- Top of page
Immunomodulation of perioval granulomas is a well-known phenomenon in schistosome-infected mice, but only little is known about granuloma modulation in other animal models of human schistosomiasis. In the present study, we explored immunomodulation of egg granulomas in the liver in a pig model of schistosomiasis japonica. Granuloma size was measured and T cells, B cells and IgG+ plasma cells in granulomas were quantified in pigs at 9, 12 and 21 weeks post infection (wpi) with Schistosoma japonicum. Granulomas were largest at 9 wpi, had decreased significantly in size at 12 wpi and remained small at 21 wpi (9 vs. 12 and 21 wpi: P < 0·05). The size of granulomas containing mature and immature eggs, respectively, did not differ significantly. The density of T (CD3ɛ+) cells and IgG+ plasma cells in granulomas was the same, irrespective of granuloma size and time points. B (CD79α+) cells were rare in granulomas. The results indicate that in pigs, S. japonicum egg granulomas in the liver are immunomodulated at an early stage of infection, and that not only mature but also immature eggs induce a marked granulomatous reaction in this species.
The zoonotic trematode Schistosoma japonicum is an important parasite of humans in endemic regions of China, the Philippines and Indonesia (1,2). The major lesions in schistosome infections are caused by the cell-mediated immune response against parasite egg antigens, resulting in perioval granuloma formation in target organs (3,4). The immune response is also considered instrumental for the development of liver fibrosis, the most serious pathological consequence of human S. japonicum and S. mansoni infections. In experimental murine schistosomiasis japonica, granuloma formation is a CD4+ T cell-dependent reaction to soluble egg antigens (SEA) in which Th2 responses play a dominant role (5,6). An important feature of murine schistosomiasis japonica and mansoni is granuloma immunomodulation, i.e. the spontaneous down-regulation of the response to newly deposited eggs with increased duration of infection (7,8). Immunomodulation has been linked to a decrease of T cell responses to egg antigens and is beneficial to the host in that it limits tissue injury and morbidity. Down-regulation by CD8+ T cells (9), immunoregulatory effects of the Th2-associated cytokine IL-10 (10), anti-inflammatory B cell activity (11), idiotypic interactions (12), and reduction of SEA-specific T cell responsiveness mediated by adhesion molecules (13) are among the mechanisms proposed to explain this probably multifactorial process in the murine S. mansoni model. In murine S. japonicum infection both cellular and humoral immune responses appear to be involved, the process being mediated by CD8+ T cells in the acute and by circulating IgG1 antibody in the chronic phase of infection (8).
As animal models in schistosomiasis research, mice suffer the drawbacks of small size and short life span, making it impossible to study either infections of different levels of intensity or long-term infections comparable to those in humans (14). Due to the development of severe hepatic pathology, murine models are suitable neither for chemotherapy nor for reinfection studies (15). In the search for other animal models of S. japonicum infection, the pig has been found to share major pathological manifestations of the disease with humans (16,17). One important question regarding the pig as an animal model of human schistosomiasis is whether or not it displays granuloma immunomodulation. If so, further studies on molecular immunopathogenesis in this species should be of great comparative interest. In order to investigate granuloma modulation in the pig we measured the size of perioval granulomas in the livers of S. japonicum-infected pigs at 9, 12 and 21 weeks post infection (wpi). We also assessed possible changes of T and B cell infiltration of granulomas with progression of infection.
Twenty-one helminth-naive, specific pathogen-free Danish Landrace/Yorkshire/Duroc crossbred pigs, aged 8–12 weeks were used. These pigs were previously used for another study and details of feeding and management conditions are described in that context (18). The pigs were treated in accordance with animal ethics legislation of Denmark. Fifteen pigs, divided into three groups of five pigs each, were inoculated intramuscularly with S. japonicum cercariae suspended in Iscove's medium as described elsewhere (19). The cercariae were of a Chinese mainland strain of the parasite and were maintained in Oncomelania hupensis hupensis snails at the DBL-Institute for Health Research and Development, Charlottenlund, Denmark. Each pig received 850 cercariae. Two groups of pigs were inoculated at week 0 and necropsied at 12 and 21 wpi, respectively, and the third group was inoculated at week 12 and necropsied at 9 wpi. Six pigs were non-infected controls, inoculated with Iscove's medium only, and killed at 12 or 21 weeks post inoculation. The pigs were killed with an overdose of pentobarbital intravenously. After vascular perfusion for recovery of schistosome worms, tissue samples from various organs were fixed in 10% neutral-buffered formalin and embedded in paraffin (18). For the present study, 4 µm-thick serial sections of liver tissue were cut from each infected pig. Examinations were performed on coded slides to blind the investigator to the identity and group affiliation of the pigs. Lymph node sections from the non-infected pigs were used as control materials in immunohistochemistry.
Only granulomas containing a single and viable egg were included in the study.
The first section of each series displaying the egg centrally in the granuloma was stained with haematoxylin and eosin (HE) and used for measurement of granuloma size and for histopathological evaluation. In each pig, at least 10 granulomas were measured. The developmental stage of each egg, i.e. a mature egg with a fully developed miracidium (20) (Figure 1a), or an immature egg containing only differentiating cells (Figure 1c), was noted. The total number of granulomas with mature and immature eggs, respectively, available for examination were 34 and 22 at 9 wpi, 34 and 18 at 12 wpi, and 42 and 15 at 21 wpi. The longest diameter of the granuloma and that perpendicular to it were measured, using an ocular micrometer, and the granuloma area was calculated (area = π × r1 × r2). The last section of each series was stained as the first and used to check granuloma identity throughout the series. The remaining sections of each series were used for immunohistochemical detection of T and B cells. Primary antibodies included polyclonal rabbit anti-human CD3ɛ (A0452, Dako Cytomation A/S, Glostrup, Denmark; diluted 1 : 100), monoclonal mouse anti-human CD79α (M 7051, Dako; 1 : 50) and polyclonal goat anti-porcine IgG-Fc (Bethyl Laboratories Inc., Montgomery, TX, USA; 1 : 9000). Pretreatment for antigen retrieval was performed with 0·05% pronase (Dako) for 20 min at 37°C for CD3ɛ and IgG-Fc and with Target Retrieval Solution (Dako) for 25 min in a water bath at 97°C for CD79α. Endogenous peroxidase was quenched with hydrogen peroxide (3% in TBS for 10 min for CD3ɛ and IgG-Fc and 1·5% in methanol for 30 min for CD79α). The Avidin–Biotin Blocking Kit (Vector Laboratories Inc., Burlingame, CA, USA) was applied for CD79α. Non-specific protein binding was minimized by incubation with normal horse serum for CD3ɛ and IgG-Fc and with normal rabbit serum for CD79α. The sections were incubated with the primary antibodies for 18 h at 4°C, after which immunoreactivity was demonstrated with the Streptavidin Biotin Complex/Horseradish Peroxidase (StreptABC/HRP, Dako) method, using as secondary antibodies biotinylated goat anti-rabbit Ig for CD3ɛ, rabbit anti-goat Ig (both Vector; 1 : 200) for IgG-Fc or rabbit anti-mouse Ig (Dako; 1 : 200) for CD79α. The primary antibodies were replaced with adequately diluted rabbit Ig, normal goat serum or mouse IgG1, respectively, for specificity control. The sections were developed with 3,3′diaminobenzidine tetrahydrochloride (Dako) and counterstained with haematoxylin.
Figure 1. Examples of hepatic perioval granulomas. (a–c) haematoxylin and eosin, (d–f) StreptABC/HRP. (a) Vigorous (non-modulated) granuloma at 9 weeks post infection with a mature, viable, central egg and with a granuloma-associated lymphoid nodule (arrow). Inset: the egg at higher magnification. (b) Vigorous (non-modulated) granuloma at 9 weeks post infection with an immature, viable egg. (c) Modulated granuloma at 21 weeks post infection with an immature, viable egg. Inset: the egg at higher magnification. (d) Granuloma (identical with that in Figure 1a) immunostained with anti-CD3α. Numerous positive cells in the periphery and in the granuloma-associated lymphoid nodule (top left). (e) Granuloma at 9 weeks post infection with a mature, viable egg immunostained with anti-IgG-Fc. Scattered positive cells in the periphery. (f) Granuloma (identical with that in Figure 1a) immunostained with anti-CD79α. Only occasional positive cells are present in the granuloma periphery. Numerous positive cells are present in the granuloma-associated lymphoid nodule (top left).
Download figure to PowerPoint
Initial light microscopic examination of the immunostained sections revealed that CD3ɛ+ cells were abundant in many granulomas, whereas CD79α+ and IgG+ cells were less numerous. CD3ɛ+ cells were counted in digital images of granulomas at 20× magnification using the Leica DC 100 and Adobe Photoshop 5·0 software. A grid with small squares outlined by thin lines was positioned on top of the images to facilitate counting. CD79α+ and IgG+ cells were counted manually with a light microscope at 40× magnification without the aid of digitalization. The cells in each granuloma were counted three times by one observer (S.L.) and the mean of the counts was calculated. The mean granuloma area and the mean cell counts for each pig formed the basis for calculations of mean group values. Statistical analyses were carried out using Statistical Analysis System computer software (SAS Institute Inc., 2000). The variations in the granuloma area and cell counts at different time points of infection were assessed using the MIXED procedure. Four statistical models were used. Model 1 dealt with fixed effects on the mean granuloma area of (1) duration of infection and (2) maturity of the egg. Pairwise comparisons between the different time points were made with the Tukey–Kramer method. Models 2 and 3 determined the mean cell counts as functions of the mean granuloma area. In these two models, Type 3 tests of fixed effects were included as part of the analyses. In the last model the possible relationship between cell counts and granuloma area was explored by regression analysis. A probability value of < 0·05 was considered significant.
The granuloma morphology was typical of S. japonicum schistosome egg granulomas in the pig (21), showing variable combinations of epithelioid macrophages, giant cells, eosinophils, lymphocytes and plasma cells surrounding the central egg, and occasional closely apposed small lymphoid nodules (Figure 1a,b, both 9 wpi). The mean granuloma area was smaller at 12 and 21 wpi, respectively, than at 9 wpi (P < 0·05), both for granulomas with mature and those with immature eggs (Figure 2a). There was no significant difference in granuloma size between 12 and 21 wpi. Typical large, vigorous granulomas are shown in Figure 1a (9 wpi, mature egg) and 1b (9 wpi, immature egg) and a typical small, modulated granuloma in Figure 1c (21 wpi).
Figure 2. (a) The area (mean ± SD) of liver granulomas with mature and immature eggs, respectively, in the pigs at different time points post infection. *12 and 21 vs 9 weeks post infection: P < 0.05, both for granulomas with mature and immature eggs. (b) The relationship between CD3ɛ+ cell counts and mean area (102 µm2) for liver granulomas with mature and immature eggs, respectively. Continuous line and ○: granulomas with a mature egg; broken line and +: granulomas with an immature egg. (c) The relationship between IgG-Fc+ cell counts and mean area (102µm2) for liver granulomas with mature and immature eggs, respectively. Continuous line and ○: granulomas with a mature egg; broken line and +: granulomas with an immature egg.
Download figure to PowerPoint
The pattern of immunostaining of CD3ɛ+, CD79α+ and IgG+ cells in the lymph nodes of control pigs was consistent with results of other studies (22,23). Typical immunostaining of granuloma cells for CD3ɛ and IgG is presented in Figure 1d and e, respectively, (both 9 wpi). The mean numbers of CD3ɛ+ and IgG+ cells per granuloma were highest at 9 wpi, had declined at 12 wpi and remained decreased at 21 wpi, both in granulomas with mature and in those with immature eggs. The mean ratios between cell numbers and granuloma area were for CD3ɛ+ cells 0·30, 0·31 and 0·28 for granulomas with mature eggs and 0·36, 0·39 and 0·35 for those with immature eggs at 9, 12 and 21 wpi, respectively. The corresponding mean ratios for IgG+ cells were for granulomas with mature eggs 0·012, 0·010 and 0·007 and for immature eggs 0·011, 0·007 and 0·006. Regression analysis showed a linear positive correlation between cell numbers and granuloma area in individual pigs (Figure 2b,c), statistically significant both for granulomas with mature eggs (CD3ɛ+ cells: P < 0·01; IgG+ cells: P < 0·05) and for those with immature eggs (CD3ɛ+ and IgG+ cells: P < 0·01). Only minimal numbers of CD79α+ cells were found in granulomas at the different time points of infection. In contrast, the small lymphoid nodules adjacent to granulomas contained numerous CD79α+ cells (Figure 1f, 9 wpi).
Immunomodulation of liver granulomas has been shown to occur between 4 and 16 weeks of S. japonicum infection in several strains of mice (24–26). The significant decrease of granuloma size between 9 and 12 wpi observed in the present study indicates that immunomodulation is a feature also of porcine schistosomiasis japonica and occurs at an early stage of infection. We studied quantitatively T and B lymphocytes in granulomas by the use of anti-human CD3ɛ and CD79α antibodies on paraffin sections. These antibodies are accepted universal markers of T and B lymphocytes, respectively, in paraffin-embedded materials and have been shown to cross-react with porcine lymphocyte antigens (22). Our finding that T cell numbers in granulomas decreased in a linear fashion along with the decrease in granuloma size show that T cell numbers influenced the size of granulomas, whereas the density of T cells appeared unchanged. In a previous study of immune cell phenotypes in S. japonicum egg granulomas of pigs, we demonstrated the occurrence of CD4+, CD8+ and γδ T cells and discussed the possible presence of CD4+CD8+ dual positive cells in the perioval response (27). Though the present study showed an unaffected density of T cells with modulation, the composition of various T cell subsets may very well have differed between the non-modulated and modulated states, as shown earlier in mice with S. mansoni (28,29). Further immunohistochemical studies to explore these matters using markers for T cell subsets in the pig/S. japonicum model are warranted.
We detected only few B lymphocytes in the granulomas at any one time point, but several B lymphocytes were present in the granuloma-associated lymphoid nodules. Such nodules appear to be a unique feature of pigs in their granulomatous response to S. japonicum eggs and are postulated to be a site of B cell activity, possibly supplying the granulomas with antibody-producing plasma cells (27). However, it was impossible to assess any changes of their content of B lymphocytes of interest in the context of the modulation process, since these small and peripherally located nodules were detected only infrequently in the centrally sectioned granulomas included in the present study. B cell activity in granulomas was also indicated by the presence of IgG-producing plasma cells, concordant with earlier observations in S. japonicum-infected pigs (27). De Brito et al. (30) discussed the possibility that in human schistosomiasis mansoni, antibodies produced by plasma cells in granulomas may function to neutralize secreted egg antigens. Antibodies produced locally may also mediate cytotoxic reactions through Fc receptors of inflammatory cells (31). The present study showed that the density of IgG+ cells in granulomas, as that of CD3ɛ+ cells, was maintained irrespective of modulation, cell numbers only diminishing along with the decrease of the granuloma area, and there was no indication of any significant increase of B cell activity in the modulated state, in contrast to findings in S. mansoni-infected mice (32).
In our study, both granulomas around mature eggs and those circumscribing immature eggs were evaluated, in contrast to the great majority of previous investigations on immunomodulation in murine models, where only granulomas with mature eggs have been explored. An unexpected finding was that the maturation status of the egg influenced neither granuloma size at the early stage of infection at 9 wpi, nor the degree of modulation at the later stages. It has been shown that in murine schistosomiasis japonica, immature eggs are markedly less antigenic than mature ones, and that the granulomatous reaction is only small until the egg displays a miracidium (20,33). Kawanaka and Carter (34) reported that although excretory–secretory proteins were released from both immature and mature S. japonicum eggs, only proteins from mature eggs were reactive with human infected sera. However, it has been demonstrated in the murine model of S. mansoni that the immunogenicity of schistosome egg antigens may vary significantly according to the genetic background of the host and may correlate with egg-induced immunopathology (35). The full-blown granuloma formation around immature S. japonicum eggs in our pigs might indicate that even minute amounts of antigens secreted during egg maturation may be recognized and lead to a delayed hypersensitivity response in this species.