Stimulation of isolated rat Kupffer cells by viable Leptospira interrogans, Treponema pallidum and Borrelia garinii elicited cellular responses resulting in the release of different tumor necrosis factor α (TNF-α) levels, depending on the spirochetes. L. interrogans induced TNF-α levels higher than those achieved with B. garinii and T. pallidum (in this order), but lower than the levels achieved with lipopolysaccharide (LPS). In contrast to L. interrogans, pretreatment of borreliae and treponemes with polymyxin B did not substantially diminish the ability of B. garinii and T. pallidum to stimulate Kupffer cells. Purified T. pallidum lipoproteins TpN47, TmpA, TpN15–TpN17, and B. garinii OspA induced TNF-α responses comparable to that achieved by LPS. This response was almost insensitive to the action of polymyxin B.
Spirochetes are the etiological agents of human infections causing diseases such as syphilis, Lyme borreliosis, relapsing fevers and leptospirosis. Despite substantial biological differences among the genera Treponema, Borrelia and Leptospira, both clinical and pathogenetic similarities have been recognized among spirochetal diseases. From the pathogenetic point of view, all spirochetoses share a spirochetemic phase during an early stage of infection. During this phase, leptospires can cause hepatitis in humans. This will result in microscopic alterations in the liver, including enlargement of Kupffer cells . In Lyme disease, caused by Borrelia burgdorferi sensu lato (s.l.), liver function test abnormalities are common but mild, most often not associated with symptoms. Systemic features of secondary syphilis include hepatitis, particularly in young adults .
It is not yet clear how spirochetes cause inflammation in tissues. In vitro and in vivo studies have shown that B. burgdorferi s.l. is rapidly phagocytosed and killed by macrophages in the absence of serum factors [3,4]. B. burgdorferi s.l. also activates several immune and non-immune cells with the release of various monocytic cell cytokines, among which the proinflammatory ones interleukin (IL)-1β and tumor necrosis factor α (TNF-α) [5,6].
Studies with Treponema pallidum have shown that the macrophages take up and kill T. pallidum and are therefore likely responsible for the clearance of the spirochete from the infected tissues. In vivo studies demonstrated that cells infiltrating primary and secondary syphilis lesions contain mRNA encoding IL-2, interferon-γ, and IL-12 . In vitro studies showed that T. pallidum lipoprotein TpN47 and B. burgdorferi lipoprotein OspA induce the synthesis of TNF-α. The stimulatory effect was dependent on lipidation and could be reproduced using synthetic acylated hexapeptides . The use of in vitro and animal models has shown that leptospires are phagocytosed by Kupffer cells [10,11], and lipopolysaccharide (LPS) is likely to play a role in inflammation in diseases caused by these spirochetes .
There are no comparative studies about the ability of T. pallidum, B. burgdorferi s.l., and Leptospira interrogans to activate the cells of the hepatic reticuloendothelial system and to induce the release of cytokines. Therefore, in this study, we analyzed the TNF-α release by rat Kupffer cells probed with treponemes, borreliae and leptospires.
Materials and methods
Bacterial strains and culture conditions
We used the following spirochetal strains: Borrelia garinii PBi (a gift of Dr. Bettina Wilske, Max von Pettenkofer-Institut für Hygiene und Medizinische Mikrobiologie, Munich, Germany), the avian strain Borrelia anserina ES (a gift of Dr. R.C. Johnson, Minneapolis, MN, USA), L. interrogans serovar icterohaemorrhagiae (a gift of Dr. M. Fabbi, Istituto Zooprofilattico Sperimentale, Pavia, Italy), and T. pallidum, Nichols strain (purchased from Statens Serum Institute, Copenhagen, Denmark). B. garinii PBi and B. anserina ES were cultured in Barbour–Stoenner–Kelly II medium at 34°C, as previously reported [13,14]. Leptospires were grown in liquid Ellighausen–McCullough–Johnson–Harris medium at 30°C under aerobic conditions, to a density of about 108 bacteria ml−1 and counted in a Petroff–Hausser counting chamber. T. pallidum was maintained by testicular passage in adult male New Zealand White rabbits. Infected rabbits were individually housed, maintained at 16–18°C and given antibiotic-free food and water. Each treponemal suspension was prepared from rabbit testicles infected with 2×107 treponemes, 10–14 days after inoculation. The treponemes were isolated from infected testicular tissue using RPMI 1640 (Gibco Laboratories, Grand Island, NY, USA), supplemented with 10% fetal bovine serum, under strict anaerobic conditions. After centrifugation at 1000×g for 10 min to remove cellular debris, the number of treponemes in the medium was determined by dark-field microscopy.
Separation of Kupffer cells
Kupffer cells were harvested following the procedure of Smedsrød and Pertoft  with some modifications. Briefly, male Sprague–Dawley rats (220–300 g body weight) were used as liver donors. The animals were anesthetized with pentobarbital sodium (50 mg kg−1, given intraperitoneally), and the liver was perfused in a non-recirculating system through the portal vein with 500 ml of calcium- and magnesium-free Hanks’ balanced salt solution, which had been pre-warmed to 37°C and buffered at pH 7.3 with 10 mM HEPES (Sigma Chemicals, St. Louis, MO, USA). The effluent from the liver was collected from the inferior vena cava immediately above the suprahepatic veins. The liver was then perfused with complete Hanks’ balanced salt solution containing 0.05% collagenase (type IV; Sigma), pre-warmed to 42°C. The perfusate flow was established at 8 ml min−1 for 15 min. The liver was then excised and placed in a Petri dish, the liver capsule peeled off and the cells were gently dispersed in calcium- and magnesium-free Hanks’ balanced salt solution. The cells were then centrifuged at 50×g at 4°C for 2 min, in a Beckman J6B centrifuge (Beckman Instruments, Palo Alto, CA, USA). The non-parenchymal cell-enriched supernatant was centrifuged at 800×g for 10 min, the pellet resuspended in 40 ml of phosphate-buffered saline (PBS), and portions of 10 ml were layered on top of a preformed two-step Percoll gradient (the bottom cushion with a density of 1066 g ml−1 and an osmolality of 310 mOsm; the overlying cushion with a density of 1.037 g ml−1 and an osmolality of 300 mOsm), and centrifuged at 400×g for 15 min at 4°C. Purified non-parenchymal cells enriched in Kupffer cells extended throughout the lower Percoll cushion. The pellets consisted of erythrocytes, liver non-parenchymal cells, and other small white cells. The Kupffer cell-enriched fraction was diluted in PBS and centrifuged at 800×g for 10 min. The resulting pellet was resuspended in culture medium (RPMI 1640 with 10% fetal calf serum) at a concentration of 2.0×106 cells ml−1. A 0.5-ml portion of cell suspension was added to 8-well culture plate (Lab-Tek, Nalge Nunc International, Naperville, IL, USA). Macrophages were selected by allowing them to adhere for 2 h at 37°C, in an atmosphere with 5% CO2. After non-adherent cells were removed by gentle washing, adherent cells were incubated with RPMI 1640 for 24 h before performing uptake or cytokine assays. More than 95% of adherent cells were esterase-positive.
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS–PAGE) was performed using 12% acrylamide gel, as described elsewhere .
Isolation of spirochetal proteins
B. garinii PBi OspA and p41, and T. pallidum TpN47, TmpA, TpN17 and TpN15 kDa, B. anserina ES p24 proteins were purified by SDS–PAGE. After separation of spirochete peptides by SDS–PAGE, the gels were stained with CuCl2, followed by rapid destaining in distilled water. The bands corresponding to the above proteins were excised from the gel and transferred into a dialysis tube and electroeluted at 200 V for 2 h. The purity of electroeluted proteins was assessed by SDS–PAGE and staining of the gels with Coomassie brilliant blue or silver nitrate solution.
Kupffer cells for TNF-α assays
Kupffer cells, purified as above reported, were added (5×105) to each well of an 8-well culture plate (Lab-Tek). After 2 h at 37°C, non-adherent cells were removed by washing three to five times with RPMI 1640 supplemented with 10% fetal calf serum, 2 mM l-glutamine. Cells were stimulated for 6 h with either LPS (Sigma), viable L. interrogans, B. garinii PBi, T. pallidum, or with purified spirochetal proteins. The supernatants were then harvested for TNF-α assays (see below).
TNF-α released into supernatants of Kupffer cells was measured by the L929 cytotoxicity bioassay. Briefly, monolayer cultures of L929 fibroblasts were grown in 75-cm2 flasks in complete Dulbecco's modified Eagle's medium. The L929 cells were trypsinized, plated in 96-microwell flat-bottomed plates and cultured overnight at 37°C in 5% CO2. All TNF-α samples (diluted 1:10 in medium) were tested in triplicate. Each test well contained 200 µl of medium (RPMI 1640 without phenol red and with 5% fetal bovine serum). After 24 h, the wells were stained with 0.5% crystal violet in 20% (v/v) methanol for 30 min. Following extensive washing with PBS, the incorporated dye was eluted by the addition of 50 µl of 0.1 M sodium citrate in 50% (v/v) ethanol (pH 4.2), and optical densities were read at 540 nm. TNF-α activity was determined by comparison of experimental optical densities to those obtained from a standard curve of recombinant murine TNF-α (Sigma). When required, preincubation of material with polymyxin B (10 µg ml−1) (Sigma) was used to abrogate the effect of LPS.
Results and discussion
TNF-α release by Kupffer cells stimulated by viable spirochetes
Stimulation of Kupffer cells by viable spirochetes elicited cellular responses resulting in the production of TNF-α levels higher than those observed in control cultures. L. interrogans induced TNF-α levels higher than those achieved with B. garinii PBi and T. pallidum, but lower than the levels achieved with LPS (Fig. 1). In contrast to L. interrogans, pretreatment of B. garinii PBi and T. pallidum with 10 µg ml−1 of polymyxin B, which binds and inactivates LPS , did not substantially diminish the ability of spirochetes to stimulate Kupffer cells.
TNF-α release by Kupffer cells stimulated by purified spirochetal proteins
To investigate the ability of treponemal and borrelial lipoproteins to activate Kupffer cells, two abundant lipoproteins, T. pallidum TpN47 and B. garinii OspA, the less abundant T. pallidum TpN15 and TpN17, TmpA, B. anserina ES p24, and the non-acylated protein p41 of B. garinii were purified (Fig. 2). In cell activation studies, spirochetal lipoproteins and LPS induced comparable TNF-α responses. A clearly lower TNF-α response was induced by non-acylated B. garinii p41. Pretreatment of LPS with polymyxin B abrogated TNF-α induction in Kupffer cells, whereas it had no effect on TNF-α induction by spirochetal lipoproteins (Fig. 2).
The innate immune response provides a rapid mechanism for recognizing and responding to microbial pathogens . Kupffer cells, the resident macrophages of the liver, are primarily responsible for this activity during bacteremic infections.
Leptospirosis, Lyme disease and syphilis are multisystem diseases where host injury is presumed to result from the inflammatory response to intrinsic spirochetal components [18,19]. While it is well established that borreliae and T. pallidum lack the potent proinflammatory molecule LPS [20,21], this latter is present in leptospires . On the other hand, borrelial and treponemal lipoproteins have been found to be potent stimulators of inflammation . LPS binds to macrophages mainly through the CD14 receptors resulting in activation of the cells and release of inflammatory mediators . Borrelial and treponemal lipoproteins have also been shown to activate peripheral blood mononuclear cells through a CD14-dependent pathway .
In this work we analyzed comparatively the proinflammatory activity of viable spirochetes and spirochetal components, by studying their ability to induce TNF-α release by Kupffer cells, which are primarily responsible for the clearance of LPS from the blood .
Leptospires, as expected, had the highest stimulatory effect on Kupffer cells to release TNF-α, while intact T. pallidum cells were less stimulatory than intact borreliae . Leptospiral stimulation of TNF-α release was reduced by 50% by polymyxin B, suggesting that it was equally due to bacterial LPS . In contrast, polymyxin B reduced only by about 10% and 15%, respectively, the TNF-α production by Kupffer cells stimulated by viable borreliae and treponemes. This observation supports the possibility that intrinsic spirochetal components other than LPS are responsible for Kupffer cell activation by intact borreliae and treponemes. In effect, the results obtained in the present study showed that several purified Borrelia and T. pallidum lipoproteins, but not spirochetal non-acylated proteins, act as potent TNF-α inducers of Kupffer cells, showing a stimulatory activity equivalent to LPS. This activity was insensitive to the action of polymyxin B. The various T. pallidum purified lipoproteins were similar in potency to each other and to borrelial preparations, confirming previous data obtained with other macrophages . Finally, it is interesting to note that lipoprotein p24 of B. anserina ES  showed a TNF-α-inducing activity similar to that of B. garinii lipoproteins, suggesting that this activity is a property shared by lipoproteins of bacteria of the genus Borrelia as well as by lipidated proteins of T. pallidum.
This study was supported by MURST grant ‘cofinanziamento 98’.