (Dako) was used to detect the biotinylated goat a-C6 antibody. Development was performed with tetramethylbenzidine dihydrochloride (Sigma). Purified protein standards (C3: Sigma; C1q, C5: Quidel) were used to calculate the exact amounts of the complement factors in the supernatants. C3a was quantified using a commercially available kit (Quidel) according to manufacturers’ instruction, with a similar procedure as described above. 2.4. Isolation of monocytes, granulocytes, macrophages, dendritic cells and microglia For phagocytosis experiments granulocytes and monocytes were isolated from buffy coat by density gradient centrifugation over Ficoll–Paque (Pharmacia, Uppsala, Sweden); peripheral blood mononuclear cells (PBMC) layer and granulocyte layer were obtained separately. The granulocytes were further purified by lysis of erythrocytes with sterile distilled water for 20 s followed by washing with PBS. The monocytes were separated from lymphocytes using gelatin/plasma-coated plates as described previously [16]. Briefly, Petri dishes were coated with sterile 2% gelatin (Merck) solution for 2 h at 37 °C followed by incubation with autologous plasma for 30 min at 37 °C. PBMC obtained from Ficoll separation (Amersham, Buckinghamshire, UK) were suspended in RPMI medium (Life Technologies, Vienna) containing 10% FCS, 5% pooled serum and 2 mM L-glutamine and loaded onto gelatin/plasma-coated plates at 4 × 106 cells/ml. After 40 min of incubation at 37 °C peripheral blood lymphocytes (PBL) were washed out with medium; adherent monocytes were detached by 5 mM EDTA (Sigma) and further cultured in complete RPMI medium at 37 °C in a humidified 5% CO2 atmosphere. Fluorescence-activated cell sorter (FACS) analysis revealed that the cell population regularly consisted of 94–96% monocytes. Macrophages were obtained by differentiation of monocytes with human AB serum (Sigma) for at least 7 days until morphological changes were clearly visible. For oxidative burst experiments a crude preparation of granulocytes was obtained from heparinized blood. Aliquots were incubated with lysis buffer (150 mM ammonium chloride, 10 mM potassium hydrogen carbonate, 0.1 mM EDTA)
1287
for 5 min at room temperature. The suspension was centrifuged for 5 min at 1000 rpm and the pellet incubated again for 5 min in lysis buffer, followed by centrifugation. The cells were washed twice in PBSg (PBS with 0.2% gelatin (Merck, Darmstadt, Germany)) and finally resuspended in 300 µl PBSg. Monocyte-derived mature dendritic cells were obtained from previously isolated monocytes. For differentiation monocytes were cultivated in complete RPMI medium supplemented with 1500 U/ml IL-4 and 1600 U/ml GM-CSF at a density of 106 cells/ml in six-well plates. After 2 days 1000 U/ml IL-4 and 1600 U/ml GM-CSF were added. On day 5, cells were collected, washed and transferred to a new plate. Cells were cultivated further for 2 days in the presence of fresh complete RPMI medium supplemented with 100 U/ml IL-4, 1600 U/ml GM-CSF and 100 ng/ml lipopolysaccharide (LPS; Sigma) to obtain mature dendritic cells. Primary microglia from human adult brain were isolated from surgery material. The tissue was dissected using enzymatic digestion with 0.5% DNase and 0.5% collagenase (Sigma) and washed twice in DMEM culture medium. The cell suspension was purified from myelin by a percoll gradient (Amersham Biosciences, Little Chalfont, UK), washed in cell culture medium and cultured in DMEM containing 10% FCS, 2 mM glutamine and 1 × non-essential amino acids (Life Technologies), supplemented with 25 ng/ml GM-CSF (Peprotech, London, UK). The percentage of microglia and astrocytes was estimated by double immunofluorescence using an 具 -CD68 antibody (Dako) for microglia and an a-GFAP antibody (Dako) for astrocytes. The microglia preparation used for the experiments presented here consisted of 70% microglia and 30% astrocytes. For analysis of phagocytosis only microglia were counted, since these two cell types have a different morphology and can be distinguished easily. 2.5. Oxidative burst in granulocytes To quantify oxidative burst, the granulocytes were loaded for 15 min with the diacetate form of 2′,7′-dichlorodihydrofluorescin (Sigma). This cell-permeable substance is deesterified intracellularly by esterases concentrating the polar non-fluorescent DCFH in the cytoplasm. During the oxidative burst, DCFH is converted to highly fluorescent 2′,7′dichlorofluorescin diacetate through oxidation mainly by H2O2. The fluorescence of the cells after stimulation was analyzed by FACS gating the neutrophils [17]. Aspergillus conidia were incubated for 60 min either in GVB at 37 °C (non-opsonized conidia) or opsonized under the same conditions in serum or CSF. Thirty conidia per cell were added to each sample and the oxidative burst was measured by FACS at different time points. As a positive control for the capacity of granulocytes to react on stimulation, 10 mM 4-a-phorbol 12-myristate 13-acetate (PMA; Sigma) was added to one sample and the following activation of the dye was analyzed in parallel. Analysis was performed using a FACScan flow cytometer (Becton Dickinson, San Diego).
1288
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
2.6. Phagocytosis assay Aspergillus conidia were stained with Calcofluor White (Polysciences, Warrington, USA) for 2 min, followed by washing with PBS. For opsonization, the stained conidia were incubated for 60 min at 37 °C with serum, CSF, or with GBV as a negative control. Thirty conidia per cell were added for 20–40 min. Uningested conidia were removed by washing and the cells were analyzed microscopically. The following indices were calculated after counting 50–100 cells [18]: % phagocytosis, determined as percentage of cells that had one or more conidia phagocytosed; phagocytic index (PI) as the average number of conidia that had been phagocytosed by each phagocytosing cell. 2.7. Quantification of fungal killing Aspergillus conidia were opsonized by incubation for 60 min at 37 °C in serum, CSF samples, or in GVB as negative control. Granulocytes isolated from buffy coat were added to the conidia in an effector/target ratio of 1:1.5 with a final volume of 300 µl. After 5 h the neutrophils were lysed by adding 5.7 ml distilled water for 5 min. Different volumes of this suspension were plated on Sabouraud agar plates and the number of colonies was counted after 24 h. 2.8. Statistical analysis Statistical analysis (Student’s t-test) was performed using the Origin 6.1 software (Microcal).
3. Results 3.1. Deposition of complement from CSF on Aspergillus hyphae To investigate whether the low constitutive complement levels in normal CSF are sufficient to obtain opsonization, Aspergillus hyphae were incubated either in buffer or in normal CSF and the binding of complement was detected subsequently by immunofluorescence. Complement deposition on the fungal surface could be shown for various complement factors, including C1q, C4, C3, C5 and C7 (Fig. 1). The signals were weak, but clearly visible compared to the hyphae incubated in buffer, indicating that complement levels in noninflammatory human CSF are sufficient for opsonization of the fungus. To compare the opsonization capacity of normal CSF with the one of human serum, different concentrations of serum in comparison to CSF were used for complement deposition experiments, and C3 on the hyphal surface was again visualized by immunofluorescence. A strong deposition signal was obtained with serum at a dilution of 1:100 and a weak signal after diluting serum 1:1200 (Fig. 2A–C). The intensity of
Fig. 1. Deposition of complement on Aspergillus hyphae. The clinical isolate A. fumigatus 14 was grown on slides, fixated and the hyphae were incubated either in GVB buffer (A, C, E, G, I) or normal CSF (B, D, F, H, K). Deposition of the complement factors C1q (A, B), C4 (C, D), C3 (E, F), C5 (G, H) and C7 (I, K) on the hyphae was detected with specific antibodies and visualized by a fluorescence-labeled second antibody.
opsonization in CSF corresponded to the one obtained in serum at a dilution between 1:100 and 1:400 (Fig. 2D), dilution of CSF abolished the opsonization signal. This corresponds to the fact that C3 concentration in serum (1200 g/ml; [19]) is approximately 400-fold higher than in CSF (2.8 g/ml) [5]).
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
1289
Fig. 2. Deposition of complement factor C3 from serum or normal CSF on Aspergillus hyphae. The clinical isolate A. fumigatus 14 was grown on slides, fixated and the hyphae were incubated either with serum diluted 1:100 (A), 1:400 (B), and 1:1200 (C) or with undiluted CSF (D). Deposition of the complement factor C3 was demonstrated by immunofluorescence using a C3d-antibody.
3.2. Modulation of phagocytic activity by opsonization of Aspergillus conidia To test that whether opsonization of the fungal surface by CSF is strong enough to support the phagocytosis of Aspergillus by infiltrating granulocytes, granulocytes were incubated with A. fumigatus conidia either mock-treated in buffer or opsonized in serum or CSF; the percentage of phagocytosing granulocytes and the phagocytic index PI (average number of ingested conidia per phagocytosing cell) were quantified microscopically. Whereas opsonization of the conidia in serum strongly stimulated phagocytosis by granulocytes with 70% of cells showing phagocytic activity and a phagocytic index of 3.5 conidia per cell, the triggering by CSF-opsonized conidia corresponded to that of non-opsonized conidia with 18% phagocytosing cells and a PI of 1.7 conidia per cell (Fig. 3A, B). Similarly, the percentage of phagocytosing monocytes, macrophages and mature dendritic cells was much lower with conidia opsonized in CSF than with serum-opsonized conidia (32% versus 68%; 28% versus 85%; 16% versus 99%, respectively) (Fig. 3C–E). Likewise, microglial cells have a much higher phagocytic activity after contact with serum-opsonized conidia than with CSF-opsonized conidia (86% versus 33%) (Fig. 3F). For monocytes, macrophages and microglia the phagocytosis rate was slightly higher for CSF-opsonized than for non-opsonized conidia, and the difference even reached significance for monocytes and microglia, indicating that the weak opsonization by CSF may slightly support the uptake. To exclude that factors others than complement products might contribute to enhanced phagocytosis after opsonization in serum, we performed control experiments, adding 20 mM EDTA to serum and CSF during opsonization of the conidia. EDTA is an efficient inhibitor of complement acti-
Fig. 3. Phagocytosis of opsonized and non-opsonized Aspergillus conidia. Human granulocytes (A,B), monocytes (C), macrophages (D), dendritic cells (E) and microglia (F) were seeded on cover slips. Aspergillus conidia were stained with calcofluor white, either mock-treated in GVB buffer or opsonized in serum or normal CSF for 45 min, and added to the cells. Four parallel samples were analyzed microscopically, the percentage of phagocytosing cells with ingested conidia and the phagocytic index (number of ingested conidia per phagocytosing cell) were determined. The statistical significance of phagocytosis of conidia incubated with serum or CSF versus phagocytosis of buffer-treated conidia was evaluated by Student’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.005.
vation. Our results showed that EDTA nearly completely abolished the positive serum effect, indicating that complement is the predominant factor responsible for the enhanced phagocytosis (data not shown). 3.3. Induction of oxidative burst by opsonized versus non-opsonized Aspergillus conidia To further clarify the role of complement in the antifungal defense of brain tissue we investigated the influence of
1290
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
Fig. 4. Oxidative burst and intracellular killing of Aspergillus conidia by granulocytes. (A) Freshly isolated granulocytes were loaded with the dye 2′,7′-dichlorodihydrofluorescin diacetate, followed by stimulation by medium (1), PMA (2), or by Aspergillus conidia either incubated in GVB buffer (3), or opsonized with serum (4) or with normal CSF (5). Induction of oxidative burst was quantified after 45 min by FACS analysis of the radical-induced conversion of 2′,7′-dichlorodihydrofluorescin diacetate within the cells into its fluorescent metabolite. (B) Conidia of A. fumigatus 14 were incubated either in GVB buffer, in human serum or in CSF. Granulocytes were added to the conidia for 5 h at 37 °C, and cells were lysed in distilled water. Different volumes of the suspension were plated on Sabouraud agar plates and the number of fungal colonies was counted after 18 h. The amount of colony-forming units (CFU) per sample is presented as mean ± standard deviation (S.D.). from two parallel samples. The statistical significance of colony numbers of conidia versus conidia incubated with granulocytes was evaluated for each opsonization condition by Student’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.005.
opsonization by CSF versus opsonization by serum or mocktreatment in buffer on the induction of oxidative burst in infiltrating granulocytes. Aspergillus conidia were incubated in buffer, in CSF or in serum, and added to freshly isolated granulocytes. Oxidative burst was quantified by FACS analysis of the H2O2-dependent conversion of the dye 2′,7′-dichlorofluorescin diacetate into its fluorescent metabolite.
Buffer-treated Aspergillus conidia were not able to induce a significant oxidative burst signal in granulocytes (Fig. 4A); fluorescence of the cells corresponded to that of the unstimulated granulocytes (0.31% versus 0.36% of the cells within an arbitrary gate; Fig. 4A1, A3); PMA as positive control generally showed the capacity of the granulocytes to respond to external triggering (99.82% of the cells within the gate)
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
(Fig. 4A2). Opsonization of the conidia with serum induced a clear shift of the fluorescence signal of the cells compared to the non-opsonized conidia with 40.36% of the cells within the gate (Fig. 4A4). In contrast, the opsonization of the Aspergillus conidia with CSF was too weak to stimulate the granulocytes and the fluorescence signal was the same as with non-opsonized conidia (0.31% of the cells in the gate) (Fig. 4A5). 3.4. Effıciency of fungal killing in correlation to the opsonization of the conidia Since the antimicrobial activity of granulocytes also includes non-oxidative mechanisms and also affects noningested pathogens by neutrophil degranulation and release of toxic factors, we analyzed the effect of complement opsonization on fungal killing. Granulocytes were added to non-opsonized conidia and to conidia opsonized in serum or normal CSF, respectively; the neutrophils were lysed after 5 h and the number of surviving conidia was quantified by plating the different suspensions on agar plates and counting of fungal colonies. Granulocytes were not able to mediate effective killing of non-opsonized Aspergillus conidia; although the number of fungal colonies was lower than without neutrophils, this difference did not reach significance (Fig. 4B). Opsonization of the conidia in human serum enabled the granulocytes to drastically lower the number of viable conidia and the reduction
1291
was highly significant. When the conidia had been opsonized in non-inflammatory CSF the reduction of viable conidia by granulocytes corresponded to that of the non-opsonized conidia, indicating that the complement levels in normal CSF were not sufficient to support effective antifungal attack by granulocytes. 3.5. Complement levels, opsonization capacity and activation of phagocytosis by CSF of a patient with cerebral aspergillosis We had the opportunity to obtain three CSF samples from a patient who was diagnosed with cerebral aspergillosis by direct culture from CSF. The available CSF samples were taken on day 18, 27 and 32 after admission to hospital with neurological symptoms; the patient deceased on day 42. The concentration of different complement factors in the CSF samples was determined by ELISA and compared to normal non-inflammatory CSF. A highly significant increase in the level of C1q, C3 and C5 was measured for all three CSF samples from the patient with cerebral aspergillosis compared to two different pools of non-inflammatory CSF (Fig. 5A–C). The amount of C1q, C3 and C5 reached a peak on day 27 and decreased again, probably due to intense destruction of the complement-producing brain tissue. The appearance of C3a, a proteolytic cleavage product of C3, had a peak on day 32 which might indicate the activity of proteolytic enzymes after predominant cell lysis (Fig. 5D). The
Fig. 5. Concentration of complement factors and complement activation products in the CSF of a patient with cerebral aspergillosis. Three CSF samples from a patient with cerebral aspergillosis were obtained on day 18, 27 and 32 after hospitalization due to neurological symptoms (CA18 d, CA27 d, CA32 d). Levels of C1q (A), C3 (B), C5 (C), C3a (D) and TCC (E) in these patient samples were determined by ELISAs and compared to levels in two different pools of non-inflammatory CSF (CSF1, CSF2). The mean concentration ± S.D. from three parallel samples are presented. The statistical significance of complement levels in inflammatory CSF versus normal CSF was evaluated by Student’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.005.
1292
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
level of the soluble terminal complement complex, a late activation product of the complement cascade, showed the same kinetic as C1q, C3 and C5 (Fig. 5E). To test whether increased complement levels in inflammatory CSF reflect an improved capacity to opsonize the fungus we incubated the hyphae with the different CSF samples and detected the deposition by immunofluorescence with specific antibodies. All three CSF samples derived from the patient with cerebral aspergillosis induced a higher complement deposition on the fungal surface than non-inflammatory CSF (Fig. 6). The intensity directly correlated with the complement levels of C1q, C3 and C5 in CSF with the sample taken after 27 days resulting in the strongest opsonization, followed by the sample taken after 32 days. However, all inflammatory CSF samples did not reach the opsonization capacity of serum. The induction of phagocytosis by conidia opsonized in the different CSF samples was also investigated. Opsonization of conidia in the inflammatory CSF samples derived from the patient with cerebral aspergillosis stimulated a higher percentage of granulocytes to ingest the conidia, compared to the non-inflammatory CSF (Fig. 7A). Similarly, the phagocytic index, i.e. the number of conidia per phagocytosing cell was slightly higher with conidia incubated with the inflammatory CSF samples (Fig. 7B). Opsonization of the conidia in serum induced the highest stimulation of phagocytosis. Furthermore, the inflammatory CSF samples induced at least partly a more efficient killing by granulocytes. Opsonization of conidia with the CSF obtained on day 27 significantly enabled the granulocytes to reduce the number of viable conidia (Fig. 8). The granulocyte-induced reduction of conidia opsonized by CSF taken on day 32 was not significant; the
Fig. 7. Phagocytosis of Aspergillus conidia opsonized by non-inflammatory or inflammatory CSF. Human granulocytes were seeded on cover slips. Aspergillus conidia were stained with calcofluor white, either mock-treated in GVB buffer or opsonized for 45 min in serum, non-inflammatory CSF or in CSF samples from a patient with cerebral aspergillosis, obtained 18 days (CA18 d), 27 days (CA27 d) and 32 days (CA32 d) after hospitalization due to neurological symptoms. The conidia were added to the cells, and four parallel samples were analyzed microscopically to determine the percentage of phagocytosing cells with ingested conidia and the phagocytic index (number of ingested conidia per phagocytosing cell). The statistical significance of phagocytosis of conidia incubated with serum or CSF versus phagocytosis of buffertreated conidia was evaluated by Student’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.005.
sample from day 18 and the non-inflammatory CSF were unable to stimulate killing by granulocytes. Again, serum was the most efficient complement source to support the killing activity of granulocytes. 4. Discussion
Fig. 6. Opsonization of fungal hyphae by the CSF from a patient with cerebral aspergillosis. The clinical isolate A. fumigatus 14 was grown on slides, fixated and the hyphae were incubated either with non-inflammatory CSF (A) or with CSF samples from a patient with cerebral aspergillosis, obtained 18 days (B), 27 days (C) and 32 days (D) after hospitalization due to neurological symptoms. Deposition of complement factor C3 was demonstrated by immunofluorescence with a specific 具 C3d-antibody.
The limitation of cerebral immune response by restricted access of peripheral immune effectors and by establishment of an antiinflammatory environment protects the sensitive brain tissue from harmful inflammation [20]. This protection takes a high price, making an effective antifungal response very difficult. The complement system is one of the few locally produced weapons that enable a fast and reliable antifungal reaction. However, the constitutive complement synthesis in the brain is low, and the high mortality rate of patients with
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
Fig. 8. Granulocyte-mediated killing of Aspergillus conidia opsonized by non-inflammatory CSF versus CSF derived from cerebral aspergillosis. Conidia of A. fumigatus 14 were incubated either in human serum, in noninflammatory CSF or in CSF samples from a patient with cerebral aspergillosis, obtained 18 days (CA18 d), 27 days (CA27 d) and 32 days (CA32 d) after hospitalization due to neurological symptoms. Granulocytes were added to the conidia for 5 h at 37 °C, followed by cell lysis in distilled water. Different volumes of the suspension were plated on Sabouraud agar plates and the number of fungal colonies was counted after 18 h. The amount of colonyforming units per sample is presented as mean ± S.D. from two parallel samples. The statistical significance of colony numbers of conidia versus conidia incubated with granulocytes was evaluated for each opsonization condition by Student’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.001.
cerebral aspergillosis indicates that the local complement attack lacks the capacity to eliminate the invading pathogen from CNS. We, therefore, investigated which antifungal immune responses can be induced or supported by the low complement amounts in CSF. Our experiments showed that the constitutive complement levels in CSF enable a clear opsonization of fungal hyphae, albeit much weaker than by serum. The ratio of opsonization intensity between serum and CSF of 1:400 corresponds approximately to the ratio of C3 levels of serum and CSF [19], but the concentration of other factors might also be limiting. Complement deposition on fungal pathogens does not result in direct killing via formation of membrane attack complexes, presumably as a consequence of their thick cell wall. Thus the importance of the complement system for the antifungal defense is supposed to be support of phagocytosis, a critical component of host defense against fungal infection [21]. Phagocytosis is complement-dependent with a clear correlation between deposition of C3 on Aspergillus and its ingestion by human neutrophils [22], since opsonization provides an improved interaction between the fungal surface and complement receptor-bearing phagocytic cells [23–25]. We could show that opsonization of Aspergillus by CSF-derived complement does not achieve the intensity to effectively support phagocytosis of Aspergillus by microglial cells; similarly, infiltrating phagocytes from the periphery (neutrophils, monocytes, macrophages, dendritic cells) are insufficiently stimulated. A slight increase in the uptake rate of CSFopsonized conidia compared to non-opsonized conidia was revealed only for monocytes and microglia. This low support of phagocytosis by cerebral complement might contribute to
1293
the immune failure in cerebral aspergillosis, since there is a correlation between the pathogenicity of Aspergillus and its capacity to evade complement attack and thus limit phagocytosis [10,25]. Following complement receptor-mediated uptake of opsonized particles neutrophils start a powerful microbicidal program with production of oxidative metabolites that destroy the engulfed material. Oxidative mechanisms damage the fungi by producing protein modifications, nucleic acid breaks and lipid peroxidation [26]. Opsonization has a profound effect on induction of oxidative burst in neutrophils [27]. Our experiments showed that, whereas opsonization with serum triggered an oxidative burst in neutrophils, the deposition of complement from CSF was not intense enough to induce this effect. Polymorphonuclear leukocytes not only destroy engulfed pathogens, but can also damage extracellular particles [28–30,8]. For that reason we also quantified directly the killing after contact of neutrophils with opsonized conidia. Whereas granulocytes effectively killed serum-opsonized conidia, fungal viability decreased minimally when granulocytes were incubated with CSF-opsonized conidia. Thus the constitutively present cerebral complement system is no effective weapon against invading Aspergillus, and this lack of efficacy presumably contributes to the high mortality in cerebral aspergillosis. Therefore, enhanced complement synthesis might improve the defense reaction. The CSF of the patient with cerebral aspergillosis contained highly elevated concentrations of C1q (up to 14-fold increase compared to non-inflammatory CSF), C3 and C5 (17-fold increase). This enhancement in complement levels can result either from a leaky blood–brain barrier or from local production in the brain. Immunohistochemical analysis of brain tissue derived from patients with cerebral aspergillosis revealed that upregulation of local production at least contributes to the elevated complement concentrations in CSF (Rambach et al., in preparation). Astrocytes, microglia, neurons and oligodendrocytes were demonstrated in vitro to produce almost all complement proteins in low amounts [12,31] that can be increased by inflammatory stimuli like lipopolysaccharides, viral infection and cytokines. The mechanism of complement upregulation in cerebral aspergillosis needs further investigation and may result from direct interaction with Aspergillus surface molecules and/or from indirect effects of pro-inflammatory cytokines. The potentially protective effects of complement upregulation for the brain are obvious: improved opsonization of fungi with better induction of phagocytosis, stimulation of oxidative burst and a higher killing rate. Similarly, higher production of C1q may support directly the triggering of oxidative burst, chemotaxis and phagocytosis via binding to its receptor on human neutrophils [32]. Furthermore, the anaphylatoxin C3a, which is present in high amounts in the inflammatory CSF samples, is a chemoattractant that recruits inflammatory cells to the site of infection [33]. C3a and C5a can also stimulate microglia and astrocytes via induction of pro-inflammatory cytokines [31] that might again contribute
1294
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
to the upregulation of complement expression in a feedback loop. In addition, C5a primes neutrophils and enhances the NADPH oxidase activity, a key enzyme of oxidative burst, in response to contact to opsonized zymosan [34]. TCC was described to have stimulatory effects on a number of target cells in the brain and to promote neuronal survival [31]. However, the amount of complement in the CSF of the patient with cerebral aspergillosis never reached the level of the serum. Whereas serum contains 1–1.5 mg/ml C3, the C3 concentration in the inflammatory CSF samples was maximally 38 µg/ml (Fig. 5). Our experiments showed that these complement concentrations enabled better fungal opsonization and improved phagocytosis and (at least partly) granulocyte-mediated killing, but the capacity was far from that of serum. The reasons for this limitation of complement synthesis in the inflamed brain tissue are unknown. One explanation might be that the restriction aims to prevent inflammation-induced tissue damage, since there is considerable evidence that high complement biosynthesis and activation contribute to the pathogenesis of neurodegenerative disorders, leading to neuronal loss and inflammation [12]. Chronic complement activation is associated with neurodegeneration in Alzheimer’s disease and multiple sclerosis. Beside induction of a harmful inflammatory reaction, complement-induced brain damage might also include opsonization of surrounding “self” cells with subsequent phagocytosis and bystander lysis through generation of the membrane attack complex. Neurons are extremely susceptible to lysis even mediated by homologous complement, since they constitutively express only low amounts of inhibitors like CD55, CD59, and MCP [35]. Complement-stimulated neutrophils are involved in tissue damage by an excessive acute inflammatory response, inducing oxidative stress to sensitive bystander cells [36]. An alternative explanation for the limited effect of complement upregulation might be that this occurs too late in the course of infection and is limited by the Aspergillus-induced tissue damage, leaving only few cells unaffected and thus capable of synthesizing complement. If this hypothesis is confirmed by further experiments, the specific manipulation of the level of biosynthesis and/or activation of the local complement system (e.g. using recombinant cytokines, or increase of expression by gene therapy) might be of potential therapeutic value and represent a supportive immune-based therapy to improve the treatment with antifungal drugs.
References [1] [2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11] [12]
[13]
[14]
[15]
[16]
[17]
Acknowledgements We gratefully acknowledge the generous gift of the monoclonal anti-C3 antibody BB5 from W. Prodinger (Department of Hygiene, Microbiology and Social Medicine, Innsbruck, Austria). This study was supported by LudwigBoltzmann-Society, FWF (project P15375), Österreichische Nationalbank (project 9374), BMSG and the State of Tyrol.
[18]
[19]
D.W. Denning, Invasive aspergillosis, Clin. Infect. Dis. 26 (1998) 781–805. R.C. Young, J.E. Bennett, C.L. Vogel, P.P. Carbone, V.T. DeVita, Aspergillosis: the spectrum of the disease in 98 patients, Medicine 49 (1970) 147–173. M.E. Hagensee, J.E. Bauwens, K. Bent, R.A. Bowden, Brain abscess following marrow transplantation: experience at the Fred Hutchinson Cancer Research Center, 1984–1992, Clin. Infect. Dis. 19 (1994) 402–408. N.P. Hailer, F.L. Heppner, D. Haas, R. Nitsch, Astrocytic factors deactivate antigen presenting cells that invade the central nervous system, Brain Pathol. 8 (1998) 459–474. P.J.H. Jongen, W.H. Doersburg, J.L.M. Ibrahim-Stappers, W.A.J.G. Lemmens, O.R. Hommes, K.J.B. Lamers, Cerebrospinal fluid C3 and C4 indexes in immunological disorders of the central nervous system, Acta Neurol. Scand. 101 (2000) 116–121. C. Speth, R. Würzner, H. Stoiber, M.P. Dierich, The complement system: pathophysiology and clinical relevance, Wien. Klin. Wochenschr. 111 (1999) 378–391. T.R. Kozel, M.A. Wilson, T.P. Farrell, S.M. Levitz, Activation of C3 and binding to Aspergillus fumigatus conidia and hyphae, Infect. Immun. 57 (1989) 3412–3417. J.E. Sturtevant, J.P. Latgé, Interactions between conidia of Aspergillus fumigatus and human complement component C3, Infect. Immun. 60 (1992) 1913–1918. R.F. Hector, E. Yee, M.S. Collins, Use of DBA/2N mice in models of systemic candidiasis and pulmonary and systemic aspergillosis, Infect. Immun. 58 (1990) 1476–1478. S. Henwick, S.V. Hetherington, C.C. Patrick, Complement binding to Aspergillus conidia correlates with pathogenicity, J. Lab. Clin. Med. 122 (1993) 27–35. B.P. Morgan, P. Gasque, Expression of complement in the brain: role in health and disease, Immunol. Today 17 (1996) 461–466. P. Gasque, Y.D. Dean, E.P. McGreal, J. VanBeek, B.P. Morgan, Complement components of the innate immune system in health and disease in the CNS, Immunopharmacology 49 (2000) 171–186. C. Speth, G. Rambach, C. Lass-Flörl, R. Würzner, P. Gasque, I. Mohsenipour, M.P. Dierich, Culture supernatants of patient-derived Aspergillus isolates have toxic and lytic activity towards neurons and glial cells, FEMS Immunol. Med. Microbiol. 29 (2000) 303–313. R. Würzner, M. Schulze, L. Happe, A. Franzke, F.A. Bieber, M. Oppermann, O. Götze, Inhibition of terminal complement complex formation and cell lysis by monoclonal antibodies, Complement Inflamm. 8 (1991) 328–340. M.A. Hess, M.G. Schwendinger, E.L. Eskelinen, K. Pfaller, M. Pavelka, M.P. Dierich, W.M. Prodinger, Tracing uptake of C3dgconjugated antigen into B cells via complement receptor type 2 (CR2, CD21), Blood 95 (2000) 2617–2623. M. Barcova, C. Speth, L. Kacani, F. Überall, H. Stoiber, M.P. Dierich, Involvement of adenylate cyclase and p70S6-kinase activation in IL-10 up-regulation in human monocytes by gp41 envelope protein of human immunodeficiency virus type 1, Eur. J. Physiol. 437 (1999) 538–546. D.A. Bass, J.W. Parce, L.R. Dechatelet, P. Szejda, M.C. Seeds, M. Thomas, Flow cytometric studies of oxidative product formation by neutrophils: a graded response to membrane stimulation, J. Immunol. 130 (1983) 1910–1917. L.E. Leigh, B. Ghebrehiwet, T.P. Perera, I.N. Bird, P. Strong, U. Kishore, K.B. Reid, P. Eggleton, C1q-mediated chemotaxis by human neutrophils: involvement of gC1qRp and G-protein signalling mechanisms, Biochem. J. 330 (1998) 247–254. W.M. Prodinger, R. Würzner, H. Stoiber, M.P. Dierich, in: W.E. Paul (Ed.), Fundamental Immunology, Lippincott-Raven, Philadelphia, 2003, pp. 1077–1103.
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295 [20] Z. Fabry, C.S. Raine, M.N. Hart, Nervous tissue as an immune compartment: the dialect of the immune response in the CNS, Immunol. Today 15 (1994) 218–224. [21] A.R. Waldorf, Pulmonary defense mechanisms against opportunistic fungal pathogens, Immunol. Ser. 47 (1989) 243–270. [22] J.E. Sturtevant, J.P. Latgé, Participation of complement in the phagocytosis of the conidia of Aspergillus fumigatus by human polymorphonuclear cells, J. Infect. Dis. 166 (1992) 580–586. [23] R.A. Calderone, L. Linehan, The role of complement in host resistance to systemic fungal infection, Immunol. Ser. 47 (1989) 225–242. [24] J.W. Murphy, Immunity to fungi, Curr. Opin. Immunol. 2 (1990) 360–367. [25] H.F. Tsai, Y.C. Chang, R.G. Washburn, M.H. Wheeler, K.J. KwonChung, The developmentally regulated alb1 gene of Aspergillus fumigatus: its role in modulation of conidial morphology and virulence, J. Bacteriol. 180 (1998) 3031–3038. [26] M.K. Mansour, S.M. Levitz, Interaction of fungi with phagocytes, Curr. Opin. Microbiol. 5 (2002) 359–365. [27] A.A. Vaporciyan, P.A. Ward, Enhanced generation of O2– by human neutrophils via complement iC3b/Mac-1 interaction, Biol. Signals 2 (1993) 126–135. [28] R.D. Diamond, R. Krzesicki, W. Jao, Damage to pseudohyphal forms of Candida albicans by neutrophils in the absence of serum in vitro, J. Clin. Invest. 61 (1978) 349–359.
1295
[29] R.D. Diamond, R. Krzesicki, B. Epstein, W. Jao, Damage to hyphal forms of fungi by human leukocytes in vitro. A possible host defense mechanism in aspergillosis and mucormycosis, Am. J. Pathol. 91 (1978) 313–328. [30] A.R. Waldorf, R.D. Diamond, Neutrophil chemotactic responses induced by fresh and swollen Rhizopus oryzae spores and Aspergillus fumigatus conidia, Infect. Immun. 48 (1985) 458–463. [31] C. Speth, M.P. Dierich, P. Gasque, Neuroinvasion by pathogens: a key role of the complement system, Mol. Immunol. 38 (2002) 669–679. [32] L.E. Leigh, B. Ghebrehiwet, T.P. Perera, I.N. Bird, P. Strong, U. Kishore, K.B. Reid, P. Eggleton, C1q-mediated chemotaxis by human neutrophils: involvement of gC1qRp and G-protein signalling mechanisms, Biochem. J. 330 (1998) 247–254. [33] S.M. Levitz, Overview of host defenses in fungal infections, Clin. Infect. Dis. 14 (Suppl 1) (1992) 37–42. [34] J.G. Bender, L.C. McPhail, D.E. Van Epps, Exposure of human neutrophils to chemotactic factors potentiates activation of the respiratory burst enzyme, J. Immunol. 130 (1983) 2316–2323. [35] S.K. Singhrao, J.W. Neal, N.K. Rushmere, B.P. Morgan, P. Gasque, Spontaneous classical pathway activation and deficiency of membrane regulators render human neurons susceptible to complement lysis, Am. J. Pathol. 157 (2000) 905–918. [36] H. Jaeschke, C.W. Smith, Mechanisms of neutrophil-induced parenchymal cell injury, J. Leukoc. Biol. 61 (1997) 647–653.
Original article
Antifungal activity of the local complement system in cerebral aspergillosis Günter Rambach a, Magdalena Hagleitner a, Iradj Mohsenipour b, Cornelia Lass-Flörl a, Hans Maier c, Reinhard Würzner a, Manfred P. Dierich a, Cornelia Speth a,* a
Department of Hygiene, Microbiology and Social Medicine, Innsbruck Medical University, and Ludwig-Boltzmann-Institute for AIDS Research, Fritz-Pregl-Str. 3, 6020 Innsbruck, Austria b Department of Neurosurgery, Innsbruck Medical University, Anichstr. 35, 6020 Innsbruck, Austria c Department of Pathological Anatomy, Innsbruck Medical University, Müllerstr. 44, 6020 Innsbruck, Austria Received 13 December 2004; accepted 22 April 2005 Available online 13 June 2005
Abstract Dissemination of aspergillosis into the central nervous system is associated with nearly 100% mortality. To study the reasons for the antifungal immune failure we analyzed the efficacy of cerebral complement to combat the fungus Aspergillus. Incubation of Aspergillus in non-inflammatory cerebrospinal fluid (CSF) revealed that complement levels were sufficient to obtain a deposition on the surface, but opsonization was much weaker than in serum. Consequently complement deposition from normal CSF on fungal surface stimulated a very low phagocytic activity of microglia, granulocytes, monocytes and macrophages compared to stimulation by conidia opsonized in serum. Similarly, opsonization of Aspergillus by CSF was not sufficient to induce an oxidative burst in infiltrating granulocytes, whereas conidia opsonized in serum induced a clear respiratory signal. Thus, granulocytes were capable of considerably reducing the viability of serum-opsonized Aspergillus conidia, but not of conidia opsonized in CSF. The limited efficacy of antifungal attack by cerebral complement can be partly compensated by enhanced synthesis, leading to elevated complement concentrations in CSF derived from a patient with cerebral aspergillosis. This inflammatory CSF was able to induce (i) a higher complement deposition on the Aspergillus surface than non-inflammatory CSF, (ii) an accumulation of complement activation products and (iii) an increase in phagocytic and killing activity of infiltrating granulocytes. However, levels and efficacy of the serum-derived complement were not reached. These data indicate that low local complement synthesis and activation may represent a central reason for the insufficient antifungal defense in the brain and the high mortality rate of cerebral aspergillosis. © 2005 Elsevier SAS. All rights reserved. Keywords: Cerebral aspergillosis; Innate immunity; Complement; Aspergillus; Phagocytes
1. Introduction Invasive aspergillosis is an opportunistic infection that contributes considerably to the morbidity and mortality among immunocompromized patients, including patients with acute leukemia, transplant recipients, patients with autoimmune diseases treated with immunosuppressive regimens, and AIDS patients [1]. A dangerous complication of invasive aspergillosis is the involvement of the brain, which occurs in 10–20% of all cases and has a mortality rate of nearly 100% [1]. Inva* Corresponding author. Tel.: +43 512 507 3405; fax: +43 512 507 2870. E-mail address: [email protected] (C. Speth). 1286-4579/$ - see front matter © 2005 Elsevier SAS. All rights reserved. doi:10.1016/j.micinf.2005.04.014
sion of the central nervous system (CNS) mostly occurs either via blood from a pulmonary focus, or directly by infection after head trauma or surgical procedures [2]. The clinical presentation includes lesions, brain abscess formation, granulomatous reactions, cerebral infarction, hemorrhage, meningoencephalitis and meningitis [1]. Prominent symptoms of the patients are fever, altered mental status, headache, hemiplegia and seizures [3]. Treatment regimens with antifungal agents are often of limited efficacy because of their poor penetration into the brain, underlining the relevance of an efficient local immune response. The reasons for the insufficient antifungal response in the brain are not yet known. In general, the brain is an immuno-
1286
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
privileged organ and thus protected from harmful inflammatory processes. This immunoprivilege is guaranteed by several factors, including (i) the blood–brain barrier that restricts the access of peripheral immune effectors like lymphocytes and antibodies; (ii) the deactivation of infiltrating granulocytes and monocytes/macrophages by resident astrocytes [4]; and the low complement levels in the cerebrospinal fluid (CSF) [5]. Complement is a central element of the innate immune system and enables a fast and effective antimicrobial defense. In general, activation of the complement cascade does not only induce lysis of invading pathogens but also opsonizes them to become a target for phagocytosis by complementreceptor bearing cells. In addition, the pro-inflammatory anaphylatoxins C3a and C5a generated by complement activation attract phagocytes to the site of infection and can activate different cell types via specific intracellular signal transduction pathways (reviewed in: [6]). Whereas the interaction of complement with Aspergillus spp. in the brain has not yet been studied, some aspects are known from the peripheral blood. Aspergillus conidia and hyphae are potent activators of the serum complement cascade by classical and alternative pathway leading to deposition of C3 on the fungal surface [7,8]. Complement deficiency is correlated with enhanced susceptibility to a disseminated infection by A. fumigatus [9]. The degree of complement deposition on Aspergillus correlates indirectly with pathogenicity, with highly virulent species binding less C3 on their surface than non-pathogenic species [10]. The CNS produces its own complement with astrocytes, neurons, microglia and oligodendrocytes being all able to synthesize complement factors [11,12]. However, the constitutive synthesis in the brain is low, making it unclear whether the complement levels in CSF are sufficient for a competent antifungal defense. We, therefore, studied the ability of cerebral complement to fulfill functions like opsonization, induction of phagocytosis and oxidative burst in phagocytic cells. Our experiments showed that the complement concentration in non-inflammatory CSF was sufficient for opsonization of fungal hyphae, but the intensity was much weaker than the opsonization in serum. Aspergillus conidia opsonized by human serum triggered an oxidative burst in granulocytes whereas the burst signal induced by CSF-opsonized conidia corresponded to that of non-opsonized conidia. Similarly, phagocytosis and lysis of fungal conidia opsonized by CSF was low in case of granulocytes, monocytes, macrophages and microglia, whereas a high percentage of these cell types ingested and killed serum-opsonized conidia. The concentration of complement in CSF was highly upregulated during the pathogenesis of cerebral aspergillosis, giving raise to increased opsonization of the fungal surface, higher phagocytosis and better killing by granulocytes. However, this upregulation might still be insufficient or occur too late to inhibit effectively the fungal infection and spreading of hyphae in the brain tissue.
2. Materials and methods 2.1. Aspergillus isolate and culture Aspergillus fumigatus isolate 14 was obtained from a hospitalized patient with cerebral aspergillosis; the patient suffered from acute myeloic leukemia and neutropenia as underlying disease [13]. The fungus was grown for at least 5 days on Sabouraud (Oxoid, Wesel, Germany) agar plates at 28 °C until sporulation was clearly visible. The conidia were swept off from sporulating colonies with phosphate-buffered saline (PBS) containing 0.05% Tween-20 (Sigma, St. Louis, USA) and kept at 4 °C. 2.2. Deposition of complement on fungal germination tubes Both human serum and CSF were used for studies of complement deposition on fungal hyphae. Human serum was obtained from five to six healthy individuals, pooled and frozen at –80 °C for further use. CSF pools were obtained from 15 individuals who were investigated for neurological noninflammatory diseases and also stored at –80 °C. For opsonization experiments 105 conidia were grown on prelubricated microscope slides at 28 °C until formation of germination tubes was seen. The adherent fungi were washed with PBS and fixated with acetone and methanol at –20 °C. Fixated fungi were incubated with serum or CSF at various concentrations, or, as a negative control, with GVB buffer (veronal buffer saline containing 0.1% gelatin, 0.15 mM CaCl2, 0.1 mM MgCl2; bioMérieux, Marcy l’Etoile, France). Deposition of complement factors was detected by standard indirect immunofluorescence procedure after 1.5 h for C1q, C4 and C3 and after 3.5 h for C5 and C7. Briefly, the slides were washed with PBS to remove serum or CSF, followed by blocking of unspecific binding with PBS/1% BSA (Sigma). The specific primary antibody was added for 1 h at 37 °C. After extensive washing the fluorescence-labeled secondary antibody was incubated for 30 min and visualized in a Zeiss Axioplan microscope. The following antibodies were used for the immunofluorescence analysis: polyclonal a-C1q, a-C4, a-C3d and a-C5 antibodies (Dako, Glostrup, Denmark); a monoclonal a-C7 [14]; a secondary goat-a-rabbit or goat-a-mouse Ig, labeled with the dye Alexa 488 (Molecular Probes, Leiden, The Netherlands). 2.3. Quantification of complement factors in CSF by ELISA Complement quantification of C1q, C3, C5 and TCC was performed by sandwich ELISA using the following antibodies: the monoclonal coating a-C3d antibody BB5 [15] and a rabbit polyclonal a-C3c (Dako, Glostrup, Denmark) for detection; for C1q a monoclonal coating antibody (Quidel, San
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
Diego, USA) and a rabbit polyclonal a-C1q (Dako); for C5 a monoclonal coating antibody (clone N19-8) [14] and a rabbit polyclonal detection antibody (Dako); for TCC a monoclonal coating antibody (clone WU 7-2) and a biotinylated polyclonal goat a-C6 [14]. Microplates were coated overnight at 4 °C with the corresponding coating antibody in 0.1 M NaHCO3, pH 9.6. Unspecific binding was inhibited by saturation with 1% BSA (Sigma) in PBS. CSF samples were added to the coated wells and incubated for 1 h at room temperature. After washing with PBS the bound complement was detected with the corresponding second antibody, followed by a peroxidaselabeled antibody (Dako). For TCC-ELISA an avidinhorseradish peroxidase conjugate (Dako) was used to detect the biotinylated goat a-C6 antibody. Development was performed with tetramethylbenzidine dihydrochloride (Sigma). Purified protein standards (C3: Sigma; C1q, C5: Quidel) were used to calculate the exact amounts of the complement factors in the supernatants. C3a was quantified using a commercially available kit (Quidel) according to manufacturers’ instruction, with a similar procedure as described above. 2.4. Isolation of monocytes, granulocytes, macrophages, dendritic cells and microglia For phagocytosis experiments granulocytes and monocytes were isolated from buffy coat by density gradient centrifugation over Ficoll–Paque (Pharmacia, Uppsala, Sweden); peripheral blood mononuclear cells (PBMC) layer and granulocyte layer were obtained separately. The granulocytes were further purified by lysis of erythrocytes with sterile distilled water for 20 s followed by washing with PBS. The monocytes were separated from lymphocytes using gelatin/plasma-coated plates as described previously [16]. Briefly, Petri dishes were coated with sterile 2% gelatin (Merck) solution for 2 h at 37 °C followed by incubation with autologous plasma for 30 min at 37 °C. PBMC obtained from Ficoll separation (Amersham, Buckinghamshire, UK) were suspended in RPMI medium (Life Technologies, Vienna) containing 10% FCS, 5% pooled serum and 2 mM L-glutamine and loaded onto gelatin/plasma-coated plates at 4 × 106 cells/ml. After 40 min of incubation at 37 °C peripheral blood lymphocytes (PBL) were washed out with medium; adherent monocytes were detached by 5 mM EDTA (Sigma) and further cultured in complete RPMI medium at 37 °C in a humidified 5% CO2 atmosphere. Fluorescence-activated cell sorter (FACS) analysis revealed that the cell population regularly consisted of 94–96% monocytes. Macrophages were obtained by differentiation of monocytes with human AB serum (Sigma) for at least 7 days until morphological changes were clearly visible. For oxidative burst experiments a crude preparation of granulocytes was obtained from heparinized blood. Aliquots were incubated with lysis buffer (150 mM ammonium chloride, 10 mM potassium hydrogen carbonate, 0.1 mM EDTA)
1287
for 5 min at room temperature. The suspension was centrifuged for 5 min at 1000 rpm and the pellet incubated again for 5 min in lysis buffer, followed by centrifugation. The cells were washed twice in PBSg (PBS with 0.2% gelatin (Merck, Darmstadt, Germany)) and finally resuspended in 300 µl PBSg. Monocyte-derived mature dendritic cells were obtained from previously isolated monocytes. For differentiation monocytes were cultivated in complete RPMI medium supplemented with 1500 U/ml IL-4 and 1600 U/ml GM-CSF at a density of 106 cells/ml in six-well plates. After 2 days 1000 U/ml IL-4 and 1600 U/ml GM-CSF were added. On day 5, cells were collected, washed and transferred to a new plate. Cells were cultivated further for 2 days in the presence of fresh complete RPMI medium supplemented with 100 U/ml IL-4, 1600 U/ml GM-CSF and 100 ng/ml lipopolysaccharide (LPS; Sigma) to obtain mature dendritic cells. Primary microglia from human adult brain were isolated from surgery material. The tissue was dissected using enzymatic digestion with 0.5% DNase and 0.5% collagenase (Sigma) and washed twice in DMEM culture medium. The cell suspension was purified from myelin by a percoll gradient (Amersham Biosciences, Little Chalfont, UK), washed in cell culture medium and cultured in DMEM containing 10% FCS, 2 mM glutamine and 1 × non-essential amino acids (Life Technologies), supplemented with 25 ng/ml GM-CSF (Peprotech, London, UK). The percentage of microglia and astrocytes was estimated by double immunofluorescence using an 具 -CD68 antibody (Dako) for microglia and an a-GFAP antibody (Dako) for astrocytes. The microglia preparation used for the experiments presented here consisted of 70% microglia and 30% astrocytes. For analysis of phagocytosis only microglia were counted, since these two cell types have a different morphology and can be distinguished easily. 2.5. Oxidative burst in granulocytes To quantify oxidative burst, the granulocytes were loaded for 15 min with the diacetate form of 2′,7′-dichlorodihydrofluorescin (Sigma). This cell-permeable substance is deesterified intracellularly by esterases concentrating the polar non-fluorescent DCFH in the cytoplasm. During the oxidative burst, DCFH is converted to highly fluorescent 2′,7′dichlorofluorescin diacetate through oxidation mainly by H2O2. The fluorescence of the cells after stimulation was analyzed by FACS gating the neutrophils [17]. Aspergillus conidia were incubated for 60 min either in GVB at 37 °C (non-opsonized conidia) or opsonized under the same conditions in serum or CSF. Thirty conidia per cell were added to each sample and the oxidative burst was measured by FACS at different time points. As a positive control for the capacity of granulocytes to react on stimulation, 10 mM 4-a-phorbol 12-myristate 13-acetate (PMA; Sigma) was added to one sample and the following activation of the dye was analyzed in parallel. Analysis was performed using a FACScan flow cytometer (Becton Dickinson, San Diego).
1288
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
2.6. Phagocytosis assay Aspergillus conidia were stained with Calcofluor White (Polysciences, Warrington, USA) for 2 min, followed by washing with PBS. For opsonization, the stained conidia were incubated for 60 min at 37 °C with serum, CSF, or with GBV as a negative control. Thirty conidia per cell were added for 20–40 min. Uningested conidia were removed by washing and the cells were analyzed microscopically. The following indices were calculated after counting 50–100 cells [18]: % phagocytosis, determined as percentage of cells that had one or more conidia phagocytosed; phagocytic index (PI) as the average number of conidia that had been phagocytosed by each phagocytosing cell. 2.7. Quantification of fungal killing Aspergillus conidia were opsonized by incubation for 60 min at 37 °C in serum, CSF samples, or in GVB as negative control. Granulocytes isolated from buffy coat were added to the conidia in an effector/target ratio of 1:1.5 with a final volume of 300 µl. After 5 h the neutrophils were lysed by adding 5.7 ml distilled water for 5 min. Different volumes of this suspension were plated on Sabouraud agar plates and the number of colonies was counted after 24 h. 2.8. Statistical analysis Statistical analysis (Student’s t-test) was performed using the Origin 6.1 software (Microcal).
3. Results 3.1. Deposition of complement from CSF on Aspergillus hyphae To investigate whether the low constitutive complement levels in normal CSF are sufficient to obtain opsonization, Aspergillus hyphae were incubated either in buffer or in normal CSF and the binding of complement was detected subsequently by immunofluorescence. Complement deposition on the fungal surface could be shown for various complement factors, including C1q, C4, C3, C5 and C7 (Fig. 1). The signals were weak, but clearly visible compared to the hyphae incubated in buffer, indicating that complement levels in noninflammatory human CSF are sufficient for opsonization of the fungus. To compare the opsonization capacity of normal CSF with the one of human serum, different concentrations of serum in comparison to CSF were used for complement deposition experiments, and C3 on the hyphal surface was again visualized by immunofluorescence. A strong deposition signal was obtained with serum at a dilution of 1:100 and a weak signal after diluting serum 1:1200 (Fig. 2A–C). The intensity of
Fig. 1. Deposition of complement on Aspergillus hyphae. The clinical isolate A. fumigatus 14 was grown on slides, fixated and the hyphae were incubated either in GVB buffer (A, C, E, G, I) or normal CSF (B, D, F, H, K). Deposition of the complement factors C1q (A, B), C4 (C, D), C3 (E, F), C5 (G, H) and C7 (I, K) on the hyphae was detected with specific antibodies and visualized by a fluorescence-labeled second antibody.
opsonization in CSF corresponded to the one obtained in serum at a dilution between 1:100 and 1:400 (Fig. 2D), dilution of CSF abolished the opsonization signal. This corresponds to the fact that C3 concentration in serum (1200 g/ml; [19]) is approximately 400-fold higher than in CSF (2.8 g/ml) [5]).
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
1289
Fig. 2. Deposition of complement factor C3 from serum or normal CSF on Aspergillus hyphae. The clinical isolate A. fumigatus 14 was grown on slides, fixated and the hyphae were incubated either with serum diluted 1:100 (A), 1:400 (B), and 1:1200 (C) or with undiluted CSF (D). Deposition of the complement factor C3 was demonstrated by immunofluorescence using a C3d-antibody.
3.2. Modulation of phagocytic activity by opsonization of Aspergillus conidia To test that whether opsonization of the fungal surface by CSF is strong enough to support the phagocytosis of Aspergillus by infiltrating granulocytes, granulocytes were incubated with A. fumigatus conidia either mock-treated in buffer or opsonized in serum or CSF; the percentage of phagocytosing granulocytes and the phagocytic index PI (average number of ingested conidia per phagocytosing cell) were quantified microscopically. Whereas opsonization of the conidia in serum strongly stimulated phagocytosis by granulocytes with 70% of cells showing phagocytic activity and a phagocytic index of 3.5 conidia per cell, the triggering by CSF-opsonized conidia corresponded to that of non-opsonized conidia with 18% phagocytosing cells and a PI of 1.7 conidia per cell (Fig. 3A, B). Similarly, the percentage of phagocytosing monocytes, macrophages and mature dendritic cells was much lower with conidia opsonized in CSF than with serum-opsonized conidia (32% versus 68%; 28% versus 85%; 16% versus 99%, respectively) (Fig. 3C–E). Likewise, microglial cells have a much higher phagocytic activity after contact with serum-opsonized conidia than with CSF-opsonized conidia (86% versus 33%) (Fig. 3F). For monocytes, macrophages and microglia the phagocytosis rate was slightly higher for CSF-opsonized than for non-opsonized conidia, and the difference even reached significance for monocytes and microglia, indicating that the weak opsonization by CSF may slightly support the uptake. To exclude that factors others than complement products might contribute to enhanced phagocytosis after opsonization in serum, we performed control experiments, adding 20 mM EDTA to serum and CSF during opsonization of the conidia. EDTA is an efficient inhibitor of complement acti-
Fig. 3. Phagocytosis of opsonized and non-opsonized Aspergillus conidia. Human granulocytes (A,B), monocytes (C), macrophages (D), dendritic cells (E) and microglia (F) were seeded on cover slips. Aspergillus conidia were stained with calcofluor white, either mock-treated in GVB buffer or opsonized in serum or normal CSF for 45 min, and added to the cells. Four parallel samples were analyzed microscopically, the percentage of phagocytosing cells with ingested conidia and the phagocytic index (number of ingested conidia per phagocytosing cell) were determined. The statistical significance of phagocytosis of conidia incubated with serum or CSF versus phagocytosis of buffer-treated conidia was evaluated by Student’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.005.
vation. Our results showed that EDTA nearly completely abolished the positive serum effect, indicating that complement is the predominant factor responsible for the enhanced phagocytosis (data not shown). 3.3. Induction of oxidative burst by opsonized versus non-opsonized Aspergillus conidia To further clarify the role of complement in the antifungal defense of brain tissue we investigated the influence of
1290
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
Fig. 4. Oxidative burst and intracellular killing of Aspergillus conidia by granulocytes. (A) Freshly isolated granulocytes were loaded with the dye 2′,7′-dichlorodihydrofluorescin diacetate, followed by stimulation by medium (1), PMA (2), or by Aspergillus conidia either incubated in GVB buffer (3), or opsonized with serum (4) or with normal CSF (5). Induction of oxidative burst was quantified after 45 min by FACS analysis of the radical-induced conversion of 2′,7′-dichlorodihydrofluorescin diacetate within the cells into its fluorescent metabolite. (B) Conidia of A. fumigatus 14 were incubated either in GVB buffer, in human serum or in CSF. Granulocytes were added to the conidia for 5 h at 37 °C, and cells were lysed in distilled water. Different volumes of the suspension were plated on Sabouraud agar plates and the number of fungal colonies was counted after 18 h. The amount of colony-forming units (CFU) per sample is presented as mean ± standard deviation (S.D.). from two parallel samples. The statistical significance of colony numbers of conidia versus conidia incubated with granulocytes was evaluated for each opsonization condition by Student’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.005.
opsonization by CSF versus opsonization by serum or mocktreatment in buffer on the induction of oxidative burst in infiltrating granulocytes. Aspergillus conidia were incubated in buffer, in CSF or in serum, and added to freshly isolated granulocytes. Oxidative burst was quantified by FACS analysis of the H2O2-dependent conversion of the dye 2′,7′-dichlorofluorescin diacetate into its fluorescent metabolite.
Buffer-treated Aspergillus conidia were not able to induce a significant oxidative burst signal in granulocytes (Fig. 4A); fluorescence of the cells corresponded to that of the unstimulated granulocytes (0.31% versus 0.36% of the cells within an arbitrary gate; Fig. 4A1, A3); PMA as positive control generally showed the capacity of the granulocytes to respond to external triggering (99.82% of the cells within the gate)
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
(Fig. 4A2). Opsonization of the conidia with serum induced a clear shift of the fluorescence signal of the cells compared to the non-opsonized conidia with 40.36% of the cells within the gate (Fig. 4A4). In contrast, the opsonization of the Aspergillus conidia with CSF was too weak to stimulate the granulocytes and the fluorescence signal was the same as with non-opsonized conidia (0.31% of the cells in the gate) (Fig. 4A5). 3.4. Effıciency of fungal killing in correlation to the opsonization of the conidia Since the antimicrobial activity of granulocytes also includes non-oxidative mechanisms and also affects noningested pathogens by neutrophil degranulation and release of toxic factors, we analyzed the effect of complement opsonization on fungal killing. Granulocytes were added to non-opsonized conidia and to conidia opsonized in serum or normal CSF, respectively; the neutrophils were lysed after 5 h and the number of surviving conidia was quantified by plating the different suspensions on agar plates and counting of fungal colonies. Granulocytes were not able to mediate effective killing of non-opsonized Aspergillus conidia; although the number of fungal colonies was lower than without neutrophils, this difference did not reach significance (Fig. 4B). Opsonization of the conidia in human serum enabled the granulocytes to drastically lower the number of viable conidia and the reduction
1291
was highly significant. When the conidia had been opsonized in non-inflammatory CSF the reduction of viable conidia by granulocytes corresponded to that of the non-opsonized conidia, indicating that the complement levels in normal CSF were not sufficient to support effective antifungal attack by granulocytes. 3.5. Complement levels, opsonization capacity and activation of phagocytosis by CSF of a patient with cerebral aspergillosis We had the opportunity to obtain three CSF samples from a patient who was diagnosed with cerebral aspergillosis by direct culture from CSF. The available CSF samples were taken on day 18, 27 and 32 after admission to hospital with neurological symptoms; the patient deceased on day 42. The concentration of different complement factors in the CSF samples was determined by ELISA and compared to normal non-inflammatory CSF. A highly significant increase in the level of C1q, C3 and C5 was measured for all three CSF samples from the patient with cerebral aspergillosis compared to two different pools of non-inflammatory CSF (Fig. 5A–C). The amount of C1q, C3 and C5 reached a peak on day 27 and decreased again, probably due to intense destruction of the complement-producing brain tissue. The appearance of C3a, a proteolytic cleavage product of C3, had a peak on day 32 which might indicate the activity of proteolytic enzymes after predominant cell lysis (Fig. 5D). The
Fig. 5. Concentration of complement factors and complement activation products in the CSF of a patient with cerebral aspergillosis. Three CSF samples from a patient with cerebral aspergillosis were obtained on day 18, 27 and 32 after hospitalization due to neurological symptoms (CA18 d, CA27 d, CA32 d). Levels of C1q (A), C3 (B), C5 (C), C3a (D) and TCC (E) in these patient samples were determined by ELISAs and compared to levels in two different pools of non-inflammatory CSF (CSF1, CSF2). The mean concentration ± S.D. from three parallel samples are presented. The statistical significance of complement levels in inflammatory CSF versus normal CSF was evaluated by Student’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.005.
1292
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
level of the soluble terminal complement complex, a late activation product of the complement cascade, showed the same kinetic as C1q, C3 and C5 (Fig. 5E). To test whether increased complement levels in inflammatory CSF reflect an improved capacity to opsonize the fungus we incubated the hyphae with the different CSF samples and detected the deposition by immunofluorescence with specific antibodies. All three CSF samples derived from the patient with cerebral aspergillosis induced a higher complement deposition on the fungal surface than non-inflammatory CSF (Fig. 6). The intensity directly correlated with the complement levels of C1q, C3 and C5 in CSF with the sample taken after 27 days resulting in the strongest opsonization, followed by the sample taken after 32 days. However, all inflammatory CSF samples did not reach the opsonization capacity of serum. The induction of phagocytosis by conidia opsonized in the different CSF samples was also investigated. Opsonization of conidia in the inflammatory CSF samples derived from the patient with cerebral aspergillosis stimulated a higher percentage of granulocytes to ingest the conidia, compared to the non-inflammatory CSF (Fig. 7A). Similarly, the phagocytic index, i.e. the number of conidia per phagocytosing cell was slightly higher with conidia incubated with the inflammatory CSF samples (Fig. 7B). Opsonization of the conidia in serum induced the highest stimulation of phagocytosis. Furthermore, the inflammatory CSF samples induced at least partly a more efficient killing by granulocytes. Opsonization of conidia with the CSF obtained on day 27 significantly enabled the granulocytes to reduce the number of viable conidia (Fig. 8). The granulocyte-induced reduction of conidia opsonized by CSF taken on day 32 was not significant; the
Fig. 7. Phagocytosis of Aspergillus conidia opsonized by non-inflammatory or inflammatory CSF. Human granulocytes were seeded on cover slips. Aspergillus conidia were stained with calcofluor white, either mock-treated in GVB buffer or opsonized for 45 min in serum, non-inflammatory CSF or in CSF samples from a patient with cerebral aspergillosis, obtained 18 days (CA18 d), 27 days (CA27 d) and 32 days (CA32 d) after hospitalization due to neurological symptoms. The conidia were added to the cells, and four parallel samples were analyzed microscopically to determine the percentage of phagocytosing cells with ingested conidia and the phagocytic index (number of ingested conidia per phagocytosing cell). The statistical significance of phagocytosis of conidia incubated with serum or CSF versus phagocytosis of buffertreated conidia was evaluated by Student’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.005.
sample from day 18 and the non-inflammatory CSF were unable to stimulate killing by granulocytes. Again, serum was the most efficient complement source to support the killing activity of granulocytes. 4. Discussion
Fig. 6. Opsonization of fungal hyphae by the CSF from a patient with cerebral aspergillosis. The clinical isolate A. fumigatus 14 was grown on slides, fixated and the hyphae were incubated either with non-inflammatory CSF (A) or with CSF samples from a patient with cerebral aspergillosis, obtained 18 days (B), 27 days (C) and 32 days (D) after hospitalization due to neurological symptoms. Deposition of complement factor C3 was demonstrated by immunofluorescence with a specific 具 C3d-antibody.
The limitation of cerebral immune response by restricted access of peripheral immune effectors and by establishment of an antiinflammatory environment protects the sensitive brain tissue from harmful inflammation [20]. This protection takes a high price, making an effective antifungal response very difficult. The complement system is one of the few locally produced weapons that enable a fast and reliable antifungal reaction. However, the constitutive complement synthesis in the brain is low, and the high mortality rate of patients with
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
Fig. 8. Granulocyte-mediated killing of Aspergillus conidia opsonized by non-inflammatory CSF versus CSF derived from cerebral aspergillosis. Conidia of A. fumigatus 14 were incubated either in human serum, in noninflammatory CSF or in CSF samples from a patient with cerebral aspergillosis, obtained 18 days (CA18 d), 27 days (CA27 d) and 32 days (CA32 d) after hospitalization due to neurological symptoms. Granulocytes were added to the conidia for 5 h at 37 °C, followed by cell lysis in distilled water. Different volumes of the suspension were plated on Sabouraud agar plates and the number of fungal colonies was counted after 18 h. The amount of colonyforming units per sample is presented as mean ± S.D. from two parallel samples. The statistical significance of colony numbers of conidia versus conidia incubated with granulocytes was evaluated for each opsonization condition by Student’s t-test. * P < 0.05; ** P < 0.01; *** P < 0.001.
cerebral aspergillosis indicates that the local complement attack lacks the capacity to eliminate the invading pathogen from CNS. We, therefore, investigated which antifungal immune responses can be induced or supported by the low complement amounts in CSF. Our experiments showed that the constitutive complement levels in CSF enable a clear opsonization of fungal hyphae, albeit much weaker than by serum. The ratio of opsonization intensity between serum and CSF of 1:400 corresponds approximately to the ratio of C3 levels of serum and CSF [19], but the concentration of other factors might also be limiting. Complement deposition on fungal pathogens does not result in direct killing via formation of membrane attack complexes, presumably as a consequence of their thick cell wall. Thus the importance of the complement system for the antifungal defense is supposed to be support of phagocytosis, a critical component of host defense against fungal infection [21]. Phagocytosis is complement-dependent with a clear correlation between deposition of C3 on Aspergillus and its ingestion by human neutrophils [22], since opsonization provides an improved interaction between the fungal surface and complement receptor-bearing phagocytic cells [23–25]. We could show that opsonization of Aspergillus by CSF-derived complement does not achieve the intensity to effectively support phagocytosis of Aspergillus by microglial cells; similarly, infiltrating phagocytes from the periphery (neutrophils, monocytes, macrophages, dendritic cells) are insufficiently stimulated. A slight increase in the uptake rate of CSFopsonized conidia compared to non-opsonized conidia was revealed only for monocytes and microglia. This low support of phagocytosis by cerebral complement might contribute to
1293
the immune failure in cerebral aspergillosis, since there is a correlation between the pathogenicity of Aspergillus and its capacity to evade complement attack and thus limit phagocytosis [10,25]. Following complement receptor-mediated uptake of opsonized particles neutrophils start a powerful microbicidal program with production of oxidative metabolites that destroy the engulfed material. Oxidative mechanisms damage the fungi by producing protein modifications, nucleic acid breaks and lipid peroxidation [26]. Opsonization has a profound effect on induction of oxidative burst in neutrophils [27]. Our experiments showed that, whereas opsonization with serum triggered an oxidative burst in neutrophils, the deposition of complement from CSF was not intense enough to induce this effect. Polymorphonuclear leukocytes not only destroy engulfed pathogens, but can also damage extracellular particles [28–30,8]. For that reason we also quantified directly the killing after contact of neutrophils with opsonized conidia. Whereas granulocytes effectively killed serum-opsonized conidia, fungal viability decreased minimally when granulocytes were incubated with CSF-opsonized conidia. Thus the constitutively present cerebral complement system is no effective weapon against invading Aspergillus, and this lack of efficacy presumably contributes to the high mortality in cerebral aspergillosis. Therefore, enhanced complement synthesis might improve the defense reaction. The CSF of the patient with cerebral aspergillosis contained highly elevated concentrations of C1q (up to 14-fold increase compared to non-inflammatory CSF), C3 and C5 (17-fold increase). This enhancement in complement levels can result either from a leaky blood–brain barrier or from local production in the brain. Immunohistochemical analysis of brain tissue derived from patients with cerebral aspergillosis revealed that upregulation of local production at least contributes to the elevated complement concentrations in CSF (Rambach et al., in preparation). Astrocytes, microglia, neurons and oligodendrocytes were demonstrated in vitro to produce almost all complement proteins in low amounts [12,31] that can be increased by inflammatory stimuli like lipopolysaccharides, viral infection and cytokines. The mechanism of complement upregulation in cerebral aspergillosis needs further investigation and may result from direct interaction with Aspergillus surface molecules and/or from indirect effects of pro-inflammatory cytokines. The potentially protective effects of complement upregulation for the brain are obvious: improved opsonization of fungi with better induction of phagocytosis, stimulation of oxidative burst and a higher killing rate. Similarly, higher production of C1q may support directly the triggering of oxidative burst, chemotaxis and phagocytosis via binding to its receptor on human neutrophils [32]. Furthermore, the anaphylatoxin C3a, which is present in high amounts in the inflammatory CSF samples, is a chemoattractant that recruits inflammatory cells to the site of infection [33]. C3a and C5a can also stimulate microglia and astrocytes via induction of pro-inflammatory cytokines [31] that might again contribute
1294
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295
to the upregulation of complement expression in a feedback loop. In addition, C5a primes neutrophils and enhances the NADPH oxidase activity, a key enzyme of oxidative burst, in response to contact to opsonized zymosan [34]. TCC was described to have stimulatory effects on a number of target cells in the brain and to promote neuronal survival [31]. However, the amount of complement in the CSF of the patient with cerebral aspergillosis never reached the level of the serum. Whereas serum contains 1–1.5 mg/ml C3, the C3 concentration in the inflammatory CSF samples was maximally 38 µg/ml (Fig. 5). Our experiments showed that these complement concentrations enabled better fungal opsonization and improved phagocytosis and (at least partly) granulocyte-mediated killing, but the capacity was far from that of serum. The reasons for this limitation of complement synthesis in the inflamed brain tissue are unknown. One explanation might be that the restriction aims to prevent inflammation-induced tissue damage, since there is considerable evidence that high complement biosynthesis and activation contribute to the pathogenesis of neurodegenerative disorders, leading to neuronal loss and inflammation [12]. Chronic complement activation is associated with neurodegeneration in Alzheimer’s disease and multiple sclerosis. Beside induction of a harmful inflammatory reaction, complement-induced brain damage might also include opsonization of surrounding “self” cells with subsequent phagocytosis and bystander lysis through generation of the membrane attack complex. Neurons are extremely susceptible to lysis even mediated by homologous complement, since they constitutively express only low amounts of inhibitors like CD55, CD59, and MCP [35]. Complement-stimulated neutrophils are involved in tissue damage by an excessive acute inflammatory response, inducing oxidative stress to sensitive bystander cells [36]. An alternative explanation for the limited effect of complement upregulation might be that this occurs too late in the course of infection and is limited by the Aspergillus-induced tissue damage, leaving only few cells unaffected and thus capable of synthesizing complement. If this hypothesis is confirmed by further experiments, the specific manipulation of the level of biosynthesis and/or activation of the local complement system (e.g. using recombinant cytokines, or increase of expression by gene therapy) might be of potential therapeutic value and represent a supportive immune-based therapy to improve the treatment with antifungal drugs.
References [1] [2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11] [12]
[13]
[14]
[15]
[16]
[17]
Acknowledgements We gratefully acknowledge the generous gift of the monoclonal anti-C3 antibody BB5 from W. Prodinger (Department of Hygiene, Microbiology and Social Medicine, Innsbruck, Austria). This study was supported by LudwigBoltzmann-Society, FWF (project P15375), Österreichische Nationalbank (project 9374), BMSG and the State of Tyrol.
[18]
[19]
D.W. Denning, Invasive aspergillosis, Clin. Infect. Dis. 26 (1998) 781–805. R.C. Young, J.E. Bennett, C.L. Vogel, P.P. Carbone, V.T. DeVita, Aspergillosis: the spectrum of the disease in 98 patients, Medicine 49 (1970) 147–173. M.E. Hagensee, J.E. Bauwens, K. Bent, R.A. Bowden, Brain abscess following marrow transplantation: experience at the Fred Hutchinson Cancer Research Center, 1984–1992, Clin. Infect. Dis. 19 (1994) 402–408. N.P. Hailer, F.L. Heppner, D. Haas, R. Nitsch, Astrocytic factors deactivate antigen presenting cells that invade the central nervous system, Brain Pathol. 8 (1998) 459–474. P.J.H. Jongen, W.H. Doersburg, J.L.M. Ibrahim-Stappers, W.A.J.G. Lemmens, O.R. Hommes, K.J.B. Lamers, Cerebrospinal fluid C3 and C4 indexes in immunological disorders of the central nervous system, Acta Neurol. Scand. 101 (2000) 116–121. C. Speth, R. Würzner, H. Stoiber, M.P. Dierich, The complement system: pathophysiology and clinical relevance, Wien. Klin. Wochenschr. 111 (1999) 378–391. T.R. Kozel, M.A. Wilson, T.P. Farrell, S.M. Levitz, Activation of C3 and binding to Aspergillus fumigatus conidia and hyphae, Infect. Immun. 57 (1989) 3412–3417. J.E. Sturtevant, J.P. Latgé, Interactions between conidia of Aspergillus fumigatus and human complement component C3, Infect. Immun. 60 (1992) 1913–1918. R.F. Hector, E. Yee, M.S. Collins, Use of DBA/2N mice in models of systemic candidiasis and pulmonary and systemic aspergillosis, Infect. Immun. 58 (1990) 1476–1478. S. Henwick, S.V. Hetherington, C.C. Patrick, Complement binding to Aspergillus conidia correlates with pathogenicity, J. Lab. Clin. Med. 122 (1993) 27–35. B.P. Morgan, P. Gasque, Expression of complement in the brain: role in health and disease, Immunol. Today 17 (1996) 461–466. P. Gasque, Y.D. Dean, E.P. McGreal, J. VanBeek, B.P. Morgan, Complement components of the innate immune system in health and disease in the CNS, Immunopharmacology 49 (2000) 171–186. C. Speth, G. Rambach, C. Lass-Flörl, R. Würzner, P. Gasque, I. Mohsenipour, M.P. Dierich, Culture supernatants of patient-derived Aspergillus isolates have toxic and lytic activity towards neurons and glial cells, FEMS Immunol. Med. Microbiol. 29 (2000) 303–313. R. Würzner, M. Schulze, L. Happe, A. Franzke, F.A. Bieber, M. Oppermann, O. Götze, Inhibition of terminal complement complex formation and cell lysis by monoclonal antibodies, Complement Inflamm. 8 (1991) 328–340. M.A. Hess, M.G. Schwendinger, E.L. Eskelinen, K. Pfaller, M. Pavelka, M.P. Dierich, W.M. Prodinger, Tracing uptake of C3dgconjugated antigen into B cells via complement receptor type 2 (CR2, CD21), Blood 95 (2000) 2617–2623. M. Barcova, C. Speth, L. Kacani, F. Überall, H. Stoiber, M.P. Dierich, Involvement of adenylate cyclase and p70S6-kinase activation in IL-10 up-regulation in human monocytes by gp41 envelope protein of human immunodeficiency virus type 1, Eur. J. Physiol. 437 (1999) 538–546. D.A. Bass, J.W. Parce, L.R. Dechatelet, P. Szejda, M.C. Seeds, M. Thomas, Flow cytometric studies of oxidative product formation by neutrophils: a graded response to membrane stimulation, J. Immunol. 130 (1983) 1910–1917. L.E. Leigh, B. Ghebrehiwet, T.P. Perera, I.N. Bird, P. Strong, U. Kishore, K.B. Reid, P. Eggleton, C1q-mediated chemotaxis by human neutrophils: involvement of gC1qRp and G-protein signalling mechanisms, Biochem. J. 330 (1998) 247–254. W.M. Prodinger, R. Würzner, H. Stoiber, M.P. Dierich, in: W.E. Paul (Ed.), Fundamental Immunology, Lippincott-Raven, Philadelphia, 2003, pp. 1077–1103.
G. Rambach et al. / Microbes and Infection 7 (2005) 1285–1295 [20] Z. Fabry, C.S. Raine, M.N. Hart, Nervous tissue as an immune compartment: the dialect of the immune response in the CNS, Immunol. Today 15 (1994) 218–224. [21] A.R. Waldorf, Pulmonary defense mechanisms against opportunistic fungal pathogens, Immunol. Ser. 47 (1989) 243–270. [22] J.E. Sturtevant, J.P. Latgé, Participation of complement in the phagocytosis of the conidia of Aspergillus fumigatus by human polymorphonuclear cells, J. Infect. Dis. 166 (1992) 580–586. [23] R.A. Calderone, L. Linehan, The role of complement in host resistance to systemic fungal infection, Immunol. Ser. 47 (1989) 225–242. [24] J.W. Murphy, Immunity to fungi, Curr. Opin. Immunol. 2 (1990) 360–367. [25] H.F. Tsai, Y.C. Chang, R.G. Washburn, M.H. Wheeler, K.J. KwonChung, The developmentally regulated alb1 gene of Aspergillus fumigatus: its role in modulation of conidial morphology and virulence, J. Bacteriol. 180 (1998) 3031–3038. [26] M.K. Mansour, S.M. Levitz, Interaction of fungi with phagocytes, Curr. Opin. Microbiol. 5 (2002) 359–365. [27] A.A. Vaporciyan, P.A. Ward, Enhanced generation of O2– by human neutrophils via complement iC3b/Mac-1 interaction, Biol. Signals 2 (1993) 126–135. [28] R.D. Diamond, R. Krzesicki, W. Jao, Damage to pseudohyphal forms of Candida albicans by neutrophils in the absence of serum in vitro, J. Clin. Invest. 61 (1978) 349–359.
1295
[29] R.D. Diamond, R. Krzesicki, B. Epstein, W. Jao, Damage to hyphal forms of fungi by human leukocytes in vitro. A possible host defense mechanism in aspergillosis and mucormycosis, Am. J. Pathol. 91 (1978) 313–328. [30] A.R. Waldorf, R.D. Diamond, Neutrophil chemotactic responses induced by fresh and swollen Rhizopus oryzae spores and Aspergillus fumigatus conidia, Infect. Immun. 48 (1985) 458–463. [31] C. Speth, M.P. Dierich, P. Gasque, Neuroinvasion by pathogens: a key role of the complement system, Mol. Immunol. 38 (2002) 669–679. [32] L.E. Leigh, B. Ghebrehiwet, T.P. Perera, I.N. Bird, P. Strong, U. Kishore, K.B. Reid, P. Eggleton, C1q-mediated chemotaxis by human neutrophils: involvement of gC1qRp and G-protein signalling mechanisms, Biochem. J. 330 (1998) 247–254. [33] S.M. Levitz, Overview of host defenses in fungal infections, Clin. Infect. Dis. 14 (Suppl 1) (1992) 37–42. [34] J.G. Bender, L.C. McPhail, D.E. Van Epps, Exposure of human neutrophils to chemotactic factors potentiates activation of the respiratory burst enzyme, J. Immunol. 130 (1983) 2316–2323. [35] S.K. Singhrao, J.W. Neal, N.K. Rushmere, B.P. Morgan, P. Gasque, Spontaneous classical pathway activation and deficiency of membrane regulators render human neurons susceptible to complement lysis, Am. J. Pathol. 157 (2000) 905–918. [36] H. Jaeschke, C.W. Smith, Mechanisms of neutrophil-induced parenchymal cell injury, J. Leukoc. Biol. 61 (1997) 647–653.