Viral IL-6 Blocks Neutrophil Infiltration during Acute Inflammation¹

Many viruses have evolved to circumvent immunological defense mechanisms by using virally encoded factors that mimic or suppress the action of immunomodulatory proteins. One such example is human herpesvirus-8 (HHV8),³ which expresses a viral equivalent of IL-6 (vIL-6) that plays a substantive role in HHV-8-associated pathology (1). Although originally isolated from tumor biopsies of AIDS-associated Kaposi’s sarcoma (2), HHV8 has now been implicated in other AIDS-related conditions such as the B cell lymphoma, primary effusion lymphoma (PEL), and the plasmacytic lymphoproliferative disease, multicentric Castleman’s disease (3). Several studies have documented a close association between these HHV8-related conditions and human IL-6 signaling (4, 5). However, expression of vIL-6 by HHV8-infected PEL cells appears to be critical for the cellular expansion of this B cell lymphoma (6). Indeed, vIL-6 seems not only to be required for the cellular proliferation, but also for rescuing cells from entering apoptosis and for immune evasion of IFN-α activity (7).

Viral IL-6 shares ∼25% homology with its human counterpart (1); however, the autocrine activation of PEL cells cannot be fully substituted by the action of its human counterpart (6). Thus, the mode of vIL-6 activation appears distinct from that of human IL-6. IL-6 responses are mediated through specific engagement of a cognate IL-6R and activation of gp130, which is the universal signal-transducing receptor for all IL-6-related cytokines (8). Although gp130 is ubiquitously expressed within the body, IL-6R expression is restricted to hepatocytes and subpopulations of leukocytes (8). IL-6 responses are however also transmitted through a mechanism known as IL-6 trans-signaling, which adopts the soluble IL-6R (sIL-6R) as a surrogate receptor for the IL-6 activation of gp130 (9). Viral IL-6 elicits cellular responses through a mechanism that is independent of the cognate IL-6R and therefore employs a method of activation that resembles IL-6 trans-signaling (10, 11, 12).

Virally encoded immunomodulators have the potential to not only orchestrate tumor progression, but to also influence the inherent control of immunological processes that may be triggered either as a consequence of the viral infection or in response to a secondary infection. Appropriate control of an inflammatory response ultimately relies on the tight control of leukocyte recruitment, activation, and clearance to ensure successful resolution of the condition. IL-6 trans-signaling is actively involved in each of these stages and governs transition from neutrophil to mononuclear cell recruitment (13), as well as influences aspects of both innate and acquired immunity (9). Consequently, vIL-6 might also act in a similar fashion, which due to the constitutive nature of its expression may affect the initial stages of inflammation. Inefficient regulation of local innate immune responses may for instance increase an individual’s susceptibility to opportunistic infections and worsen the prognosis of HHV8-infected patients. Since IL-6 trans-signaling orchestrates the effective clearance of neutrophils from sites of infection (13), studies have examined the ability of vIL-6 to regulate local chemokine levels and the inflammatory influx of neutrophils during activation of an acute inflammatory episode. These findings represent the first in vivo demonstration that vIL-6 may significantly influence immunological responses.

Recombinant IL-6 and soluble IL-6 receptor (sIL-6R) were purchased from R&D Systems. Recombinant vIL-6 was produced in the EBNA1-293 kidney epithelial cell line (Invitrogen Life Technologies), stably transfected with an episomal vector encoding vIL-6 as a His-tag fusion protein (pCEP-Pu-flag-vIL-6-his). Viral IL-6 was purified on a freshly prepared Ni-NTA-agarose column following elution using 0.5 M imidazole.

HPMC were isolated by tryptic digestion of omental tissue obtained from consenting patients undergoing elective abdominal surgery and cultured as previously described (14). Before experimentation, HPMC monolayers were growth arrested for 48 h in the absence of serum. All stimulations were performed in the absence of serum on isolates no older than the second passage. Culture supernatants were rendered cell free and stored at −70°C. Stimulation of HPMC with varying doses of LPS (0–500 ng/ml, in the presence of 0.1% FCS) showed no induction of CXCL8, CCL2, and IL-6 and is consistent with the TLR4⁻CD14⁻ phenotype of HPMC (data not shown).

Inflammatory mediator concentrations were quantified using sandwich ELISA techniques. Human CCL2 (MCP-1) and CXCL8 (IL-8) were quantified using matched Ab pair OptEIA ELISA sets (BD Pharmingen). Human IL-6 was measured using a matched Ab pair (Mab206 and Baf206) from R&D Systems. Human vascular endothelial growth factor (VEGF) and the murine chemokine KC were analyzed using Duoset ELISA kits (R&D Systems).

Cytosolic and nuclear extracts were prepared from HPMC using a rapid technique for the extraction of nuclear proteins (15). Briefly, cells were harvested in ice-cold PBS and pelleted by centrifugation. Cells were resuspended in ice-cold buffer A containing 0.5 mM DTT, 0.2 mM PMSF, 1 mM sodium orthovanadate, 50 mM sodium fluoride, and proteinase inhibitors (diluted 1/1000; Sigma-Aldrich) and incubated on ice for 15 min. The nuclei were pelleted by centrifugation and the cytosolic fraction was removed. The nuclear pellet was resuspended in ice-cold buffer C containing 0.5 mM DTT, 0.2 mM PMSF, 1 mM sodium orthovanadate, 50 mM sodium fluoride, and proteinase inhibitors (diluted 1/1000; Sigma-Aldrich) and incubated on ice for 20 min to allow for high salt extraction. Cellular debris was removed by brief high-speed centrifugation, and the resulting supernatants (nuclear extract) were collected. Protein concentrations were determined using the Bradford method.

EMSAs were performed using 3 μg of nuclear extract as previously described (16). Oligonucleotides containing the SIE STAT-binding IFN-γ activation site elements were annealed and labeled with [α³²P]dTTP (Amersham Pharmacia) using the Klenow fragment of DNA polymerase I. The composition of STAT-DNA binding complexes was determined by supershift assays using rabbit polyclonal Abs specific to STAT-1 and STAT-3 (Santa Cruz Biotechnology). Densitometry analysis was performed using the Bio-Rad gel documentation system and QuantityOne software (Bio-Rad). Detected STAT DNA binding was expressed relative to control levels (equal to 1).

Cytosolic proteins (8 μg) were separated by SDS-PAGE on a 10% polyacrylamide gel and transferred to Hybond ECL nitrocellulose membrane (Amersham Biosciences). Membranes were probed with rabbit polyclonal Abs raised against phosphorylated ERK1/2 (p42/p44 MAPK Thr²⁰²/Tyr²⁰⁴) and subsequently stripped (62.5 mM Tris-Cl (pH 6.8), 2% SDS, 100 mM 2-ME) at 55°C for 30 min before being reprobed with an Ab against total ERK1/2/p42/p44 MAPK (both from Cell Signaling Technology). Both Abs were diluted 1/1000 in 5% BSA/TBS with 0.2% Tween 20, and immunolabeling was visualized by chemiluminescence (ECL detection reagents; Amersham Biosciences). Densitometry analysis was performed using the Bio-Rad gel documentation system and QuantityOne software. Phosphorylated ERK1/2 was normalized against the staining intensity of total ERK1/2 and expressed relative to control levels (equal to 1).

Experiments were performed in 8- to 12-wk-old in-bred C57BL/6J IL-6-deficient (IL-6^−/−) mice (17). Procedures were conducted following Home Office approval under project license number PPL-40/2131.

A lyophilized cell-free supernatant Staphylococcus epidermidis-2 derived cell-free supernatant (SES) was prepared from a clinical isolate of Staphylococcus epidermidis as described previously (18). Peritoneal inflammation was established in IL-6^−/− mice by i.p. injection of a defined 500-μl dose of SES. Control populations were administered with sterile PBS. At designated time intervals, mice were sacrificed and the peritoneal cavity was lavaged with 2 ml of ice-cold PBS. Leukocyte numbers were assessed by differential cell counts and inflammatory mediators were quantified by ELISA. Viral IL-6 was administered as indicated in the figure legends.

To address the potential impact of vIL-6 during the course of an inflammatory episode, studies used a peritoneal model of acute inflammation. PEL is a B cell lymphoma, which develops in pleural and peritoneal body cavities. Recurrent effusions of these body cavities are frequently observed in HHV8-associated conditions, and atypical HHV8-infected mesothelial cells have been identified in HIV⁺ individuals with pleural effusions (19, 20). Since PEL effusions also contain high levels of vIL-6 (∼70 ng/ml) (21), in vitro studies were performed on primary human mesothelial cells, which line pleural and peritoneal cavities. Primary peritoneal mesothelial cells express undetectable levels of IL-6R, but high levels of gp130 (16, 22). They are therefore unresponsive to IL-6 alone and require the addition of sIL-6R to activate gp130 signaling (13). Stimulation of mesothelial cells with vIL-6 resulted in the activation of STAT3 (Fig. 1, A and B). The activation of STAT was specifically blocked by an anti-gp130 mAb, suggesting vIL-6-mediated responses are directly elicited through gp130 (data not shown). Membrane-bound gp130 expression was also transiently down-regulated by IL-6 plus sIL-6R and to a lesser extent vIL-6. However, surface levels returned to baseline by 24 h after stimulation (data not shown). Stimulation with vIL-6 did not however regulate STAT1 (Fig. 1,B), and this is in general agreement with other reports of gp130 activation in mesothelial cells (16). Activation of STAT3 also coincided with the phosphorylation of ERK1/2, and maximal activation of both proteins was observed 30 min after stimulation with vIL-6 (Fig. 1, C and D). Regulation of these signaling molecules showed that vIL-6 was less potent than IL-6 and sIL-6R (Fig. 2). Indeed, vIL-6 appeared to preferentially promote STAT3 activity with physiologically relevant doses of vIL-6 inducing a 3–6 change in STAT3 activation (Fig. 2,A). In contrast, only the highest dose of vIL-6 (500 ng/ml) had any appreciable impact on ERK1/2 phosphorylation (Fig. 2 B).

Viral IL-6 has previously been shown to regulate inflammatory mediators (11, 23, 24, 25). To establish the consequence of mesothelial cell activation by vIL-6, initial studies examined changes in the expression of human IL-6 and VEGF. These factors are known to be induced by vIL-6 in cell lines and are linked to the pathology of HHV8-associated conditions (4, 5, 23, 24, 25, 26, 27). Viral IL-6 promoted a 4-fold increase in the secretion of both mediators (Table I). In each case, the threshold of activation (50 ng/ml) related to the dose of vIL-6 required to elicit STAT3 activation (data not shown).

Transition from innate to acquired immune responses is defined as a switch from an initial infiltration of neutrophils to a more sustained population of mononuclear cells. This step represents a critical event in the resolution of an inflammatory episode and is governed by IL-6 trans-signaling (13). Mesothelial control of human IL-6 secretion by vIL-6 will ultimately influence the balance of this sIL-6R-regulated response; however, the ability of vIL-6 to directly activate gp130-mediated activities implies that this virokine may be able to circumvent the inherent control of this immunological switch. In vitro studies therefore examined the ability of vIL-6 to regulate mesothelial chemokine expression (Fig. 3).

Consistent with the activities of IL-6 and its soluble receptor, vIL-6 triggered a dose-dependent increase in the expression of CCL2, but not that of CXCL8 (Fig. 3,A). However, IL-6 trans-signaling is known to suppress CXCL8 production driven by proinflammatory cytokines (13, 16). In accordance with this activity, vIL-6 was found to suppress IL-1β-mediated release of CXCL8 (Fig. 3,B). Combined stimulation of mesothelial cells with IL-1β and vIL-6 however resulted in an accumulative increase in CCL2 production (Table II). In all cases, the potency and efficacy of vIL-6 was deemed to be lower than that of IL-6 and its soluble receptor.

To equate these in vitro observations to the potential regulation of inflammation by vIL-6, studies adopted a peritoneal model of inflammation that closely resembles the clinical alterations associated with acute bacterial peritonitis (13). Utilization of an IL-6-deficient background ensured that IL-6 trans-signaling did not influence the experimental setup and enabled reconstitution of IL-6 activity to be solely achieved via i.p. administration of vIL-6. As shown in Fig. 4, i.p. administration of a cell-free supernatant derived from a clinical isolate of S. epidermidis (termed SES) promoted an early, but transient increase in neutrophil recruitment. IL-6 deficiency does not affect the initial rate of neutrophil recruitment, but results in a heightened neutrophil infiltration and a delay in their clearance as compared with wild-type mice (13, 28). Intraperitoneal administration of SES in combination with vIL-6 significantly blocked both neutrophil recruitment and peritoneal expression of the neutrophil-activating chemokine KC (Fig. 4). Blockade of neutrophil infiltration by vIL-6 may therefore impede innate immune responses.

By encoding a variety of immunomodulatory agents that efficiently bypass the host’s ability to control immunological responses, viruses have adapted sophisticated mechanisms for immune evasion (29). These strategies appear fundamentally designed to advance the cause of the virus; however, studies presented in this current communication show that HHV8-encoded vIL-6 has the capacity to influence the course of an acute inflammatory response. Viral IL-6 plays a prominent role in progression of HHV8-associated pathology (6, 7); however, by mimicking IL-6 trans-signaling as a mechanism of IL-6 activation, vIL-6 was found to modify the chemokine-directed recruitment of leukocytes during acute inflammation.

HHV8-associated primary effusion lymphomas typically invade pleural and peritoneal body cavities, and development of these conditions are associated with high vIL-6 levels that may affect tumor expansion, cellular apoptosis, and IFN-α activity (6, 7, 21). Within this context, studies examined the direct action of vIL-6 on human peritoneal mesothelial cells. Mesothelial cells are the principle peritoneal lining cells and are considered the major producers of inflammatory mediators within the peritoneal cavity (30). Although mesothelial cells lack a functional cognate IL-6R, they express gp130 and during recurrent HHV8-associated effusions are likely to be exposed to vIL-6.

Stimulation of mesothelial cells with vIL-6 directly triggered gp130-mediated activation of STAT3, but not STAT1. This response is analogous to the mesothelial response elicited by IL-6 and its soluble receptor, suggesting that vIL-6 adopts a trans-signaling-like mechanism of cellular activation. The threshold of STAT3 activation by vIL-6 (>50 ng/ml) was however higher than that observed for IL-6 and sIL-6R. This difference in efficacy and potency might relate to the low affinity of vIL-6 for gp130 (K_d 2500 nM) compared with that of IL-6/sIL-6R (K_d 1.7 nM) (31, 32). Although this affinity was determined using a bacterially expressed form of vIL-6 that lacks the N-linked glycosylation required for optimal activity (33), the low affinity of vIL-6 for gp130 might be critical in maintaining a threshold between the beneficial consequences of its action (e.g., lymphoma proliferation) and the more detrimental outcomes associated with disrupting the host’s immune response. This notion may relate to the poor activation of ERK1/2 by vIL-6, which showed only a limited degree of activation at the highest vIL-6 concentration (500 ng/ml) tested. In this respect, genetic approaches involving knock-in mice that exhibit altered gp130-mediated signaling capacity have shown that the quality of the signal relayed via STAT1/3 and ERK1/2 effects the nature of the cellular response (34, 35). Collectively, these studies suggest that vIL-6 predominantly signals via gp130-mediated STAT3 activation. Constitutive STAT3 activity is widely associated with tumor development, where it accounts for increased cellular proliferation and resistance to apoptosis (36). Although vIL-6 might therefore contribute to STAT3 activity in HHV8-associated tumors, STAT3 signaling within resident tissue cells may also affect events known to influence the immunological response (37).

To test whether this is a likely scenario, studies specifically focused on the vIL-6 regulation of leukocyte recruitment, which is modified by the activity of IL-6 and its soluble receptor (13, 16). Specifically, IL-6 trans-signaling promotes transition from an initial influx of neutrophils to a more sustained population of mononuclear cells (13). Regulation of this switch is orchestrated by differential control of inflammatory chemokine expression and appropriate activation of cellular apoptosis (13). Consistent with this inflammatory response, vIL-6 was found to augment mesothelial-derived CCL2 secretion and to inhibit the IL-1β-induced expression of CXCL8 by HPMC in vitro. Suppression of CXCL8 was also evident in vivo, where vIL-6 was found to dose-dependently block secretion of the murine CXC chemokine KC following inflammatory activation. KC is a prominent neutrophil chemoattractant in vivo (16), and a significant reduction in neutrophil recruitment was observed when vIL-6 was given in combination with SES. Elevated levels of vIL-6 may therefore lead to decreased neutrophil recruitment and impaired innate immune function.

Although monocyte/macrophage-rich body cavity effusions are frequently seen in individuals with HHV8-associated conditions (19) and vIL-6 was found to promote CCL2 expression in vitro, vIL-6 did not dramatically influence the number of mononuclear cells recruited in response to SES (data not shown). Active recruitment of monocytes and B cell may be advantageous to the persistence of HHV-8 by allowing de novo infection of new cell populations. Viral regulation of leukocyte recruitment is also an efficient immune evasion strategy used by many viruses (38). HHV-8 encodes several chemokine homologues, which direct T cell migration and preferentially recruit Th2-polarized cells instead of Th1 (39). In this respect, CCL2 has been implicated in directing Th2 polarization (40), and vIL-6-induced CCL2 may ultimately influence the nature of the T cell response associated with HHV8 infections. Future studies will therefore need to consider the role of vIL-6 in terms of mononuclear cell recruitment and activation. This may be particularly pertinent to HHV8 persistence and immune evasion.

An immunocompromised state, where viral replication proceeds unchecked, is required for the development of HHV-8-associated disease. Our data show that vIL-6 has the potential to influence inflammatory events at clinically relevant concentrations (21). Indeed, vIL-6 is able to direct IL-6-like responses in resident tissue cells lacking a functional cognate IL-6R. Through differential control of inflammatory cytokines, growth factors, and chemokines, vIL-6 may alter the eventual outcome of an inflammatory episode by affecting the inherent control of leukocyte recruitment. Disruption of innate immune responses by vIL-6 may ultimately contribute to the development of opportunistic infections in HHV-8-associated disease.

The authors have no financial conflict of interest.