Eta-1 (Osteopontin): An Early Component of Type-1 (Cell-Mediated) Immunity

See allHide authors and affiliations

Science  04 Feb 2000:
Vol. 287, Issue 5454, pp. 860-864
DOI: 10.1126/science.287.5454.860


Cell-mediated (type-1) immunity is necessary for immune protection against most intracellular pathogens and, when excessive, can mediate organ-specific autoimmune destruction. Mice deficient in Eta-1 (also called osteopontin) gene expression have severely impaired type-1 immunity to viral infection [herpes simplex virus–type 1 (KOS strain)] and bacterial infection (Listeria monocytogenes) and do not develop sarcoid-type granulomas. Interleukin-12 (IL-12) and interferon-γ production is diminished, and IL-10 production is increased. A phosphorylation-dependent interaction between the amino-terminal portion of Eta-1 and its integrin receptor stimulated IL-12 expression, whereas a phosphorylation-independent interaction with CD44 inhibited IL-10 expression. These findings identify Eta-1 as a key cytokine that sets the stage for efficient type-1 immune responses through differential regulation of macrophage IL-12 and IL-10 cytokine expression.

The development of cell-mediated (type-1) immune responses is necessary for protection against the growth of many infectious pathogens and, when excessive, can mediate autoimmune host tissue destruction. Although macrophage activation by microbial pathogens (1, 2) and foreign body reactions (3) are associated with type-1 immunity, the cellular and molecular events that imprint this response are not fully understood. An essential early step in this process is macrophage production of IL-12 at sites of infection, whereas early IL-10 production inhibits this response (4). Although IL-12 responses can be triggered by an interaction between the CD40 ligand on activated T cells and CD40 on macrophages (4), this interaction also induces the inhibitory IL-10 cytokine (5, 6), and its transient nature may not suffice for sustained IL-12 induction in vitro (7) or in vivo (8).

A gene product that may play an important role in the development of type-1 immunity is the T cell cytokine Eta-1 (for early T lymphocyte activation–1), also known as osteopontin (Opn) (9). The Eta-1 gene is expressed in T cells early in the course of bacterial infections (within 48 hours), and interaction of its protein product with macrophages can induce inflammatory responses (10). Genetic resistance to infection by certain strains ofRickettsia may depend on Eta-1–dependent attraction of monocytes into infectious sites and acquisition of bacteriocidal activity (11); the granulomatous responses characteristic of sarcoidosis and tuberculosis are associated with high levels of Eta-1 expression (12).

Granuloma formation in these human diseases is a cellular consequence of type-1 immunity (12), and sarcoid-type granulomas can be induced in mice after injection of polyvinyl pyrrolidone (PVP) (13). Because certain murine models of parasite-induced granulomas may reflect a mixture of type-2 and type-1 immunity (6), we first established the importance of IL-12–dependent type-1 immunity in this murine model of granuloma formation. An intense granulomatous response was provoked shortly after (subcutaneous) injection of PVP into C57BL/6 (+/+) but not C57BL/6 nu/nu strains of mice. This response was diminished by 70 to 80% in C57BL/6 IL-12−/− mice and was enhanced two- to threefold in C57BL/6 IL-10−/− mice (Fig. 1, A and B). Because C57BL/6nu/nu mice coinjected with PVP and purified Eta-1 displayed a granulomatous reaction, this gene product can partially substitute for activated T lymphocytes in this setting (Fig. 1, A and B).

Figure 1

Granuloma formation in cytokine-deficient mice. Granulomas were measured 5 days after subcutaneous injection of PVP into mice carrying the indicated mutations of cytokine genes (38). (A) Results are expressed as the mean number of granulomas per high-power field (HPF) (×200 magnification) (left); as the mean number of cells per granuloma after examination of 50 to 80 HPF per mouse (middle); and as the product of these two indices, shown as granuloma burden (right). Error bars indicate 1 SEM. (B) Photomicrographs at indicated magnifications showing histologic analysis of tissue sections at PVP injection sites of Eta-1+/+ (left) or Eta-1−/− (right) mice showing granulomatous infiltrations of mononuclear cells in subcutaneous dermal and subdermal areas 5 days after injection of PVP, PBS, or PVP + 5 μg purified Eta-1 (40). (C) Analysis of surface antigens expressed by cells within granulomas in the indicated mouse strains was done with monoclonal antibodies to Mac-1, B220, CD3, and BP-55 (a neutrophil marker). Error bars indicate 1 SEM. (D) Cytokine expression by cells from lymph nodes draining the site of granulomas 5 days after PVP injection was measured 48 hours after incubation with PVP (2 × 106 cells per well). Data are representative of three independent experiments.

We then asked whether mice deficient in Eta-1 secondary to targeted gene mutation (14) formed granulomas after PVP injection. Eta-1−/− mice did not develop a detectable granulomatous response after challenge with PVP; the response was partially restored by coinjection of purified Eta-1 with PVP (Fig. 1, A and B). Histologic analysis of granulomas formed in Eta-1+/+ mice and in Eta-1−/− mice reconstituted with purified Eta-1 revealed a similar macrophage-dominant cellular infiltrate: About 85% of granulomatous cells in both cases were Mac-1+, whereas 5 to 10% were CD3+ T cells or B220+ B cells. BP-55+ neutrophils, which were only a minor component (1 to 2%) of granulomas in these mice, increased 5- to 10-fold in the granulomas of IL-10−/− mice (Fig. 1C). Eta-1−/− mice also displayed defective granulomatous responses to injection of collagen and latex, consistent with reports that human T cells resident in sterile granulomas have high expression of Eta-1 (12). Restimulation of lymph nodes draining subcutaneous sites of PVP injection in Eta-1−/− mice and control mice with PVP revealed impaired IL-12 and interferon-γ (IFN-γ) responses: The IL-12 response was reduced by ∼95%, and the IFN-γ response of Eta-1−/− mice was reduced by 90% in comparison to Eta-1+/+ controls (Fig. 1D).

We next defined the role of Eta-1 in the immune response to herpes simplex virus–type 1 (HSV-1) (KOS strain) infection. Eta-1−/− mice infected by HSV-1 [4 × 106 plaque-forming units (PFU) via the cornea] did not develop a significant tuberculin-type delayed-type hypersensitivity (DTH) response after footpad challenge with HSV-1 (105PFU), in contrast to the strong DTH response of Eta-1+/+controls (Fig. 2A, left). Although the numbers of T cells and proportions of T cell subsets in the thymus and peripheral lymphoid tissues of Eta-1−/− mice were similar to Eta-1+/+ littermates (15), defective antiviral DTH responses might reflect a subtle alteration in lymphocyte or macrophage development. We therefore tested the effects of acute in vivo depletion of Eta-1 with a neutralizing antibody. Administration of antibody to Eta-1 (LF-123) (16) immediately before and repeatedly after HSV-1 infection efficiently inhibited the DTH response upon rechallenge (Fig. 2A, right).

Figure 2

The role of Eta-1 in immunity to HSV-1 (KOS strain). (A) (Left) Defective HSV-1–specific DTH responses in Eta-1−/− mice. C57BL/6 × 129 strain mice with or without a targeted disruption of the Eta-1 gene (–/–) or controls (+/+) were infected in the right eye with 4 × 106 PFU of HSV-1 (KOS) and challenged 5 days later in the left footpad with 1 × 105 PFU of UV-inactivated HSV-1 (KOS). The right (control) and left (HSV-1) footpads of each mouse were measured 24 hours later with a micrometer. Each data point represents the mean and standard error (error bars) of three mice per group. (Right) Inhibition of the anti–HSV-1 DTH response by acute depletion of Eta-1. The neutralizing antisera LF-123 (16) or control normal rabbit serum were injected at 25 μg per dose per day, starting 2 days before injection. On day 0, mice were infected with HSV-1 (KOS) and rechallenged 5 days later. The right and left footpads of each mouse were measured 24 hours after rechallenge, and specific swelling (left versus right footpad) is shown. (B) Development of HSK in Eta-1−/− mice. The right eyes of Eta-1−/− and Eta-1+/+ mice were infected with 4 × 106 PFU of HSV-1 (KOS), and disease was assessed on days 11 and 14 after infection, as described (17). The severity of clinical stromal keratitis was scored on the basis of the percentage of corneal opacity: ≤25%, 1; ≤50%, 2; ≤75%, 3; and 75 to 100%, 4. Each point represents at least 16 mice and is the mean of three independent experiments. (C) Differential cytokine profile of draining lymph node cells and splenic macrophages from Eta-1+/+ or Eta-1−/− mice after infection with HSV-1. Cytokine levels after restimulation of draining lymph node cells (from mice 15 days after HSV-1 infection in vivo) by 4 × 107 PFU of UV-inactivated HSV-1 using 48-hour supernatants were determined by ELISA (19). Viral restimulation of mixtures of purified lymph node T cells from virus-infected donors and syngeneic (nonimmune) adherent cells yielded less than one-third of the IL-10 response of mixtures of immune T cells and macrophages from draining lymph nodes of infected donors (20). The proliferative response of lymph node cells from HSV-1–infected Eta-1+/+ and Eta-1−/− mice measured by [3H]thymidine incorporation at 72 hours was 20.9 × 103 and 18.7 × 103 cpm, respectively.

Corneal HSV-1 infection can also lead to a destructive type-1 autoimmune inflammatory reaction, herpes simplex keratitis (HSK), initiated by CD4 cells that recognize a viral peptide mimic (17). This inflammatory response depends on the production of IL-12 and is inhibited by IL-10 (18). Within 10 to 14 days after corneal HSV-1 infection, ∼65% of control Eta-1+/+ mice developed HSK, whereas HSV-1–infected Eta-1−/− mice did not readily develop this disease (Fig. 2B). Analysis of cells from the draining lymph nodes of virus-infected Eta-1−/− and Eta-1+/+ mice indicated that they responded similarly to HSV-1 according to [3H]thymidine incorporation after viral restimulation in vitro (19). However, draining lymph node cells from virally infected Eta-1−/− mice produced exaggerated amounts of IL-10 and IL-4 and reduced IL-12 in comparison with Eta-1+/+ controls (Fig. 2C). In contrast with this sterile granulomatous response, IFN-γ levels were not reduced in Eta-1−/− mice after HSV-1 viral infection (20), consistent with an IL-12–independent pathway to IFN-γ production that may depend on virally induced IFN-α/β production (21).

We then investigated the ability of Eta-1−/− mice to mount a protective immune response after bacterial infection. The murine response to Listeria monocytogenes is an experimental cornerstone of our understanding of the early events leading to type-1 immunity after microbial infection (1) and depends on early macrophage production of IL-12 and downstream expression of IFN-γ (22). Eta-1−/− mice were defective in their ability to clear L. monocytogenes after systemic infection, similar to the defect in IL-12−/− mice (23) (see Web figure 1, available Restimulation of spleen cells from Eta-1−/− and Eta-1+/+mice with heat-killed bacteria revealed that cells from the former mice had reduced IFN-γ responses: 25.5 ± 6.5 ng/ml of IFN-γ were produced by spleen cells from Eta-1+/+ mice in comparison with 3.2 ± 1.2 ng/ml of IFN-γ from Eta-1−/− mice (24).

Thus, Eta-1 expression may affect type-1 immunity through regulation of the IL-12 and IL-10 cytokine ratio. To define the effect of Eta-1 on IL-10 and IL-12 production by macrophages in vitro, we incubated resident peritoneal macrophages with increasing concentrations of purified Eta-1 in serum-free medium (Fig. 3A). This resulted in the secretion of as much as 400 pg/ml of IL-12 at 48 hours, whereas IL-10 production was not detected (Fig. 3, A and B). The failure of Eta-1 to induce IL-10 was somewhat surprising because other cytokines that activate macrophages, including tumor necrosis factor–α, IL-1, IL-2, IL-3, and IL-6, all stimulate IL-10 secretion (25), and lipopolysaccharide (LPS) stimulation of these resident peritoneal macrophages induced both IL-12 (∼250 pg/ml) and IL-10 (∼100 pg/ml) (Fig. 3B). Further analysis showed that Eta-1 actively suppressed the LPS-dependent IL-10 response of resident peritoneal macrophages by 80 to 95% (Fig. 3C).

Figure 3

Differential regulation of macrophage IL-12 and IL-10 responses by purified Eta-1. (A) Dose-dependent induction of IL-12 (open circles), but not IL-10 (solid circles) production, from macrophages by Eta-1. Resident peritoneal macrophages obtained from C57BL/6 mice (41) were incubated for 48 hours (5 × 105 macrophages per milliliter) with purified Eta-1 (40), and IL-10 and IL-12 p70 concentrations in the supernatant were determined by ELISA. Assays were done in quadruplets, and each point represents the mean and standard error (error bars) of three independent experiments. (B) Time course of IL-12 (open circles) p70 and IL-10 (solid circles) expression by resident peritoneal macrophages (5 × 105 macrophages per milliliter) after incubation with 5 nM Eta-1 or LPS (30 ng/ml). Assays were performed in quadruplets, and each data point represents the mean and standard error (error bars) of two independent experiments. (C) Inhibitory effect of Eta-1 on macrophage IL-10 production. Macrophages were activated with LPS (30 ng/ml) for 1 hour before addition of Eta-1 (5 nM) for an additional 48 hours and consecutive measurement of IL-12 and IL-10 by ELISA. Assays were performed in quadruplets, and each point represents the mean and standard error (error bars) of two independent experiments.

The interaction of Eta-1 with macrophages is mediated through two functional receptors. Engagement of CD44 mediates chemotactic migration (26), and interaction with αVβ3 integrin causes haptotaxis, adhesion, and spreading (10, 27). We asked which receptor was responsible for the regulation of macrophage cytokine production by Eta-1. Induction of IL-12 is inhibited by GRGDS (28) peptide (but not GRADS peptide) (29) and by antibody to the integrin β3 subunit (but not by antibody to CD44), and macrophages from CD44−/− mice (30) display an unimpaired IL-12 response (Fig. 4, A to C). Moreover, Eta-1–dependent induction of IL-12 secretion from macrophages was not due to contamination with endotoxin: Limulus lysate analysis indicated that purified Eta-1 contained <1 ng/g of endotoxin, and the IL-12 response of macrophages from C3H.HeJ mice (which are defective in endotoxin receptor-mediated signaling) was not obviously impaired in comparison to other strains (Fig. 4C). In contrast to IL-12 induction, inhibition of IL-10 depends on engagement of the CD44 receptor: Eta-1–dependent inhibition of IL-10 is blocked by antibody to CD44 but not by antibody to integrin β3, and macrophages from CD44−/− mice are resistant to Eta-1 inhibition of the IL-10 response (Fig. 4, A to C). To further characterize the RGD-dependent interaction with the macrophage integrin receptor, we analyzed fragments from an Eta-1 Lys-C digest and identified a proteolytic fragment from the NH2-terminal portion of Eta-1, which contains the integrin binding site (termed NK10) that is sufficient to induce macrophage IL-12 expression (Fig. 4A).

Figure 4

Regulation of macrophage IL-12 and IL-10 expression by distinct Eta-1 receptors. (A) Secretion of IL-12 by macrophages is mediated by a 10-kD peptide (NK10) derived from the NH2-terminal fragment of Eta-1 (42) and is inhibited by a blocking antibody to integrin β3 (1 μg/ml), but is unaffected by antibody to CD44 (1 μg/ml). Error bars indicate 1 SEM. (B) Induction of IL-10 production by IL-4 (500 U/ml) in the presence or absence of purified Eta-1 (5 nM) and the effects of antibodies to CD44 [Km81 purified from TIB241 (26)] and the β3 integrin (2C9.G2, PharMingen) are shown. Error bars indicate 1 SEM. (C) Production of IL-12 and IL-10 in response to Eta-1 by peritoneal macrophages from C57BL/6 mice, C57BL/6 mice that are deficient in CD44 gene expression (C57BL/6–CD44−/−), and cells from C3H.HeJ mice. Mean values and standard errors (error bars) from at least four data points are shown.

Eta-1 is secreted in nonphosphorylated and phosphorylated forms (31). Phosphorylation may allow Eta-1 to associate with the cell surface rather than the extracellular matrix (32), through a contribution to integrin binding. In contrast, serine phosphorylation of recombinant Eta-1 is not required for CD44-dependent interactions leading to chemotactic migration (26). We investigated whether phosphorylation of Eta-1 might affect its ability to regulate IL-12 and IL-10 expression. Dephosphorylation of purified, naturally produced Eta-1 abolished IL-12 stimulatory activity; phosphorylation of recombinant Eta-1 at specific sites restored activity (33) (see Web figure 2, available Although recombinant Eta-1–lacking phosphate groups could not induce IL-12, this molecule retained inhibitory activity for the macrophage IL-10 response (33). Thus, serine phosphorylation can provide molecular information that regulates the biological activity of a secreted protein.

Our data indicate that expression of Eta-1 represents an essential early step in the pathway that leads to type-1 immunity. Previous studies have established the importance of macrophage production of IL-12 in this pathway (1, 4, 22). Our experiments suggest that production of Eta-1 by activated T cells is an essential proximal event that potentiates the macrophage IL-12 response through integrin engagement and dampens the IL-10 response through CD44 engagement, leading to up-regulation of type-1 cytokines. The latter inhibitory effect on IL-10 may account for enhanced granulomatous responses of CD44−/− mice (30) and the finding that impairment of the granuloma response noted here is somewhat greater than might be anticipated from the response of IL-12–deficient mice (Fig. 1). These findings fill a logical gap in our understanding of the early molecular events that lead to type-1 immunity. Although down-regulation of CD40 ligand expression by IFN-γ and soluble CD40 occurs within 24 hours after viral infection, IL-12 is detected in serum over the next 7 to 10 days (8). Our experiments suggest that replacement of the CD40L signal by Eta-1 may potentiate the IL-12 response while dampening the IL-10 activity to allow full maturation of type-1 immunity, as judged by cellular responses and expression of downstream effector cytokines such as IFN-γ. The ability of an antigen to induce Eta-1 production after T cell receptor ligation may thus determine the ensuing duration and intensity of type-1 immune responses. Eta-1 imprinting of the IL-12 and 1L-10 response after appropriate peptide stimulation (34) may also increase the likelihood of autoimmune sequelae as shown here (Fig. 2), through pathways that do not invariably require IFN-γ (35).

Eta-1–dependent regulation of two early cytokine checkpoints that dictate development of type-1 or type-2 immunity also suggests new therapeutic approaches to several diseases. Eta-1 analogs that mediate CD44-dependent inhibition of IL-10 may inhibit sepsis in burn patients (36), and Eta-1 antagonists may ameliorate the clinical course of bacterial arthritis (37). Engineered forms of Eta-1 that imprint type-1 responses after immunization may also be valuable components of viral and cancer vaccines.

  • * These authors contributed equally to this work.

  • Present address: Division of Radiation and Cancer Biology, New England Medical Center, 750 Washington Street, Boston, MA 02111, USA.

  • To whom correspondence should be addressed. E-mail: Harvey_Cantor{at}


Stay Connected to Science

Navigate This Article