Induction of resistance to Toxoplasma gondii in human macrophages by soluble lymphocyte products

Autoradiography and light microscopy were used to study the effects of lymphocyte culture supernatants, prepared under a variety of conditions, on the course of intracellular Toxoplasma gondii infection in human monocyte-derived macrophages in vitro. Supernatants prepared by incubating lymphocytes of dye test- (DT) positive subjects with T. gondii lysate antigen (TLA), lymphocytes of DT/negative subjects with streptokinase-streptodornase (SK-SD), or both populations of lymphocytes with concanavalin A (Con A) were capable of activating macrophages to inhibit or kill intracellular T. gondii. Supernatants prepared with the homologous antigen (TLA) to the target organism appeared more active in conferring resistance to infection with T. gondii on macrophages than those prepared with a heterologous antigen (SK-SD) or mitogen (Con A). The number of lymphocytes was critical in preparing active supernatants. These results suggest that soluble lymphocyte mediators can activate human macrophages in vitro to inhibit or kill T. gondii.     

Effects of Helicobacter pylori vacuolating cytotoxin on primary cultures of human gastric epithelial cells

BACKGROUND: Many Helicobacter pylori strains produce a cytotoxin that induces cytoplasmic vacuolation in various types of eukaryotic cells. In contrast with the marked cell vacuolation that occurs in vitro in response to this cytotoxin, comparatively little epithelial vacuolation has been observed in the gastric mucosa of H pylori infected persons. AIMS: Experiments were performed to determine the susceptibility of human gastric epithelial cells in vitro to H pylori vacuolating cytotoxin activity. METHODS: Human gastric epithelial cells, harvested from upper gastrointestinal endoscopic biopsy specimens, were incubated overnight with broth culture supernatants from either a wild type cytotoxin producing (tox+) H pylori strain or an isogenic mutant strain that lacks cytotoxin activity. RESULTS: Prominent cytoplasmic vacuolation occurred in response to tox+ supernatant, but not supernatant from the isogenic mutant strain. Primary human gastric epithelial cells were significantly more sensitive to H pylori vacuolating cytotoxin activity than were either HeLa or AGS cells. Exposure of human gastric epithelial cells to high concentrations of tox+ supernatant for 48 hours caused lethal cell injury. CONCLUSIONS: These studies indicate that primary human gastric epithelial cells are highly sensitive to H pylori vacuolating cytotoxin activity.     

Studies on Pasteurella multocida. III. In vitro assay for cell-mediated immunity

Peripheral blood lymphocytes (PBL) from turkeys immunized against fowl cholera with a bacterin or a live avirulent vaccine (strain CS-148) were cultured in vitro with various antigenic preparations from Pasteurella multocida (strain P-1059). The degree of lymphocyte stimulation (blastogenesis) was quantitated by measurement of the uptake of (3H) thymidine. Higher stimulation indices were obtained with immune lymphocytes rather than nonimmune lymphocytes. Stimulation was specific since PBL from turkeys immunized against P. multocida failed to react with Escherichia coli or Mycoplasma synoviae antigens. These differences were statistically significant as analyzed with the student's t-test. The lymphocyte transformation assay was emphasized as a convenient and useful in vitro indicator of cell-mediated immunity that should help define the role of cell-mediated immunity in P. multocida infections of turkeys.     

Human transfer factor: effects on lymphocyte transformation

Transfer factor preparations from 57 different donors have been compared for effects on mitogen- and antigen-induced lymphocyte transformation. Nine of the preparations were mitogenic when added to cultured lymphocytes although the magnitude of this activity was relatively low. The majority of the preparations (48/57) did not affect PHA-induced lymphocyte transformation although augmentation (6 of 57) and suppression (3 of 57) was observed with some. In addition we observed that most of the preparations tested suppressed ConA stimulation and augmented the PWM response. When selected preparations were evaluated on antigen-responsive cells, there was a correlation between the magnitude of antigen responsiveness and the magnitude of TF augmentation of antigen-induced lymphocyte transformation (p less than 0.005). Cultures that were not responsive to antigen (KLH-negative or BUdR-treated) could not be stimulated by TF from immune donors and antigen. These data suggest that TF preparations contain either stimulatory or inhibitory components and that TF is not capable of activating naive lymphocytes to undergo transformation in response to antigen.     

B-lymphocyte proliferation during bovine leukemia virus-induced persistent lymphocytosis is enhanced by T-lymphocyte-derived interleukin-2

Bovine leukemia virus (BLV)-induced persistent lymphocytosis is characterized by a polyclonal expansion of CD5+ B lymphocytes. To examine the role of the cytokine microenvironment in this virus-induced B-lymphocyte expansion, the expression of interleukin-2 (IL-2), IL-4, IL-10, and gamma interferon (IFN-gamma) mRNA, was measured in stimulated peripheral blood mononuclear cells from persistently lymphocytotic BLV-infected cows, nonlymphocytotic BLV-infected cows, and uninfected cows. IL-2 and IL-10 mRNA expression and IL-2 functional activity were significantly increased when peripheral blood mononuclear cells from persistently lymphocytotic cows were stimulated with concanavalin A (ConA). Additionally, during persistent lymphocytosis, peak IL-2 and IL-10 mRNA expression was delayed, and elevated expression was prolonged. To determine the potential biologic importance of increased IL-2 and IL-10 expression, the response of isolated B lymphocytes from persistently lymphocytotic cows to human recombinant cytokines and to cytokine-containing supernatants from isolated T lymphocytes was examined. While recombinant human IL-10 (rhIL-10) did not consistently induce detectable changes, rhIL-2 increased viral protein (p24) and IL-2 receptor expression in isolated B lymphocytes from persistently lymphocytotic cows. Additionally, rhIL-2 and supernatant from ConA-stimulated T lymphocytes enhanced B-lymphocyte proliferation. The stimulatory activity of the T-lymphocyte supernatant could be completely inhibited with a polyclonal anti-rhIL-2 antibody. Finally, polyclonal anti-rhIL-2 antibody, as well as anti-BLV antibody, inhibited spontaneous proliferation of peripheral blood mononuclear cells from persistently lymphocytotic cows, demonstrating that the spontaneous lymphoproliferation characteristic of BLV-induced persistent lymphocytosis is IL-2 dependent and antigen dependent. Collectively, these findings strongly suggest that increased T-lymphocyte expression of IL-2 in BLV-infected cows contributes to development and/or maintenance of persistent B lymphocytosis.     

Biochemical and serological investigations of flagellae of Campylobacter fetus (author's transl)

Flagellae of Campylobacter fetus group O, types 1, 2 and 7 were prepared. First they were separated from cell bodies using an ultramix. The suspension was then centrifuged for 20 mins. at 10000 rpm and the supernatant frozen at -40 degrees C. This is a simple method for the enrichment of preparations of flagellae, as they become tangled up and accumulate in the inferior third part of the frozen liquid. The physicochemical basis of this phenomenon was discussed. After thawing and spinning for 20 mins. at 5000 rpm, the sediment was suspended in 0.9% of NaC1. The purity of the preparation was checked by electron microscopy. Antibodies to this antigen showed no cross-reaction with O-Antigen, when tested by tube agglutination. The amino acid composition of flagellae from different O-antigen serotypes was different (see Tab. 1). Cysteine could not be detected and proline only in traces. After breakdown with urea followed by gel filtration on Sephadex G 200, breakdown products of diminishing molecular size were obtained (see Fig. 2). Discelectrophoresis after ultrasonic gave 8 zones (see Fig. 3). Irrespective of serotype, thin-layer chromatography of trypsin-hydrolysed flagellae always showed 9 ninhydrin-positive spots (see Fig. 1). Only breakdown products of ultrasonic reacted with antibody. After absorption of flagellar antibody with heterologous antigen, agglutination only occurred with homologous antigen (tab. 2-5). This showed that there were different flagellar antigens. Further experiments using immunoprecipitation demonstrated two common antigenic components, a and c, and a partially common antigenic factor bb (Fig. 4), and was the basis for a classification by three groups. The three antigenic components could be separated by gel electrophoresis and detected by immunoprecipitation (Figs. 5,6).     

Cell-mediated immune response to products of Actinomyces viscosus cultures

Hartley strain guinea pigs were sensitized with 0.5 ml of concentrated cell-free Actinomyces viscosus culture supernatant fluids mixed with Freund complete adjuvant. Fourteen to 16 days later the animals were challenged by intradermal injection with 0.1 ml of the culture supernatant, and the reactions were observed at 4, 8, 16, 24, and 48 h. Peritoneal exudate cells from sensitized animals were used for determination of migration inhibition factor, and guinea pig peripheral blood served as a source of cells for determining the induction of mitogenesis by antigenic material. Skin responses were consistently positive to challenge with the test material, whereas reactions to noninoculated culture medium were negative. Sensitized cells, challenged with antigen, resulted in 60% or greater inhibition of migration of indicator cells in migration inhibition factor experiments. Tests for mitogenesis showed a greater than fourfold increase in isotope uptake when sensitized cells were challenged with test material. The data are consistent with the suggestion that A. viscosus culture supernatants contain substances that induce cell-mediated immune responses in guinea pigs.     

Quantitative microanalysis of cell walls of Saprolegnia diclina Humphrey and Tremella mesenterica Fries

Cell walls of the fungi Saprolegnia declina Humphrey and Tremella mesenterica Fries were analyzed quantitatively. Particular attention was paid to the hydrolysis and analysis of neutral sugars, amino sugars and amino acids. These components, together with total lipids, total uronic acids and the ashed residue, accounted for more than 90% by weight of the original dry cell wall preparation. There were substantial losses of amino acids during hydrolysis; however, analytical recovery approached 100% when total protein was calculated from the total nitrogen analysis. The analytical procedures were reproducible (+/- 3% for amino acids and amino sugars, and +/- 5-10% for other components) when applied to individual cell wall preparations. However, even under carefully standardized conditions, different cell wall preparations from the same species showed variable composition. Glucose was the predominant neutral sugar in the cell wall polymers of both species. The amino acid compositions were remarkable in that neither species contained detectable levels of cyst(e)ine. Hydroxyproline was detected in both species. The report from Tremella mesenterica is the first for this amino acid from the cell wall of a Basidiomycete.     

Lymphocytes produce IL-1beta in response to Fcgamma receptor cross-linking: effects on parenchymal cell IL-8 release

Neutrophils mediate tissue injury in response to immune complexes, although the factors that induce their recruitment are incompletely understood. We have reported that lymphocytes may be important regulators of monocyte and macrophage IL-8 release in the presence of immobilized IgG. Since tissue parenchymal cells are important local producers of IL-8 but are not directly stimulated by FcgammaR cross-linking, we hypothesized that lymphocytes may also regulate parenchymal IL-8 release. Supernatants from lymphocytes incubated on immobilized IgG induced primary human fibroblasts and human mesangial cells to produce IL-8 (17 +/- 3.5 and 44 +/- 8 ng/ml, respectively). Fibroblast and mesangial cell IL-8 mRNA levels were similarly increased by the conditioned lymphocyte supernatant. Immobilized anti-human FcgammaRIII, but not FcgammaRI or FcgammaRII Abs, could stimulate this IL-8-inducing activity in lymphocytes, suggesting that FcgammaRIII-bearing lymphocytes were responsible. Supernatants from lymphocytes incubated on immobilized IgG contained 2.2 +/- 0.8 ng/ml of IL-1beta, while enriched monocyte preparations from the same donors incubated on immobilized IgG released only 0.1 +/- 0.04 ng/ml of IL-1beta (p = 0.05). Consistent with the identification of IL-1beta as the lymphocyte factor, fibroblast or mesangial cell IL-8 release induced by the IgG-stimulated lymphocyte supernatants was inhibited by 1) the combination of IL-1R antagonist and soluble type II IL-1R, 2) an IL-1-converting enzyme inhibitor, or 3) anti-IL-1beta but not preimmune Abs. These data suggest that targeted deposits of IgG can stimulate FcgammaRIII-bearing lymphocytes to produce IL-1beta, which induces parenchymal cell IL-8 release.     

In vitro activation of mouse macrophages by rat lymphocyte mediators

Macrophage-activating factors (MAF)3 were released by presensitized rat lymphocytes stimulated in vitro with the appropriate antigens. Different supernatants of presensitized rat lymphocytes specifically stimulated in vitro with several different mouse, dog, and rat tumor or normal cells were capable of rendering normal rat and mouse macrophages nonspecifically cytotoxic in vitro to their respective syngeneic tumor cells. The release of active mediators by rat lymphocytes sensitized in vivo was dependent upon immunologically specific recognition of an antigen in vitro. When rat lymphocytes were incubated in vitro with antigens unrelated to the in vivo sensitizing antigens, no release of MAF occurred. Once rat MAF was released, it activated both syngeneic (rat) and xenogeneic (mouse) macrophages to kill tumor cells in vitro. These activated marcophages destroyed all syngeneic tumor targets. Such cytotoxicity was obtained even when the cells used to elicit release of MAF were totally unrelated to the target tumor cells. The data thus demonstrated that MAF can cross strain and even species specificities and can activate macrophages to kill tumors in a nonspecific manner. The cytotoxicity mediated by in vitro activated mouse macrophages decreased with time once the macrophages were removed from MAF; and by 7 days postactivation, the macrophages were not cytotoxic. However, when incubated again with MAF, significant reactivation was observed. This suggested that activation of macrophages in vivo may be a continuous process of lymphocyte-macrophage interaction.     

Immune modulation of human B lymphocytes by gene transfer with recombinant Epstein-Barr virus amplicons

We described previously a novel mode of gene transfer by infection of human B lymphocytes with recombinant Epstein-Barr virus (EBV) amplicons. This system was explored for its potential use in expressing various recombinant genes, including the cytokine IL-4, the HIV envelope glycoprotein (gp120) and a suicide and gag gene. Recombinant genes were present as multiple copy episomes and stable, high level recombinant gene expression could be detected by antigenic and functional assays. Amplicon-infected B cells secreted high levels of recombinant cytokine and efficiently presented recombinant antigens through classes I and II MHC-restricted antigen processing pathways. Thus, recombinant EBV amplicons can be used to express components of the immune system or heterologous genes for immune recognition in human B cells. Combining gene transfer with EBV infection may provide unique advantages for in vitro and in vivo gene transfer.     

Chinese and American college students'' body-image: perceived body shape and body affect

We established an in vitro model of the phagocytosis of Mycobacterium tuberculosis by human peripheral blood monocytes to evaluate the subsequent inhibition of intracellular replication of the organism. Highly purified T cells (94% CD3(+)/CD16(-)) or natural killer (NK) cells (96% CD16(+)/CD3(-)) isolated by Percoll discontinuous density gradient of peripheral blood mononuclear cells were incubated with M. tuberculosis-infected monocyte monolayers. Monocytes were lysed immediately and at 4, 7, and 10 d after infection for quantification of intracellular replication, which was assessed by quantitative plating techniques as colony-forming units (CFU). Whereas control monocytes permitted intracellular replication, T cells activated monocytes to kill 77% (p < 0.01) of intracellular M. tuberculosis compared with control monocytes by 10 d after infection. NK cells activated monocytes to kill 84% (p < 0.01) of M. tuberculosis in comparison with control monocytes. Lymphokine (IL-2)-activated-killer (LAK) cells were capable of activating monocytes to kill 97% (p < 0.01) of the intracellular organisms compared with control monocytes. In purified protein derivative (PPD)-positive donors, PPD-specific-CD4(+) lymphocytes stimulated monocytes to kill intracellular M. tuberculosis in a Class II major histocompatibility complex-restricted manner. In contrast, in PPD-negative donors, CD4(-) lymphocytes activated monocytes in a genetically unrestricted manner. Both T cell supernatant and NK cell supernatant generated from cocultivation with M. tuberculosis-infected monocytes also activated monocytes to augment mycobactericidal function. In conclusion, T cells, NK cells, LAK cells, and their supernatants activated mycobactericidal function of monocytes, although these pathways of activation differed in terms of antigenic specificity and genetic restriction.     

Allogeneic and semiallogeneic immunizations with a strong transplantation antigen (Ag-B)

A comparison has been made between the alloantibody response evoked by graded doses of cells being allogeneic and semiallogeneic with the host. The responses were measured by hemagglutination and complement-dependent lymphocytotoxicity. In the strain combination of rats used (DA leads to HO), a marked difference in the optimal dose was found after i.v. injection of lymphocytes. Although semiallogeneic cells gave the better antigenic stimulus at lower cell doses (0.3 and 3 X 10(6) cells), allogeneic cells were better at higher cell doses (30 and 150 X 10(6) cells). Immunization experiments with allogeneic and semiallogeneic erythrocytes or mitomycin- or heat-treated lymphocytes indicated that the antigenicity of the cells was not directly related to the cell surface concentration of antigen and was independent on the proliferative capacity of the lymphocytes; it was, however, curtailed by heat treatment of these cells. The 3 X 10(6), but not the 0.3 or 30 X 10(6) dose of semiallogeneic cells primed efficiently for a secondary hemagglutinin and cytotoxic response. I.v. injected 51Cr-labelled allogeneic and semiallogeneic lymphocytes showed different localized patterns. Between 4 and 24 hr after injection allogeneic lymphocytes were apparently more rapidly lost from the recipient lymph nodes. Collectively, these data indicate that the different immunizing properties of i.v. injected allogeneic and semiallogeneic lymphocytes are not simple consequences of the antigen dose transferred, but suggest that different localization patterns might influence the ability of these cells to induce cytotoxic and hemagglutinating alloantibody formation.     

The bacterial superantigen Staphylococcal enterotoxin B stimulates lymphocyte locomotor capacity during culture in vitro

The bacterial superantigen Staphylococcal enterotoxin B (SEB) was investigated for its effects on lymphocyte locomotion in vitro. Culture of peripheral blood mononuclear cells (PBMC) for 24-72 hr in SEB (1-100 micrograms/ml) increased the proportion of lymphocytes in locomotor (polarized) morphology and capable of invading collagen gels, to the same extent as the established locomotor activator, anti-CD3 (alpha-CD3), though the conventional antigen, tetanus toxoid was ineffective. The cells responding to SEB were predominantly T cells. SEB had no effect on lymphocyte locomotion in short-term (45 min) assays, thus its effect is to stimulate growth-related locomotor capacity and it does not act as a chemoattractant. During culture of PBMC in SEB, the chemokines interleukin-8 (IL-8) and macrophage chemotactic protein-1 (MCP-1) were released into the culture medium. The presence of anti-IL-8, but not of anti-MCP-1, either during culture or added to SEB culture supernatants and tested in short-term assays, inhibited the development of polarization suggesting that IL-8, which is a lymphocyte chemoattractant, also plays a key role in SEB-induced locomotor activation. Among SEB-activated lymphocytes, CD45RO+CD45RA- lymphocytes showed enhanced locomotor responses, but a relation was not found between locomotor activity and the presence of cell surface CD69.     

A 37-kilodalton glycoprotein of Babesia divergens is a major component of a protective fraction containing low-molecular-mass culture-derived exoantigens

The supernatants of in vitro cultures of Babesia divergens Rouen 1987 in human erythrocytes, obtained by using a semidefined medium based on human high-density lipoproteins, were fractionated by gel filtration chromatography into four fractions, F1 to F4. The crude supernatant as well as each fraction adjuvanted with Quil-A protected gerbils from mortality due to a homologous infectious challenge. Analysis of the humoral response of the 10 protected gerbils with fraction F4, containing major proteins with molecular masses lower than 50 kDa, showed that a few antigens (from 50 to 17 kDa) could be important candidates for an improved vaccine against B. divergens babesiosis. As an immunodominant response was directed against the 37-kDa antigen (Bd37) in two different B. divergens strains tested, a polyclonal antibody directed against Bd37 was produced in a rabbit. In an immunofluorescence assay, the anti-Bd37 antiserum strongly labelled small internal vesicles of the merozoites and the cell surface was diffusely labelled after fixation, whereas on live merozoites, this labelling was not observed. [3H]glucosamine-radiolabelling experiments demonstrate that Bd37 is a glycoprotein. The Bd37 protein can also be labelled with [14C]palmitate but not with [3H]myristic acid. In Triton X-114 temperature phase partitioning of B. divergens-infected erythrocyte extracts, Bd37 was exclusively found into the detergent phase, indicating that the palmitoylated Bd37 protein was in the membrane fraction. In the in vitro supernatant, the glycoprotein Bd37 was found in a nonpalmitoylated form, indicating excretion and/or release of the glycoprotein from the merozoite.     

Lung cancer surgery in elderly patients

Immuno-electron microscopic methods revealed that the protective antigen (PAg) of Leptospira interrogans serovar lai exists on the outer envelope sheathing the leptospiral cell body. PAg lost its protective activity after treatment by hydrolysis with 2 M formic acid at 100 C for 2 hr, or oxidation with periodate at 4 C for 40 hr. The antigenic oligosaccharide fraction was further purified from the hydrolyzed PAg by immunoaffinity column coupled with protective monoclonal antibody, LW2, and by gel filtration of HPLC. The antigenic oligosaccharide fraction contained two unknown sugars and 4-O-methylmannose (molar ratio 3:5:1). These findings suggested that these sugars are components of an antigenic determinant contributing to the protective immunity against serovar lai infection.     

Cortical dysplasia: an immunocytochemical study of three patients

The role that T and B lymphocytes play in the clearance of Giardia muris in the mouse model is well known, but the cytokines produced by CD4+ T cells in response to Giardia antigenic stimulation are unknown. In this study, we have determined how Giardia trophozoite antigenic crude extract and T cell mitogens can trigger the production of cytokines by Peyer's patch and spleen cells removed from infected animals. When Giardia trophozoite proteins were used to challenge the cells in vitro, IL-4, IL-5 and IFN-gamma were not detected in the culture supernatant. When the cells were challenged with Con-A, all three cytokines were released in vitro. However, the level of each cytokine released by the spleen or Peyer's patch cells varied with the latent, acute and elimination phases of the infection. The high levels of IL-4 and IL-5 released by Peyer's patch cells confirm the importance of IgA in the control of the infection. However, we propose that the relative success of G. muris in completing its life cycle in a primary infection might be due, in part, to the stimulation of a Th2-type response (IL-4, IL-5). A stronger Th1 response (IFN-gamma) may lead to a better control of the primary infection.     

Stress and the immune response in rats

The in vitro response of sensitized splenic lymphocytes to antigen (thyroglobulin) was increased by crowding and decreased by isolation in female rats. Both isolated and crowded male rats responded by a decrease in the in vitro reactivity of lymphocytes to antigen. The response of the lymphocytes to PHA was not altered in any consistent manner. Similar animals, both control and those immunized with thyroglobulin, were tested for an effect of in vivo injections of epinephrine on the in vitro reactivity of lymphocytes; epinephrine was given intraperitoneally 30 min before the rats were killed for removal of spleens. Incorporation of 3H-thymidine by lymphocytes was greater in control cultures (neither PHA nor antigen present) but there was a decreased response to either PHA or antigen when epinephrine had been injected.     

Human immunodeficiency virus type 1 Nef does not alter T-cell sensitivity to antigen-specific stimulation

We have developed an in vitro model to study the influence that human immunodeficiency virus type 1 (HIV-1) may have on the ability of T cells to respond to antigenic challenge. We have examined consequences of HIV-1 gene expression on T-cell activation in antigen-dependent T cells that have stably integrated copies of replication-defective proviral HIV-1. Virus production by HIV-infected, antigen-dependent T cells was induced in response to antigenic stimulation and then decreased as infected cells returned to a state of quiescence. Contrary to the predictions of models proposing that Nef alters signal transduction pathways in T lymphocytes and thereby alters cellular activation, Nef expression in antigen-dependent T-cell clones did not influence their proliferative responses to low or intermediate concentrations of antigen and did not affect other measures of T-cell activation, such as induction of interleukin 2 receptor alpha-chain expression and cytokine production. In addition, we found no evidence for alteration of T-cell responsiveness to antigen by the gag, pol, vif, tat, or rev gene of HIV-1.     

Antigenic variation and cross-reactivity in Bacteroides forsythus clinical isolates detected by western blot

Bacteroides forsythus is one of the etiologic agents of destructive periodontal diseases. Determining which antigenic components of the bacterium are recognized in the immune response of periodontitis patients is an important step in assessing strategies for vaccine development. The aim of this study was to identify the major strain-variable and cross-reactive antigens of B. forsythus clinical isolates recognized by serum IgG from patients with early-onset rapidly progressive periodontitis. Ten patient sera with measurable IgG against antigenic components of the species were identified by Western blot. Positive sera were tested by checkerboard ELISA to identify those most responsive to strain-variable antigens in nine clinical isolates and ATCC strain 43037. Correlation analysis of the ELISA data suggested that different subsets of isolates were preferentially recognized by different sera. Western blots revealed that certain sera also recognized major shared components across all the isolates, but preferential recognition of different isolate subsets by different patients was clearly confirmed. To determine if the variable antigens recognized were nonprotein, proteinase K-digested isolates were compared to undigested controls by Western blot. The main strain-variable antigens were proteinase resistant, while proteins at 200 and 210 kDa were identified as the major shared components. Two-dimensional SDS-PAGE revealed that these proteins are the quantitatively dominant heat-modifiable components of the cell envelope. Even though variable antigens are prominent in the immune response of patients, a cross-protective vaccine based on the shared envelope proteins of B. forsythus seems feasible in light of these observations.