Regulation of Granulomatous Inflammation in Experimental Models of Schistosomiasis

Schistosomiasis is a parasitic disease affecting more than 200 million people. Its major pathology is granulomatous inflammation, a cellular immune response to antigens secreted by schistosome ova. The murine models of this disease have been widely studied because they permit a variety of genetic and other experimental approaches. The murine infections differ from the human, monkey, and baboon infections in many ways, especially in the number of adult worms per unit of body weight and the distribution of ova between the liver and mesenteric circulation. However, all of these species develop hepatic granulomatous (HG) inflammations that have similar dynamics and cellular compositions and are spontaneously downmodulated. Therefore, knowledge about the mechanisms of regulation of these inflammations in the mouse models is relevant to the other species. Most of the available evidence is related to HG inflammation resulting from infection with cercariae. In occasional studies workers employ another model, pulmonary granulomas (PG) induced by injection of ova or beads coated with schistosomal egg antigens (SEA) into naïve or schistosome-infected mice.

This review differs from previous reviews of this disease in that it synthesizes relevant older and newer studies into a sequence of microenvironmental, cellular, molecular, and immunological events resulting in granuloma formation and eventually downregulation. It also presents different viewpoints and new questions about some controversial and/or confusing subjects, including mechanisms of regulation by Th1- and Th2-type cytokines, chemokines, and other types of molecules and their receptors; signal transduction pathways; different types of regulatory cells; the role of gut-associated lymphoid tissue (GALT), B cells, and Fc gamma receptors (FcγR); and finally, the balance between T effector (T_E) and T regulatory (T_R) cells in the control of immunity and pathology.

SCHISTOSOME LIFE CYCLE

Schistosomiasis is an ancient and chronic disease of humans, nonhuman primates, other mammals, and birds that is caused by a number of species of flatworms (Platyhelminthes). In this review I focus on the murine model diseases caused by two major human pathogens, Schistosoma mansoni and Schistosoma japonicum. The mammalian host is infected during contact with freshwater contaminated with infectious cercariae produced by various species of snails. The cercariae penetrate intact skin and through a series of morphological, membrane, biochemical, and antigenic changes transform into schistosomulae. After days in subcutaneous tissue, the somulae travel to the lung, where they undergo adaptations for intravascular migration. From the lung the somulae are distributed to all organs. Most somulae eventually reach the liver, where they attain sexual maturity and enter the portal venous system. The adult worms mate and then travel to the small mesenteric venules. The females release ova, which migrate though the venule wall, the lamina propria, and the gut epithelium into the lumen and ultimately to the outside environment with the feces. (The embryos in the ova, the miracidia, are released and infect snails, in which they develop into infectious cercariae.) Some ova are retained in local tissues, but others go to the portal system, where they lodge in the sinusoids and induce HG. In infected mice about 75% of the ova are in the liver and the rest are in other tissues, including the mesentery (143).

GENETIC CONTROL OF CELLULAR AND HUMORAL IMMUNE RESPONSES

Table 1 summarizes experiments which examined genetic control of various parameters of cellular and humoral immune responses. These experiments showed that there is H-2 genetic control of epitope recognition (6, 56), the antibody response (7, 74), and development of Th1 and Th2 cytokine-expressing subsets (7), but there is non-H-2 control of HG size, portal hypertension, and fibrosis (20, 21, 36, 43). The apparent discrepancy between the H-2 gene control of Th1-Th2 subset development and the non-H2-control of granulomatous inflammation, which is CD4⁺T cell dependent, remains to be explained.

TABLE 1.

Genetic control of immune responses in schistosomiasis

Parameter	Genetic loci	Observation(s)	Reference(s)
Epitope recognition	H-2	C3H and CBA Th cells show strong responses and C57BL/6 Th cells show a weak response to Sm-p40, a major SEA immunogen	56
		Opposite responses by these cells to 62-kDA SEA component	6
Antibody response	H-2	Different levels of antibody responses to different SEA immunogens in different mouse strains	74
Cellular immune responses	H-2	Th1-Th2 subset development differed in different mouse strains	7
	Non-H-2, more than one gene	HG size, fibrosis, and portal hypertension different in different mouse strains	20, 21, 36, 43

GRANULOMA FORMATION

Dynamics of involvement of different microenvironments, cell populations, cellular activation, differentiation, and expansion.

SEA-reactive major histocompatibility complex class II (MHC-II)-bearing CD4⁺ (αβ⁺) T memory-effector (T_M/E) cells are required for HG formation in mice (5, 62, 90). During the early weeks of an infection, T-lymphocyte activation and T_M/E cell expansion and differentiation required for HG formation occur in extrahepatic microenvironments, including bronchial, mediastinal, and hepatic lymph nodes (LN) and the spleen, where somulae and their antigens are present (75, 76). Activation, expansion, and differentiation of SEA-reactive B_M/E cells culminating in antibody formation also occur in these extrahepatic lymphoid sites (75, 76) and mucosal tissues (38). Somular antigens which cross-react with SEA (85) must induce this early development and expansion of SEA-reactive lymphocytes. Thus, in a mouse 4 weeks postinfection, when egg laying begins and several weeks before HG appear, SEA-reactive lymphocytes are primed and proliferate in the spleen (78). These SEA-reactive lymphoblasts are then recruited into the HG (121).

Baboons infected by a single exposure to cercariae of S. mansoni developed HG whose size and eosinophilia peaked 6 weeks after infection (44). After multiple cercarial exposures, HG size and eosinophilia peaked at 9 weeks. In both groups the size of the HG diminished by week 13, but downsizing occurred more rapidly in the baboons with multiple cercarial exposures (44).

Granulomas also develop in the intestines of schistosome-infected mice. The granuloma sizes in the liver, colonic mucosa, and ileal Peyer's patches peak 8 weeks postinfection and then spontaneously decrease (143). There are differences in cellular composition; HG contain the largest number of T and B lymphocytes, eosinophils, and mast cells, whereas ileal granulomas consist mainly of macrophages (145). There are different patterns of distribution of T and B lymphocytes within granulomas in different tissues (145).

Roles of costimulatory, adhesion, and chemokine molecules.

The binding of costimulatory B7 ligands on antigen-presenting cells to CD28 receptors on T cells can enhance activation and proliferation of T lymphocytes (54). B7-2 expression by HG cells in S. mansoni-infected mice has been implicated in early T-lymphocyte activation, expansion, and differentiation because it was elevated 6.5 weeks postinfection (113). B7-2 was also involved in PG formation elicited by injection of ova; anti-B7-2 treatment greatly inhibited this process (133). However, HG size was not significantly different in S. mansoni-infected B7-1, B7-2, or double knockout (KO) and wild-type (WT) mice (57), suggesting that there is only a marginal requirement for B7-1 and B7-2 molecules. The significance of elevated expression of B7-1 and CD40 by splenic CD8α⁺ dendritic cells (DC) 8 weeks after S. mansoni infection (132) remains to be assessed.

Antibodies to the adhesion molecules intercellular adhesion molecule 1 (ICAM-1), lymphocyte function-associated antigen 1 (LFA-1), and very late antigen 4 (VLA-4) inhibited proliferation and interleukin-2 (IL-2) and IL-4 production by splenic and HG lymphocytes of acutely infected mice (82). Injection of SEA-coated Sepharose beads into S. mansoni-infected mice stimulated expression of ICAM-1, LFA-1, VLA-4, and VLA-6 around HG, suggesting that these adhesion molecules participate in initiation and maintenance of HG (64). Eight weeks postinfection ICAM-1, LFA-1, and VLA-4 were upregulated in ileal colonic granulomas, and syndecan 1 immunoreactive B lymphocytes were present close to SEA-laden macrophages in inner areas of ileal and colonic granulomas (65).

The numbers of adult worms in S. japonicum-infected mast cell-deficient mice were similar to the numbers in WT mice, but the HG size was significantly reduced in the former mice (105). The reduction was attributed to deficient production of eosinophil chemotactic factor. The levels of chemokine monocyte chemoattractant protein 1 (MCP-1) mRNA and protein were increased in PG induced by injection of S. mansoni ova into infected or noninfected mice (33). MCP-1 expression was greatest within microvascular adventitial cells or pericytes, as well as mononuclear cells associated with PG. Injection of antibodies to MCP-1 inhibited PG formation. MCP-1 was expressed in HG and branches of the hepatic artery in S. mansoni-infected mice inoculated with SEA-coated beads (64). The chemokine macrophage inflammatory protein 1α (MIP-1α) was identified in macrophages in primary PG (86). Treatment with antibody to MIP-1α decreased PG formation. When schistosome ova were injected into infected mice, there were distinct coordinated patterns of chemokine expression in the liver and draining LN (107).

Cytokines and chemokine effects were interrelated. Thus, neutralization of tumor necrosis factor alpha (TNF-α) by antibody inhibited chemokine production by pulmonary cells (111). The level of monocyte chemotactic protein 3, which binds to CCL7 receptor, was elevated by IL-4, and monocyte chemotactic protein 3 contributed to eosinophil recruitment (124).

Cytokines.

The Th1 and T-cytotoxic 1 T-cell subsets produce type 1 cytokines (IL-2 and interferon gamma [IFN-γ]), whereas the Th2 and T-cytotoxic 2 T-cell subsets produce type 2 cytokines (IL-4, IL-5, and IL-10) (99, 100). Eight weeks after S. mansoni and S. japonicum infection Th2-type cytokine production by spleen cells (SC) and HG cells was predominant compared with Th1-type cytokine production (37, 55, 151). There appeared to be cross-regulation of cytokine production: as Th2 cytokine production increased, Th1 cytokine production decreased (55, 126, 151).

These observations led to the prevalent dogma that Th2-type cytokine expression is predominant and is correlated with HG formation and that increased Th2 cytokine production is accompanied by reduced Th1 cytokine production. However, this view was challenged (78) because it was largely based on in vitro cytokine production by exogenous SEA-stimulated SC, ex vivo cytokine production induced by endogenous SEA was not examined, and the comparative dynamics of ex vivo cytokine responses of SC and HG cells to endogenous SEA also were not examined. Moreover, based on a previous study (95), it was predicted that exogenous SEA would dynamically and quantitatively alter the cytokine responses of SC and HG cells. Thus, the previously observed Th2 cytokine-dominated responses to exogenous SEA might have been skewed compared to the responses induced by endogenous SEA and, therefore, would not be expected to correlate with HG formation.

The validity of this challenge and these predictions was supported by the results of a study (78) in which the ex vivo endogenous SEA-induced cytokine responses by SC and HG cells were dynamically and quantitatively different from the responses stimulated by exogenous SEA. Unexpectedly, SC and HG cell responses were also dramatically different. Ex vivo HG cells showed a >10-fold-greater frequency of IL-4-, IL-5-, and IFN-γ-secreting HG cells than SC. Eight weeks postinfection HG cells ex vivo secreted much higher concentrations of IL-4 and IL-5 but much lower concentrations of IFN-Υ than SC secreted. Endogenous SEA induced HG cells to produce much larger amounts of IL-4, IL-5, and IL-10 than SC produced. The ex vivo responses were better correlated with the dynamics of HG formation than the exogenous SEA-stimulated responses were. Whereas exogenous SEA induced SC to produce larger amounts of cytokine, exogenous SEA downregulated cytokine production by HG cells (78), confirming previous findings (95).

Endogenous SEA-evoked ex vivo and constitutive cytokine expression (both numbers of cytokine-producing cells and amounts of cytokine produced) by CD4⁺ SC preceded cytokine expression by CD4⁺ HG cells and was coordinate (78). The data are compatible with the development of SEA-reactive cytokine-producing cells in the spleen and the subsequent migration of these cells into liver (78, 121).

Table 2 shows the effects on HG size of introduction during infection of blocking antibodies to some cytokines and cytokine receptors. By this criterion, IL-2 (22), TNF-α (71), and the IFN-γ receptor (119) in S. mansoni-infected mice and IL-5 (28) and the IL-2 receptor (27) in S. japonicum-infected mice participated in HG pathogenesis. In S. mansoni-infected mice IL-2 (22), IL-5 (125), IL-10 (15, 40), IFN-γ (125), and the IFN-γ receptor (119) and in S. japonicum-infected animals the IL-2 receptor (27) were also involved in determining the cellular composition of the HG.

TABLE 2.

Regulation of HG inflammation in murine schistosomiasis by injection of antibodies to cytokines and cytokine receptors

Species	Antibody specificity	Effects on granulomas			Reference
		Size	Cellular composition	Fibrosis
S. mansoni	IL-2	Decrease	Eosinophils decrease		22
S. mansoni	IL-4	None		Decrease	26
S. japonicum	IL-4	Increase		Decrease	23
S. japonicum	IL-5	Decrease		None	25
S. mansoni	IL-5	None	Eosinophils decrease	None	125
S. mansoni	IL-10	Increase	Eosinophils increase		40
S. mansoni	IL-10	Increase	Eosinophils increase		15
S. mansoni	IFN-γ	None	Eosinophils increase	None	125
S. japonicum	IFN-γ	Decrease		None	28
S. mansoni	TNFα	Decrease			71
S. mansoni	IFN-γ receptor	Decrease	Decrease in epithelioid cells		118
S. mansoni	IL-2 receptor	None		None	22
S. japonicum	IL-2 receptor	Increase	Decrease in eosinophils	None	27

Table 3 shows that the introduction during infection of some recombinant cytokines, including recombinant IL-2 (91), recombinant IL-7 (146), and recombinant TNF-α (3, 71), enhanced HG size. Recombinant TNF-α occasionally did not restore HG formation (24), suggesting that this cytokine was only marginally required.

TABLE 3.

Regulation of granulomatous inflammation by injection of recombinant cytokines

Cytokine	Injection schedule and/or recipients	Effects on granulomas	Reference
IL-2	Daily during acute or chronic infection	Increase HG size	91
IL-7	Before infection	Increase HG size, fibrosis	146
IL-10	4 to 8 weeks postinfection; mice receiving ova intravenously	Decrease PG size	48
IL-10Fc	4 to 8 weeks postinfection	Decrease HG size	48
TNF-α	SCID mice	Restore HG size	2
TNF-α	During chronic infection	Increase HG size	71
TNF-α	SCID mice	None on HG	24
IL-12	Mice receiving ova intravenously	Decrease PG size	150
IL-12	Daily 5 to 7.5 weeks postinfection	Decrease HG size, fibrosis	16

Table 4 shows that infected mice with KO of some genes for cytokines and cytokine receptors developed smaller HG than WT mice developed, implicating IL-4 (95), the combined effects of IL-4 and IL-13 (42), and the IL-4α (70) and IFN-γ (104, 118) receptors in HG pathogenesis. On the other hand, deletion of the IL-13 gene did not affect this process (42).

TABLE 4.

Regulation of HG inflammation in mice with deletions of genes for T and B cells, cytokines, and Fcγ and substance P receptors

Deletion genes	Effects on granulomas
	Acute	Downregulation	Cellular composition	Fibrosis	Reference
B cell	Increase			Increase	47
B cell	Increase	Abolished		Increase	67
CD8+ T	None	None		None	154
SCID	Decrease				2
IL-4	Decrease		Fewer eosinophils, no mast cells		95
IL-10	Increase	None			149
IL-13	None		None	Decrease	42
IL-4/IL-13	Decrease		Fewer eosophils	Decrease	42
IFN-γ				None	2
IFN-γ	None	None		Decrease	154
IL-4α receptor	Decrease				70
IFN-γ receptor	None		None	None	1
IFN-γ receptor	Decrease		Reduction of infiltrating cells		118
IFN-γ receptor	Decrease		More immature, proliferating monocytes	Increase	104
FcɛR	Increase			Increase	68
FcγR	Increase	Abolished			68
Substance P receptor	Decrease		None		11

Additional evidence further dissociated the Th2 cytokine and HG responses (81). Eight weeks after mice deficient in the third component (C3) of complement were infected with S. mansoni, SEA-specific Th2 cytokine (IL-5, Il-6, IL-10, IL-13) production by the SC was significantly reduced. Nevertheless, the HG sizes in WT and C3 KO mice were indistinguishable. It was not determined whether the cellular compositions of the HG were similar or different in the WT and deficient mice.

In summary, diverse evidence strongly implicates both Th1 cytokines (IL-2, IFN-γ) and a Th1 cytokine receptor (IFN-γ) and Th2 (IL-4, IL-4-IL-13, IL-5, TNF-α) and a Th2 cytokine receptor (IL-4α) in determination of HG size. Evidence obtained by King et al. (78) and the failure of the granulomatous response to switch to Th1 in IL-4-deficient mice (95) indicate that cross-regulation of the Th1 cytokine response by Th2 cytokines does not occur. Other data show that the Th1 and Th2 cytokine responses are interdependent: the IL-4 response influences the IL-2 response (153), and IL-2 contributes to the IL-5 response (96). There is also evidence (Tables 2 and 4) that Th1 cytokines (IL-2, IFN-γ) and a Th1 cytokine receptor (IFN-γ) and Th2 cytokines (IL-4, IL-5) are involved in determination of the cellular composition of the HG.

Signal transduction pathways.

Mice with deletions of signal transducer and activator of transcription 4 (STAT4) and STAT6, involved in signal transduction pathways for Th1 and Th2 cell development, respectively, were inoculated with ova or infected with cercariae of S. mansoni. Seven weeks later they were inoculated intravenously with ova and then sacrificed 8 days later (73). In the group inoculated with ova the PG size was reduced significantly in the STAT6 KO mice but not in the STAT4 KO mice. In the the group inoculated with infected ova the HG and PG sizes were reduced in the STAT6 KO mice but not in the STAT4 KO mice. Therefore, Th2 cytokines, but not Th1 cytokines, appeared to be required for egg-elicited HG and PG formation. However, the relevance of STAT4 and STAT6 to HG formation was not established because infected mice were also inoculated with ova.

SC or mesenteric LN from infected STAT6 or IL-4α receptor KO mice produced low levels of IL-4 and IL-5 compared to the levels produced by WT mice (69, 73). Mesenteric LN from these mice also had smaller numbers of CD4⁺ T cells producing these cytokines, suggesting that IL-4 and STAT6 signaling determined the Th2 cell frequency but were not essential for Th2 cell differentiation. The HG from these mice made IL-4 or IL-5, but their T cells produced only small amounts of IFN-γ (94).

Signals through the T-cell receptor (TCR) and cytokine receptor initiate Th1 cytokine expression via STAT4 and induction of transcription factor T-bet (116). Signals that activate STAT6 induce transcription factor GATA3, leading to Th2 cytokine expression. Coordinate Th1 and Th2 cytokine expression in the spleen and HG (78) might result from simultaneous triggering of the Th1 and Th2 cytokine pathways in different CD4⁺ T-cell populations by different egg antigen epitopes (17), reflecting the great heterogeneity of the TCR repertoire (59). Greater production of IL-4, IL-5, and IL-10 suggests that there is greater expression of GATA3 by HG than by SC. Greater production of IFN-γ by SC indicates that there is greater expression of T-bet by SC. Simultaneous expression of Th1 and Th2 cytokines by both SC and HG cells suggests that both transcription factors are produced by different populations of SEA-specific CD4⁺ T cells.

Neuropeptides.

As shown in Table 4, infection of mice with deletion of the substance P receptor decreased HG size, involving this receptor and, by association, its ligand, substance P, in HG formation (11).

Cellular requirements.

HG formation is initiated by delayed-type hypersensitivity (DTH) mediated by SEA-reactive CD4⁺ (αβ⁺) MHC-II-dependent T cells (5, 59, 90). The gene KO experiments (Table 4) showed that HG formation does not require CD8⁺ T cells (154) or B cells (47, 67). Another study showed that B cells are required for Th2 cytokine responses but not for HG formation (58).

Consistent with the requirement of CD4⁺ T lymphocytes for HG formation, cytokines whose expression correlated with HG size were produced by CD4⁺ T cells in the spleen and HG (78). However, in contrast to the spleen, where the IL-4- and IL-5-producing cells were all CD4⁺ T cells, about 50% of the HG cells secreting these cytokines were CD4⁺ T cells (78). IL-4 and IL-5 are produced by cells other than CD4⁺ T cells, including NK cells (101), CD8⁺ T cells (109), γ/δ cells (115), activated eosinophils (120), basophils (41), and non-B, non-T cells (123). HG cells produce other cytokines implicated in HG pathogenesis, including IL-10 (95), TNF-α (13, 84), and transforming growth factor β (TGF-β) (13).

HG formation could be initiated by a few T cells recruited to the liver that then expand locally or by T cells that are activated and expanded and whose receptors (TCR) are diversified extrahepatically and then home into the hepatic DTH inflammatory site induced by ova. The finding that the TCR repertoire of a single HG was very diverse indicated that most of the T cells recruited to these lesions were activated and expanded and that their TCR repertoire was diversified in extrahepatic microenvironments (59). This study also showed that some non-SEA-specific T cells homed into the HG.

IgE.

HG size was significantly decreased in IgE-deficient SJA/9 mice compared to the HG size in C57BL/6 or SJL/J mice infected with S. japonicum (106). In another study S. mansoni-infected mice from which the immunoglobulin E (IgE) gene was deleted developed smaller HG despite increased worm burdens (77). This effect was attributed to IgE-mediated activation of mast cells, which are present in HG, leading to release of inflammatory mediators and cytokines that promote HG formation. However, inasmuch as infected FcɛR KO mice develop larger HG than WT mice develop (68), immune complexes (IC) of IgE antibody and SEA might react with FcɛR on various cells to promote HG formation by as-yet-undetermined pathways.

FIBROSIS

Fibrosis develops in granulomas during the chronic phase of granulomatous inflammation in murine schistosomiasis. A variety of molecules stimulate the differentiation of stellate cells into myofibroblasts that secrete extracellular matrix proteins, including collagens, fibronectin, and glycosaminoglycans. Fibrogenesis is a dynamic process regulated by cytokines produced by macrophages, lymphocytes, and fibrocytes (34, 147). Experiments performed with antibodies to cytokines (Table 2) implicated IL-4 in upregulation of fibrosis (23, 26). Injection of recombinant cytokines (Table 3) resulted in association of IL-12 with upregulation of this process (16). KO experiments (Table 4) inconsistently implicated IFN-γ (2, 154) in fibrosis.

TGF-β is associated with hepatic fibrosis based on a variety of evidence. Treatment of cultured hepatic cells with TGF-β1 increased the level of type I procollagen mRNA, and an increase in TGF-β1 gene expression preceded the increase in collagen synthesis (39). This cytokine was present in normal mouse hepatocytes, and the levels were markedly increased in schistosome-infected mice at the periphery of HG and on Kupffer cells in parenchyma (39). At this time deposition of heparan sulfate proteoglycan within the HG was prominent.

Peripheral blood fibrocytes produce type I collagen, and HG in S. japonicum-infected mice contain fibrocytes in areas where there is connective tissue matrix deposition, suggesting that these cells contribute to fibrosis (34).

There is diverse evidence that regulation of HG fibrosis is independent of the regulation of HG. Thus, the degree of HG fibrosis was unrelated to HG size in a number of S. japonicum-infected mouse strains (20). Injection of antibody to IL-4 increased HG size but decreased fibrosis in mice infected with S. japonicum (28). Splenectomy of S. mansoni-infected mice 8 weeks postinfection enhanced HG formation but did not affect fibrosis (60). Adoptive transfer of SC from mice chronically infected with S. japonicum into acutely infected mice reduced the portal pressure and HG size but did not affect fibrosis (102). Finally, TGF-β promoted fibrosis (39), but its appearance correlated with a decrease in HG size (97, 139).

DOWNREGULATION OF GRANULOMATOUS INFLAMMATION

Correlation with expression of costimulatory, adhesion, prostaglandin, and cytokine molecules.

Expression of the costimulatory molecule B7-2 was enhanced during the acute phase but was sharply diminished concurrent with a decrease in HG size (113). On the other hand, MHC-II expression by HG cells was constant throughout murine S. mansoni infection (113). HG were enlarged in mice inoculated daily with a monoclonal antibody to inducible costimulatory molecule (ICOS) for 3 weeks beginning 4 weeks postinfection (122). ICOS is associated with antigen-primed T cells, binds B7-related protein 1 (B7RP-1), and regulates the differentiation of CD4⁺ T cells (92). Injection of antibody to ICOS results in a great increase in IFN-γ production by SEA-stimulated HG, mesenteric LN, and purified CD4⁺ T cells. These observations link the ICOS-B7RP-1 pathway to downregulation and IFN-γ to upregulation of HG formation.

The dynamics of ex vivo expression of IL-2, IL-4, IL-5, and IFN-γ by SC and HG cells correlated with the formation and downmodulation of HG (78). Infection of IL-10 KO mice (Table 4) (149) and injection of antibodies to IL-10 (Table 1) (15) increased HG size. On the other hand, injection of IL-10 and IL-10:Fc (49) and IL-12 (16, 150) reduced PG size.

Administration of prostaglandin E1 inhibited PG formation (31). Macrophages from mice infected for 8 weeks constitutively produced prostaglandins (31). Injection of methyl prostaglandin E1 into infected mice resulted in general immunosuppression, including reduced HG and PG size, splenomegaly, B-cell proliferation, and IL-2 production (32).

Downregulation of the proliferative response.

Splenic T-cell proliferation and HG downregulation decreased concurrently in chronically infected mice (50, 78, 112, 152).

Role of neuropeptides.

HG eosinophils produce vasoactive intestinal peptide, and HG T lymphocytes have vasoactive intestinal peptide receptor (96). Vasoactive intestinal peptide decreased SEA-induced T proliferation and CD4⁺ T-cell-dependent IL-2 production (96). CD4⁺ T lymphocytes produce somatostatin, and somatostatin decreases IFN-γ secretion by SEA-stimulated cells (12). Somatostatin is a product of HG macrophages (143). These observations implicate neuropeptides and their receptors in granuloma downmodulation through downmodulation of T-cell expansion and of production of cytokines participating in this process.

Role of B cells, macrophages, IgG1, immune complexes, and FcγR.

The gene KO experiments (Table 4) indicated that B cells and FcγR, but not CD8⁺ T cells, were required for HG downregulation (47, 67, 68, 154). HG downmodulation by B cells and FcγR could be initiated by IC of SEA and the IgA, IgE, IgG1, and IgG2a antibodies to SEA produced by B cells in the spleen, mediastinal, mesenteric, and hepatic LN and later in the HG and mucosal intestinal tissues (1, 14, 38, 68, 75, 76, 77, 95, 128). Circulating IC are present in human schistosomiasis patients (83) and presumably in mice with the disease. In sites where IC form there are FcR-bearing macrophages, DC, B cells, basophils, mast cells, and neutrophils. The IC-FcγR reactions could inhibit HG formation by generating production of immunoinhibitory molecules, such as IL-10 and prostaglandins, by one or more of these types of cells (10, 72, 117).

Alternatively, but not exclusively, the IC-FcγR reaction could inhibit inflammation by suppressing expression of surface MHC-II (138) and IL-1 (137), molecules required for antigen presentation. Thus, IC from patients with chronic intestinal schistosomiasis inhibited expression of MHC-II (histocompatibility locus antigen DR) by B cells in vitro (117).

FcγR are present on macrophages in the HG (4). The reaction of IC with FcγR IIB on macrophages inhibits a variety of immunologically induced inflammatory responses by as-yet-undefined mechanisms (114). The engagement of these receptors on cells in the HG and other tissues might downregulate production of cytokines and chemokines that participate in HG formation and fibrosis. For instance, the homing of cells into liver to form HG might be reduced by the generation of chemokines (MIP-1α, MIP-1β, RANTES) by activation of cells through their FcγR (46). When the chemokine is bound, its receptor would be internalized and, therefore, not able to mediate chemotaxis.

Injection of the IgG1 fraction of serum from mice infected for 30 weeks with S. japonicum into acutely infected recipients reduced the HG size and portal pressure (103). This fraction suppressed SEA-induced blastogenesis of SC from acutely infected mice (50). It is not known how IgG1 caused these effects; one possibility is binding to SEA to form IC, which induce production of immunomodulatory molecules (10, 72, 117).

Roles of regulatory T cells and non-T cells.

Granulomatous inflammation was downregulated by the adoptive transfer into acutely infected animals of SC and LN cells from chronically infected mice (35). Adoptive transfer of both CD4⁺ and CD8⁺ cells reduced the HG size (29). CD8⁺ cells from mice chronically infected with S. japonicum or S. mansoni suppressed SEA-stimulated cell proliferation and migration inhibitory factor and IL-2 production by SC or HG cells from acutely infected animals (112, 129, 130, 131, 144). CD8⁺ cells also reduced the HG or PG size (103, 112).

Reduced expression of proinflammatory cytokines and increased production of immunoregulatory molecules were temporally correlated with the appearance of regulatory cells in extrahepatic and hepatic microenvironments. The T_R and other regulatory cells might reduce granuloma formation by inhibiting SEA-stimulated cell proliferation and cytokine production by SC or HG cells from acutely infected mice (50, 78, 95, 112, 129, 130, 152). Regulatory macrophages downregulated proliferation by secreting IL-10 (48). In S. mansoni-infected baboons downregulation of HG was temporally correlated with TGF-β production (97).

Regulatory cells appear in the spleen and HG when HG first appear. Thus, in vitro addition of SEA to HG cells from mice infected for 6 weeks with S. mansoni reduced IL-4, IL-5, and IL-10 production (78), and adding SEA to HG cells from mice infected for 8 weeks reduced IL-4 and IL-5 production (95). By adoptive transfer, SC that regulate HG formation were found 7 weeks postinfection (35). The appearance of regulatory cells as early as cellular and humoral immune responses (50) supports the recent provocative suggestion that generation of T_R cells may be an integral feature of the immune response (110).

In some studies it was not determined that the CD8⁺ regulatory cells were T cells (129, 130). This is an important issue because in nonschistosomal systems CD8⁺ non-T cells can induce or serve as regulatory cells. Thus, CD8⁺ CD11c⁺ lymphoid-derived DC cross-presented trinitrophenylated antigen to T_R cells, resulting in immunologic unresponsiveness (45). Plasmacytoid DC and other DC subsets rendered CD4⁺ and CD8⁺ T cells unresponsive and induced differentiation of native T cells into IL-10-producing T_R cells (80). Human CD8⁺ T_R cells were generated by CD40-activated DC which produced IL-10 (53).

The role of CD8⁺ DC in schistosomiasis is not clear. Schistosomal lysophosphatidylserine activated Toll-like receptor 2 (TLR2) and caused DC to induce the development of IL-10-producing T_R (136). However, 8 weeks after S. mansoni infection the numbers of splenic CD8α⁺ DC increased, and these cells were more activated with respect to MHC-II, C80, and CD40 than the cells in naïve mice (132). Then, depending on the nature of pathogen-derived signals and host-derived cytokines, these DC activated Th1 or Th2 cytokine responses. Therefore, CD8⁺ DC might promote rather than inhibit immune responses in schistosomiasis. Thus, lymphocytes from patients chronically infected with Schistosoma hematobium showed reduced SEA-induced proliferation and cytokine production, but addition of DC isolated from these patients overcame this hyporesponsiveness (135). In this case TLR signaling might have enhanced T_E cell responses by overcoming CD4⁺ CD25⁺ T_R-cell suppression (108).

It is not known whether regulatory cells in schistosomiasis inhibit immune responses, including proliferative and cytokine responses and granuloma formation, directly by cell-cell contact (134) or indirectly by downregulating the activity of antigen-presenting cells (18, 137) and/or by production of immunoregulatory molecules, such as IL-4, IL-10, TGF-β, and prostaglandins (15, 30, 31, 32, 65, 97, 127, 149).

Multiple extrahepatic microenvironments.

Adoptive transfer of CD4⁺ and CD8⁺ SC or LN cells from chronically infected mice to acutely infected mice downregulated HG formation. However, these regulatory cells might be activated and expanded in organs other than the spleen and LN, including mucosal tissues, and then homed into HG. Thus, when S. mansoni ova were injected into surgically fashioned cecal pouches of mice infected for 4 weeks, 4 weeks later these mice had significantly smaller HG than control mice that were not inoculated with ova or were inoculated with ova intraperitoneally or subcutaneously (142). HG size was also reduced by adoptive transfer of splenocytes from mice which were infected for 8 weeks and which had received ova intracecally, but not intraperitoneally or subcutaneously, 4 weeks earlier. Suppression of HG was abolished when, before transfer, the cells were treated with antibodies to the T-cell markers Thy-1 and CD4, but not to CD8, and complement. These CD4⁺ T cells inhibited in vitro migration of peritoneal exudate cells; i.e., they exhibited macrophage migration inhibitory activity (migration inhibitory factor) (141).

The TGF-β gene is expressed in HG (79), and TGF-β is produced by HG cells (13). In the baboon model SEA-induced elevated TGF-β production by peripheral blood mononuclear cells during chronic infection was correlated with diminished HG size, suggesting that this cytokine mediated HG downmodulation (97). Finally, TGF-β KO mice had larger PG than WT mice had (139).

Oral tolerance to a variety of antigens associated with development of T_R cells is generated by exposure of cells in the GALT to high concentrations of antigen that cross the epithelial cells from the lumen (140). Two subsets of T_R cells, Th3 and CD4⁺ CD25⁺, producing TGF-β and IL-10, respectively, were implicated in suppression of immune responses in infections with Leishmania major (8, 9), Pneumocystis carinii (61), and Bordetella pertussis (93) and in hepatitis C (89) and chronic retreoviral infections (63). In the last study the expression of surface CTLA-4 (CD152) appeared to be correlated with reduced blast cell formation. It is noteworthy that only in the study of B. pertussis (93) was it established that the T_R cells were pathogen specific. This is in contrast to evidence for CD4⁺ SEA-specific T_R cells in schistosomiasis (141).

Role of apoptosis.

Numerous large clusters of apoptotic cells appeared in the spleen and in inflammatory infiltrates around eggs in the liver 6 weeks after S. mansoni infection (40). Antibodies to IL-10 prevented TCR-induced T-cell apoptosis and enhanced TCR-stimulated secretion of Th1 cytokines. In another study 10 to 12 weeks postinfection 6% of splenic lymphocytes and 53.7% of HG lymphocytes were apoptotic (121). Apoptosis of splenic B cells and CD4⁺ and CD8⁺ T cells was observed during murine infection (88). Apoptosis of splenic CD4⁺ T cells was significantly increased 6 weeks postinfection when egg laying was initiated in the liver. This apoptosis peaked at 8 weeks, decreased by 12 weeks, and then increased at 16 weeks. Almost 30% of HG CD4⁺ T cells were apoptotic 8 and 16 weeks postinfection. SEA stimulated apoptosis of splenic and HG CD4⁺ T helper lymphocytes prepared 6 to 16 weeks postinfection. This apoptosis was mediated by FasL⁺ T and B cells, with FasL expression lowest before egg deposition (4 weeks), rising to a peak at 8 to 12 weeks (coincident with peak HG development), and then decreasing when HG were downregulated (16 weeks). Added SEA stimulated FasL expression by freshly isolated splenic B cells and CD4⁺ and CD8+T lymphocytes from mice infected for at least 6 weeks. However, SEA did not enhance FasL expression by HG T or B lymphocytes. CD4⁺ HG T lymphocytes from mice that were infected for 8 weeks exhibited significantly higher FasL expression than splenic CD4⁺ T cells. Finally, SEA-stimulated lymphocytes induced lysis of Fas-bearing target cells. CD4⁺ T-cell apoptosis was mediated by FasL-expressing B-Ia⁺ cells (87).

These studies did not establish the significance of apoptosis with regard to regulation of granulomatous inflammation. Table 5 raises some questions that deal with this issue.

TABLE 5.

Questions for further study

Subject	Key observations	Questions
Genetic control	Genetic control of HG formation and downregulation, fibrosis, epitope recognition, and Th1 and Th2 subset development	How are the processes genetically controlled? What genetic loci are involved?
Innate immunity	Schistosomal lyso-PS binds TLR2 on DC and induces IL-10-producing TR cellsa	Which schistosomal components bind TLR and induce innate immunity? Are TLRs and cells involved? What are the consequences for HG formation and downregulation?
Granuloma formation
CD4+ T cells	α/β, MHC-II dependent cells develop and expand and TCR diversify in extra hepatic microenvironments; many cells produce Th1 and Th2 cytokines coordinately but produce only low concentrations of IFN-γ; experiments with STAT4 and STAT6 KO mice concluded that Th2 cytokines, but not Th1 cytokines, are required for HG formation (done in infected mice also inoculated with ova)	Are microenvironments involved, and are GALT included? What are the roles of costimulatory, cytokine, chemokine, and receptor molecules in activation and expansion? What are the mechanisms of cell homing into granulomas? What molecular mechanisms account for coordinate Th1-Th2 cytokine production but low IFN-γ production by HG cells? What are the dynamics of TCR repertoire diversification in different microenvironments? What are the effects of STAT4 and STAT6 deletions in infected mice not inoculated with ova?
Other cells	Many other cell types are present in granulomas, including CD8+ T cells, DC, eosinophils, mast cells, and fibrocytes	What are the contributions of these cells to expression of cytokines, chemokines, neuropeptides, and their receptors? What are the dynamics of expression versus granulomas?
Neuropeptides and receptors	Substance P receptor is involved in HG formation; other neuropeptides and receptors affect cytokine production and T-cell proliferation	What is more precise information about mechanisms of HG regulation by these molecules?
Fibrosis	Controlled by variety of molecules or cells; contradictory results (e.g., IFN-γ)	What are the roles and interactions among the molecules? What is the sequence or hierarchy of activity?
Downregulation of granulomatous inflammation	Apparent coordinate downregulation of granulomatous inflammation, T-cell expansion and expression of many molecules involved in T-cell activation, expansion, and homing; evidence for roles for T cells originating in extrahepatic compartments, HPA hormones, apoptosis of CD4+ T cells mediated by FasL+ la+ B cells; B cells and FcγR requiredb	What are the roles of TR cells, macrophages, DC, and other cells? What are the roles of immunomodulatory molecules and cell-cell contact? What is the pathogen specificity of the TR cells? What are the mechanisms of hormonal involvement? What are the dynamics phenotypes, and interactions of different cell types (e.g., DC) in apoptosis and granuloma downregulation?
Balance between effector and regulatory T cells	In leishmaniasis balance between TE and TR cells appears to control parasite numbers in skin	Does a balance between SEA-reactive TE and TR cells control immunity, granuloma formation, and fibrosis in schistosomiasis?

^aLyso-PS, lyso-phosphatidylserine.

^bHPA, hypothalmic-pituitary-adrenal.

Role of retardation in development or destruction of ova.

Viable miracidia in ova secrete SEA that trigger the DTH which initiates granuloma formation. Therefore, retardation of embryonation or accelerated destruction of ova should halt SEA secretion and, consequently, granuloma formation. The evidence for retardation of embyronation is sparse and unconvincing (51, 52). Egg destruction occurs in the course of schistosomiasis in mice (19) and rhesus monkeys (25). In vitro egg destruction by cytokine-activated eosinophils requiring SEA-specific antibody (66) has been observed, but its relevance to what happens in vivo is not clear. Moreover, in the mouse the extent of ovum destruction was not correlated with the downregulation of HG: the level of egg destruction was highest 7 weeks postinfection, when the inflammation was at its height, and was reduced when the inflammation had subsided 12 to 41 weeks postinfection (148). The relationship between the extent of egg destruction and granulomatous inflammation was not examined in infected rhesus monkeys (25).

Role of hormones of the hypothalamic-pituitary-adrenal axis.

Serum levels of hypothalamic-pituitary-adrenal axis hormones, including corticotropin-releasing hormone, adrenocorticotropic hormone, dehydroepiandrosterone sulfate (DHEA-S), and cortisol, were assayed in S. mansoni-infected baboons for 12 weeks (98). During the primary infections the numbers of worms and eggs were high, and HG were large. As the infections progressed, the levels of these hormones, especially DHEA-S, decreased and the HG size decreased. In reexposed baboons the opposite pattern was observed: there were low numbers of worms and eggs, small HG, and hormone levels similar to or higher than the levels in uninfected baboons. Similar observations, including reduced levels of DHEA-S and cortisol, were made with primarily infected mice.

Further examination of the mechansims by which hormones of the hypothalamic-pituitary-adrenal axis regulate granulomatous inflammation should result in a deeper understanding of immune mechanisms. For instance, how are these mechanisms related to regulation of inflammation by T_R cells, macrophages, B cells, and FcγR and immunoregulatory molecules, including cytokines, chemokines, and prostaglandins?

DOWNREGULATION OF FIBROSIS

Experiments in which recombinant cytokines were injected (Table 3) implicated IL-7 in the downregulation of fibrosis (146). Studies of cytokine KO mice (Table 4) showed that B cells (47, 67) and Fcɛ-R-bearing cells (68) were involved in downregulation of this process. Daily intramuscular injections of IFN-γ for 4 weeks beginning 4 weeks postinfection decreased collagen deposition (39). This result contrasts with the inconsistent effects on fibrosis (1, 2, 154) of deletion of the IFN-γ and IFN-γ receptor genes (Table 4). The reasons for these discrepancies are not clear.

QUESTIONS FOR FURTHER STUDY

In this review I discuss a number of poorly understood or controversial observations and questions derived from these observations. Table 5 summarizes some of these observations and questions. Included are subjects which are in embryonic development (innate immunity); poorly understood (fibrosis, regulation by T cells, B cells, and FcγR, hormones, neuropeptides, chemokines); well developed at some levels but not at others (granuloma formation and downregulation); and speculation about how the balance between the parasite and the host is maintained.

SUMMARY

Many of the cellular and molecular elements required for development of granulomatous inflammation are present or appear soon after infection in extrahepatic tissues, such as mediastinal, mesenteric, and hepatic LN and spleens. Innate immune responses of different TLR to different schistosomal components might activate the function and/or expression of some of these elements, including antigen-presenting cells, costimulatory cytokine and chemokine molecules and their receptors, integrins on homing cells, and adhesion molecules on endothelial cells. Priming and expansion of the SEA-reactive CD4⁺ T lymphocytes and the TCR diversification required for granuloma formation are stimulated by somular antigens that are cross-reactive with SEA. Presumably, distinct patterns of coordinate chemokine expression in the liver and draining LN determine which cell types are recruited into inflammatory sites initiated by the lodging of ova in the portal circulation and other microenvironments. HG formation and granuloma formation in other sites are initiated by SEA-specific, MHC-II-dependent, α/β⁺ CD4⁺ T-lymphocyte-induced DTH. Other SEA-reactive and nonreactive T cells and other cell types are present or are also recruited to the HG and presumably other granulomas. The other cells include CD8⁺ T lymphocytes, B lymphocytes, activated eosinophils, mast cells, NK cells, basophils, macrophages, neutrophils, γ/δ⁺ cells, non-B non-T cells, and fibrocytes.

HG formation is temporally correlated with the endogenous SEA-induced production of both Th1 and Th2 cytokines in the spleen and HG. SEA epitope recognition and the development of Th1 or Th2 cytokine responses are under genetic control of the H-2 locus. The extent of HG inflammation is under non-H-2, multigene control. In response to high concentrations of endogenous SEA in the liver, there is terminal differentiation of CD4⁺ T lymphocytes marked by cell cycle-independent cytokine gene expression (i.e., cytokine secretion by HG cells without appreciable lymphocyte proliferation). Many cells secrete IFN-γ, but at about 8 weeks postinfection HG cells produce very low concentrations of this cytokine.

Downmodulation of granulomatous inflammation, including its sequela, portal hypertension, is temporally associated with reduced expression of a variety of molecules involved in and/or required for activation, expansion, T- and B-lymphocyte antigen receptor diversification, induction of DTH, and, finally, homing of many cell types to form granulomas. Downregulation appears to be initiated in extrahepatic microenvironments, such as the spleen and perhaps the GALT. Downregulation requires B cells, FcγR, and possibly immune complexes reacting with these receptors to induce production of immunoregulatory molecules. It is not known whether downmodulation is mediated by cell-cell interactions and/or production or expression by T_R cells, macrophages, DC, and other cells with immunoregulatory molecules, such as IL-4, IL-10, TGF-β, CTLA-4, and prostaglandins. Downmodulation involves complex interactions among cytokines, chemokines, prostaglandins and their receptors, and other molecules, including CTLA-4, induced by reactions of schistosomal components and/or immune complexes with SEA-reactive T and B lymphocytes, TLR, and FcγR.

Eventually fibrosis appears in the granulomas. Fibrosis is under complex control by a number of cytokines produced by several cell types, but its regulation is clearly independent of regulation of granuloma formation.

The usual lack of serious deleterious effects on the host despite the presence of many adult worms and ova in the portal and intestinal circulation for months or years could depend upon the simultaneous appearance of and equilibrium in the liver and other tissues between SEA-specific T_E and T_R cells. Presumably, the T_E cells promote granuloma formation and protect the host against the deleterious effects of schistosomal products, and the T_R cells inhibit these T_E functions.