Exaggerated Human Monocyte IL-10 Concomitant to Minimal TNF-α Induction by Heat-Shock Protein 27 (Hsp27) Suggests Hsp27 Is Primarily an Antiinflammatory Stimulus¹

Systemic inflammatory responses, as well as exaggerated local inflammatory cytokine production, have been implicated in mediating multiple organ failure and rheumatoid arthritis (1, 2). During shock inflammatory stress, heat shock proteins (hsp),³ which are stress response proteins found in all species, are up-regulated (3, 4, 5). These hsp are thought to play a pivotal role in protecting cells during stress and inflammatory responses (3, 4, 5, 6). Recently, the large hsp have also been suggested as danger signals that first activate monokine production, then stimulate and/or regulate the magnitude of the immune response (7, 8). Immunization of mice with hsp65 protects against pristane-induced arthritis by inducing IL-10- and IL-4-producing CD4 T cells (9). Both IL-4 and IL-10 are potent down-regulators of monocyte production of proinflammatory mediators, such as TNF-α, IL-8, IL-1, and PGE₂ (10, 11, 12). These data suggest that some large autologous hsp may stimulate antiinflammatory cytokine activity. This antiinflammatory function of hsp is controversial, however, because hsp60 has also been shown to induce TNF-α in a human monocyte cell line and TNF-α, as well as IL-15 and IL-12, in murine bone marrow-derived macrophage (7, 13).

Hsp27, an important member of the small hsp family, has been investigated primarily for its role as a circulating protein marker of increased malignancy in breast cancer (14). Hsp27 has been shown to down-regulate reactive oxygen intermediate (ROI) production, thereby protecting from TNF-α-mediated apoptosis (15). The glutamine induction of rat hsp25, the analogue of human hsp27, has been shown to correlate with protection from lethal endotoxin shock (16). Human monocytes from patients with systemic inflammatory response syndrome have significantly elevated hsp27 expression (17). These data suggest that exogenous hsp27 may also have some antiinflammatory or immune modulatory capacities on monocytes. IL-10 can also down-regulate ROI activity in monocytes and macrophages, but, unlike the large hsp, hsp27 has not been previously shown to exogenously induce production of either pro- or antiinflammatory cytokines (18, 19). However, hsp27 is a substrate for mitogen-activated protein kinase (MAPK)-activated protein kinase-2 (MAPKAPK-2), an important member of the p38 MAPK cascade that is both activated by cytokine treatment and critical in monocyte production of cytokines (10, 20, 21, 22). Recently, the activation (phosphorylation) of p38 MAPK and its substrate, MAPKAPK-2, has been shown to be crucial to LPS induction of IL-10 in human monocytes, further suggesting that hsp27 could play an antiinflammatory role in monocytes (10). Circulating hsp27 is present in the serum of cancer patients and, in some cases, induces in vivo hsp27 Ab production, suggesting that hsp27 can stimulate as an exogenous protein (23, 24). Phosphorylated hsp27 has also been identified as being associated with cell membranes of lamellipodia in migrating cells, suggesting a possible hsp27 surface expression (25). Although hsp27 phosphorylation after MAPKAPK-2 activation is necessary for LPS induction of monocyte IL-10, the effect of exogenous hsp27 on increasing production of IL-10 or any monokine is unexplored. Administration of IL-10 has been shown to suppress lethal endotoxemia and reduce serum TNF-α levels (26). Because of its antiinflammatory properties, IL-10 has been suggested as a possible therapeutic agent for inflammatory conditions, such as rheumatoid arthritis and inflammatory bowel disease (26). Consequently, any monocyte-modulating activity of hsp27 in increasing IL-10 levels without concomitantly highly inducing proinflammatory monokines such as TNF-α could also have therapeutic implications.

In this study, hsp27 has been assessed for a novel ability to induce IL-10 and/or TNF-α in human monocytes when added exogenously. We demonstrate that human hsp27 is a potent inducer of IL-10 in human monocytes, but only a modest inducer of TNF-α. We have also explored the preferential involvement of different MAPK pathways during hsp27-induced monocyte IL-10 production to determine whether hsp27 induced IL-10 via the same p38, MAPKAPK-2, hsp27 pathway as LPS stimulation. Although modest levels of TNF-α were induced by hsp27, only a small portion of the monocyte hsp27 induction of IL-10 was due to its prior induction of TNF-α, a known enhancer of IL-10 in monocytes (10, 27). Hsp27 independently induced high levels of monocyte IL-10, while concomitantly stimulating only minimal levels of TNF-α. Hsp27 induction of monocyte IL-10 is totally dependent on the activation of the p38 MAPK pathway and, unlike monocyte TNF-α production, independent of the ERK pathway activation, further supporting a TNF-α-independent IL-10 induction by hsp27.

FBS was purchased from Sigma (St. Louis, MO). Culture media and other supplements were purchased from Irvine Scientific (Santa Ana, CA). Muramyl dipeptide (MDP) was generously provided by CIBA-GEIGY (Basel, Switzerland). Staphylococcal enterotoxin B (SEB) and zymosan A were purchased from Sigma, and polymyxin B was purchased from Calbiochem (La Jolla, CA). The mAbs, My4 (CD14) FITC and IgG2b FITC were purchased from Coulter (Hialeah, FL). Recombinant human hsp27 was purchased from Stressgen Biotechnologies (Victoria, Canada). Polyclonal Ab against hsp27 was purchased from Santa Cruz Biotechnology (Santa Cruz, CA), mAb against hsp27 from Stressgen Biotechnologies, and mAb against TNF-α from Endogen (Woburn, MA). SB203580 and PD98059 were purchased from Calbiochem (San Diego, CA). Human rTNF-α was kindly provided by The National Institute for Biological Standards and Control (Potters Bar, U.K.). Phosphoplus p38 MAPK, p44/42 (ERK 1/2) MAPK, and stress-activated protein kinase (SAPK)/c-Jun N-terminal kinase (JNK) kits were purchased from New England Biolabs (Beverly, MA). Anti-phosphoserine mAb was purchased from Calbiochem. MAPKAPK-2 IP-kinase assay kit was purchased from Upstate Biotechnology (Lake Placid, NY). ³²P and ECL reagents were purchased from NEN Life Science Products (Boston, MA).

PBMC were first isolated from venous blood of healthy volunteers by Ficoll-Hypaque density centrifugation. Monocytes were separated from PBMC by selective adherence to microexudate-coated plastic surfaces, as described (28). Adherent monocytes (>95% purity, as checked by flow-cytometric analysis) were collected by treatment with 10 mM EDTA, suspended in IMDM medium, supplemented with 10% FBS, 50 U/ml penicillin G, 50 μg/ml streptomycin, 50 μg/ml gentamicin, 2.5 μg/ml fungizone, 4 mM l-glutamine, 1 mM sodium pyruvate, and 1% MEM nonessential amino acids. Endotoxin contamination was less than 12 pg/ml in the culture medium and FBS. Polymyxin B was added (20 U/ml) to all the washing and culture media to block the effect of any contaminating LPS. In some experiments, polymyxin B was used at a higher concentration (200 U/ml) in monocyte culture. Monocytes were cultured (1 × 10⁶ cells/ml) for 16–18 h in the presence or absence of 20 μg/ml of MDP + SEB (0.5 μg/ml) or human hsp27 (2 μg/ml). Culture supernatants were harvested and stored at −80°C until they were tested for IL-10 or TNF-α. In some experiments, monocytes were also stimulated with zymosan A (50 μg/ml), a potent inducer of monocyte IL-10 and TNF-α as an additional positive control. In some experiments, monocytes were stimulated with rTNF-α (2.5 ng/ml) alone or in combination with hsp27 (2 μg/ml). In selected experiments, hsp27 was first incubated with α-hsp27 polyclonal Ab (20 μg/ml) for 3 h before its addition to monocyte culture or α-TNF-α mAb (10 μg/ml) was added, together with hsp27, to monocyte culture. In some experiments, monocytes were first treated with SB203580 (10 μM), or PD98059 (10 μM), or the DMSO control (solvent used for dissolving both the reagents) for 2 h before addition of hsp27 to the culture.

A total of 2 × 10⁶ monocytes was stimulated in the presence or absence of MDP (20 μg/ml) + SEB (0.5 μg/ml) or hsp27 (2 μg/ml) for 8–9 h. Total cytoplasmic RNA was isolated using Tri-reagent (Molecular Research Center, Cincinnati, OH), according to manufacturer’s instructions. Antisense probes were labeled with [³²P]UTP (NEN Life Science Products) using the Riboquant in vitro transcription labeling kit (PharMingen, San Diego, CA), according to manufacturer’s instructions. A mixture of probes, Riboquant hCK-1 (PharMingen), was used to facilitate the simultaneous quantification of several RNA species. The antisense probes generated using this probe set include the controls, GAPDH and L32 and the human cytokine IL-10 and some other human cytokines, IL-5, IL-4, IL-14, IL-15, IL-9, IL-2, IL-13, and IFN-γ. The RNase protection assays were performed using the Riboquant RNase protection assay kit (PharMingen), according to manufacturer’s instructions. In brief, molar excesses of labeled probes were incubated with RNA derived from cells in hybridization buffer supplied by the manufacturer for 16–48 h at 56°C. Hybridized samples were then digested with 5 U of RNase A/T1 mixture for 45 min at 30°C. Subsequent to digestion, the protected fragments were separated from digested probe by electrophoresis on an 8 molar urea, 5% polyacrylamide Tris borate EDTA gel. The gels were then dried, exposed directly to film, and developed. The band intensities were quantitated using the National Institutes of Health image software. IL-10 mRNA levels were adjusted according to L32 and GAPDH levels (used as loading controls).

Monocytes (1.5 × 10⁶ cells) were cultured in serum-free medium for 2 h and then stimulated with hsp27 (2 μg/ml) for different time periods (1 min to 3 h). Western blot analysis was performed, essentially as described previously (29). Briefly, cells were lysed using a buffer consisting of 1% Nonidet P-40, 50 mM HEPES (pH 7.2), 100 mM NaCl, 2 mM EDTA, 1 mM pyrophosphate, 2 mM Na₃VO₄, 10 mM NaF, 1 mM PMSF, 10 μg/ml leupeptin, and 10 μg/ml aprotinin. Postnuclear supernatants were harvested after centrifugation of the lysate for 15 min at 14,000 × g at 4°C. Equal amounts of postnuclear lysates were boiled for 5 min in the presence of SDS sample buffer (reducing) and subjected to SDS-12% PAGE and then transferred to nitrocellulose membrane (Millipore, Bedford, MA) in transfer buffer (25 mM Tris, 192 mM glycine, pH 8.3, 20% (v/v) methanol). Membranes were first rinsed in TTBS (TBS with 0.1% Tween-20) and then blocked for 1 h at room temperature in TTBS-5% w/v nonfat dry milk. The membrane was then incubated overnight at 4°C with antiphospho-p38 MAPK Ab (rabbit polyclonal; New England Biolabs) (1/1000 dilution in TTBS-1% BSA). Ab-Ag complexes were detected with the aid of HRP-conjugated anti-rabbit secondary Ab (1/2000 dilution) (New England Biolabs), followed by detection of the bands with ECL reagent (NEN Life Science Products). The same membranes were used for detection of several other proteins, such as phospho-ERK 1/2 (p44/42), phospho-SAPK/JNK, total p38 MAPK, total ERK 1/2, and total SAPK/JNK by sequential stripping of Abs, by incubation of the membrane for 30 min at 50°C in a specific buffer (2% SDS, 100 mM 2-ME, 62.5 mM Tris-HCl, pH 6.7), and then reprobing the blot with respective Ab (all Abs; New England Biolabs) using the procedure as mentioned above for the assessment of phospho-p38 MAPK.

Activation of endogenous hsp27 was assessed, as described above, for activation of the MAPKs. In brief, monocytes were activated with recombinant hsp27 (2 μg/ml) for different time periods (5–120 min). Cells were washed twice at 4°C to remove recombinant hsp27 from the culture and then lysed as above. Equal amounts of postnuclear lysates were then subjected to SDS-15% PAGE and immunoblotted with anti-phosphoserine mAb (1/500 dilution) for the detection of phospho-hsp27. Recombinant hsp27 was used as a positive control. Presence of equivalent amounts of hsp27 in the cell lysates was confirmed by stripping the blot and reprobing with anti-hsp27 mAb.

Monocytes (1.5 × 10⁶) were cultured in serum-free medium for 2 h and then stimulated with MDP (20 μg/ml) + SEB (0.5 μg/ml), hsp27 (2 μg/ml), or UV (as positive control) for 30 min. Postnuclear lysates were prepared as described above. Protein (A + G) (20 μl of beads/sample) (Santa Cruz Biotechnology) was first washed twice with ice-cold PBS and then the MAPKAPK-2 assay was performed as described, using a specific kit (Upstate Biotechnology) (10). In brief, washed protein (A + G) was incubated with anti-MAPKAPK-2 sheep polyclonal Ab for 1 h at 4°C. In some experiments, protein (A + G) was incubated with sheep IgG for the Ab control. Ab-bound protein (A + G) was then washed twice with ice-cold PBS, followed by incubation with the postnuclear lysate sample for 2 h at 4°C in ice-cold RIPA buffer (50 mM Tris, pH 7.5, 1 mM EDTA, 1 mM EGTA, 1 mM Na₃VO₄, 0.1% 2-ME, 1% Triton X-100, 5 mM sodium pyrophosphate, 10 mM sodium glycerophosphate, 0.1 mM PMSF, 1 μg/ml aprotinin, 1 μg/ml leupeptin, and 50 mM NaF) with thorough mixing. The protein (A + G)-enzyme immune complex was washed once with ice-cold RIPA buffer containing 0.5 M NaCl, and then twice with ice-cold RIPA buffer and once with kinase assay buffer (20 mM MOPS, pH 7.2, 25 mM β-glycerol phosphate, 5 mM EGTA, 1 mM Na₃VO₄, 1 mM DTT). The beads were resuspended in 10 μl of kinase assay buffer, followed by addition of 10 μl of 1 mM hsp27 peptide sequence KKLNRTSVA (used as substrate). Reactions were initiated by the addition of 10 μl of [γ-³²P]ATP (10 μCi/assay) diluted in magnesium/ATP mixture (75 mM magnesium chloride and 500 mM ATP in kinase assay buffer). The reaction was allowed to proceed for 30 min at 30°C before termination. This was achieved by spotting the assay mixture onto squares of p81 paper and then placing them in 0.75% orthophosphoric acid. The squares were washed three times in the acid and once in acetone before scintillation counting.

Cell phenotype verification in our monocyte populations was conducted using anti-CD14 mAb. IgG2b FITC was used as the isotype control. Fluorescent measurements were done on the Coulter Epics XL flow cytometer. Briefly, 5 × 10⁵ cells were incubated with conjugated mAb or with the appropriate isotypic control for 1 h at dilutions suggested by the manufacturer. Samples were washed twice with PBS and resuspended in 500 μl PBS for fluorescent analysis.

IL-10 and TNF-α levels in the culture supernatants were determined by specific ELISA kit (Endogen), according to the instructions of the manufacturer. The sensitivity of the assay was 5 pg/ml.

Results are expressed as mean ± SEM. Statistical significance was calculated by the Student’s t test (paired) using the StatView program. Statistical significance was accepted for p < 0.05.

To investigate pro- or antiinflammatory monokine induction by hsp27, we treated human monocytes with recombinant human hsp27 and assessed the culture supernatants for IL-10 and TNF-α levels. One recurring problem with using recombinant proteins to induce monokines is that even picogram amounts of contaminating LPS can augment other stimuli, confusing LPS independent and dependent monokine induction by exogenous recombinant proteins. A combination of SEB plus MDP, rather than LPS, was used as a positive control for induction of monokines. MDP, or MDP + SEB, induce significant quantities of IL-10 and TNF-α in monocyte/macrophage, even in the presence of polymyxin B (28, 30). Consequently, MDP + SEB was used as a control stimuli and 20 U/ml of polymyxin B was included in all media to detect monokine production induced exclusively by hsp27. Hsp27 induced significantly (p = 0.0009) higher amounts of IL-10 as compared with adherence-stimulated, untreated monocytes (10-fold higher) or even SEB + MDP-stimulated monocytes (3- to 4-fold higher) (Fig. 1,A). Even increasing the added polymyxin B to 200 U/ml did not diminish monocyte IL-10 induction by recombinant hsp27. The hsp27 induction of IL-10 protein was maximal (≈10-fold increase) at 16–18 h. MDP + SEB-induced monocyte IL-10 levels continued to slightly increase up to 40 h of culture, but still remained 3-fold less than hsp27-induced IL-10 levels. Hsp27 increased both the kinetics of monocyte IL-10 induction as well as the absolute levels, suggesting hsp27 induction was direct and not working through induction of a second monokine. Combination of hsp27 with SEB + MDP did not significantly increase IL-10 induction over hsp27 alone (≈3400–3700 pg/ml), suggesting maximal monocyte IL-10 levels were induced by hsp27. In a separate set of experiments, zymosan A, a stimuli of IL-10, reported equal in potency to LPS, induced ×845 pg IL-10 vs ×4272 pg by hsp27 (30). Hsp27-induced monocyte IL-10 production was dose dependent, with 1–5 μg/ml being the optimum concentration (Fig. 1,B). Finally, to further demonstrate the specificity of hsp27 direct induction of monocyte IL-10, hsp27 was treated with specific anti-hsp27 Ab for 3 h before its addition to the monocyte cultures. Anti-hsp27 treatment, but not addition of irrelevant Ab, could abolish the hsp27 induction of monocyte IL-10 (Fig. 1 C). These findings suggest that hsp27 itself potently induces monocyte IL-10.

Adherence alone induces some IL-10 in human monocytes in the absence of any other stimulants (28). Consequently, hsp27 might only be augmenting IL-10 protein translation of adherence-stimulated, already transcribed IL-10 mRNA rather than increasing hsp27 mRNA levels. To explore this possibility, we assessed monocyte IL-10 mRNA expression with the RNase protection assays (Fig. 2). Hsp27 induced almost 7.2-fold increases in mRNA levels, as compared with only adherence-stimulated monocytes. Hsp27-induced IL-10 mRNA levels were 3.2-fold higher than the control, MDP + SEB-induced IL-10 mRNA levels, again demonstrating the potency of hsp27 as an IL-10 inducer.

Recently, hsp60 has been shown to induce ≈750 pg/ml TNF-α in Mono Mac-6, a human monocyte cell line (7). Moreover, TNF-α is a potent augmentor of IL-10 production in human monocytes (10, 27). TNF-α induction occurs before IL-10 induction in human monocytes after LPS stimulation (31). Thus, exogenously added hsp27 could first induce high levels of monocyte TNF-α, which only subsequently autocrine stimulated the monocytes to produce IL-10. Maximal LPS induction of monocyte IL-10 has been reported to require prior induction of endogenous TNF-α (27, 32). Because of the rapidity of monocyte TNF-α induction, it was possible that hsp27 induction of monocyte IL-10 similarly was dependent on TNF-α, despite the observed increase in IL-10 induction kinetics. We, therefore, assessed hsp27-induced TNF-α production in human monocytes. Hsp27 significantly (p = 0.0003) induced TNF-α levels in human monocytes (Fig. 3,A). However, in contrast to hsp27’s exaggerated monocyte IL-10-inducing potential (10-fold vs 3-fold, as compared with monocyte IL-10-inducing potential of SEB + MDP), hsp27 and SEB + MDP induced almost identical levels of monocyte TNF-α (Fig. 3 A). In addition, the combination of SEB + MDP + hsp27 further significantly (p = 0.002) increased monocyte TNF-α production from 483 ± 74 for hsp27 alone to 1737 ± 267 pg/ml. These data are in contrast to the failure of the same combination (SEB + MDP + hsp27) to increase monocyte IL-10 production over maximal IL-10 production (≈3400 pg/ml) induced by hsp27 alone.

In the next experiments, we added anti-TNF-α Ab, along with hsp27, to the monocyte culture to delineate any critical role of endogenously produced TNF-α levels during hsp27-induced monocyte IL-10 production. As can be seen in Fig. 3,B, anti-TNF-α Abs did partially (≈40%) inhibit hsp27-induced IL-10 production. This anti-TNF-α Ab completely inhibited TNF-α-induced IL-10 production. To further test a requirement for TNF-α in the hsp27 induction of IL-10, we compared the monocyte IL-10 induction by 2500 pg of TNF alone or in combination with hsp27 to hsp27 alone. As previously reported, TNF-α alone induced only minimal monocyte IL-10 levels (10 ; Fig. 3,C). As illustrated in Fig. 3,C, addition of 2500 pg/ml of exogenous TNF-α, along with 2 μg/ml hsp27, failed to augment monocyte IL-10 induction over hsp27 alone (3098 ± 660 vs 3021 ± 729 pg). This 2500 pg of TNF-α is five times the levels of TNF-α induced by hsp27 (Fig. 3,A). Even addition of 10,000 pg/ml of TNF-α failed to augment the hsp27 induction of IL-10 (data not shown). As a further examination of hsp27 potency in inducing IL-10, but its minimal TNF-α induction, we compared zymosan A induction of these two monokines with hsp27 induction. As illustrated in Table I, zymosan A induced high levels of monocyte TNF-α (12,738 ± 3,747), but more modest levels of IL-10 (845 ± 320), while hsp27 had the opposite pattern, inducing exaggerated levels of IL-10 (4272 ± 217 pg/ml) vs modest TNF-α levels (301 ± 190 pg/ml). In total, these data suggest that hsp27 is an unusually potent stimuli for monocyte IL-10, but only a modest stimuli for monocyte TNF-α, and that TNF-α induction is not necessary for maximal hsp27 induction of monocyte IL-10.

LPS-induced monokine production involves the activation of three different MAPK, p38, p44/42 (ERK 1/2), and p46/54 (SAPK/JNK-1/2). However, LPS induction of monocyte IL-10 requires only activation of the p38 pathway, while TNF-α induction requires both ERK and p38 (10, 22, 33). Therefore, hsp27 induction of IL-10 could involve only the p38 pathway or both the ERK and p38 pathway if TNF-α production is required. In addition, endogenous hsp27 is itself a substrate in the required p38 activation pathway by which LPS induces IL-10. Consequently, the hsp27 preferential induction of monocyte IL-10 might represent just adding more hsp27 as a substrate, not increasing the overall activation of the MAPK activation pathways. To assess whether there was activation (phosphorylation) of these MAPKs after hsp27 addition, human monocytes were stimulated at different time points, and activated p38, ERK 1/2, and JNK 1/2 measured using the respective Abs against their phosphorylated forms. As can be seen in Fig. 4, hsp27 activated all three MAPKs pathways, but to different degrees. Phosphorylation of ERK 1/2, JNK 1/2, as well as p38 MAPK was clearly increased at 20 min after addition of hsp27. Maximal stimulation was observed at 40 min after hsp27 addition for all three MAPKs. However, activation of p38 MAPK persisted up to 180 min when P-p44 ERK, as well as P-p54 and P-p46 JNK were clearly declining (Fig. 4). In addition, hsp27 induction of P-p54 JNK was minimal compared with its activation of p38 and ERK. Activation of MAPKAPK-2 (a substrate of p38 MAPK) has been shown as necessary to LPS induction of IL-10 in human monocytes (10). Therefore, we also assessed the activation of MAPKAPK-2 during hsp27-induced activation and IL-10 production of human monocytes by in vitro kinase assay, using a sequence of hsp27 (KKLNRTSVA) as the substrate. As can be seen in Fig. 5,A, the immunoprecipitate (using α-MAPKAPK-2 sheep polyclonal Ab; control sheep IgG had no MAPKAPK-2 activity) from hsp27-activated monocyte lysate had significantly increased MAPKAPK-2 activity vs that from adherence-stimulated untreated monocyte lysate, suggesting that exogenously added hsp27 could activate (phosphorylate) human monocyte endogenous hsp27. To test this possibility, we stimulated human monocytes at different time points with recombinant hsp27 and measured activated (phosphorylated at serine residues) endogenous hsp27 in the cell lysates by Western blotting, using an Ab against phosphoserine. The presence of equivalent amounts of hsp27 in the cell lysates was tested by reprobing the stripped blot with anti-hsp27 Ab. As can be seen in Fig. 5,B, exogenous hsp27 addition could activate monocyte endogenous hsp27. Moreover, just as hsp27 induced prolonged p38 activation (Fig. 4), endogenous monocyte hsp27 remained activated for prolonged periods (up to 2 h tested) after addition of exogenous recombinant hsp27. These data suggest that hsp27 is a potent inducer of IL-10 in human monocytes because it differentially activates the MAPK pathways that play critical roles in inducing monokine production. The next sets of experiments examined which of the different MAPKs had critical roles in hsp27-induced IL-10 and TNF-α production by monocytes.

Although the p38 MAPK pathway is critical in LPS induction of IL-10, the differential involvement of MAPK pathways in cytokine production has been shown to be stimulus dependent and might contribute to hsp27 preferential induction of IL-10 (10, 21, 33, 34). To assess any essential role of differential MAPK activation for induction of monocyte IL-10 or TNF-α by hsp27, we added different MAPK inhibitors to the monocyte culture before addition of hsp27. SB203580 was used to block the effect of p38 MAPK, whereas PD98059 was used to inhibit the effect of MEK 1/2 (the enzyme responsible for activation of ERK 1/2) (10, 33). SB203580 could significantly (p = 0.002) block hsp27-induced IL-10 production (Fig. 6). Monocyte IL-10 production was inhibited ≈80% by SB203580, which also blocked 90% of the TNF-α activity induced by hsp27, indicating a potential critical role of p38 MAPK pathway during induction of both monocyte IL-10 and TNF-α production by hsp27. However, even in the presence of SB203580, hsp27 induced a small amount of IL-10, which was still significantly (p = 0.002) increased over that of adherence only-stimulated monocytes (Fig. 6). In addition, the SB203580 treatment had no effect on the hsp27 induction of ERK activation in these monocytes, indicating that the inhibition of hsp27-induced monocyte IL-10 was not a result of general loss of monocyte viability, nor did the DMSO control affect hsp27-induced monocyte IL-10 or TNF-α production (data not shown). In contrast to the inhibitory effects of SB203580, PD98059 had no inhibitory effect on hsp27-induced monocyte IL-10 production, but blocked 68% of the TNF-α induced by hsp27 (Fig. 6). These data suggest that activation of the ERK 1/2 pathway is not required for induction of monocyte IL-10 by hsp27, but that both the ERK 1/2 and p38 pathways are involved in hsp27 induction of monocyte TNF-α. These data further support that monocyte TNF-α production induced by hsp27 is not required for maximal IL-10 production.

This is the first demonstration of a novel role for exogenous hsp27 in the preferential induction of monocyte IL-10 production. Hsp interactions in the immune response have been a focus of many recent investigations (6, 13). Primarily, the large hsp (6, 7) have been examined for their stimulation or regulation of T cells and their augmentation of Ag presentation (6, 9, 13, 35). However, one of the small hsp, hsp27, has been implicated as interacting with granzyme A in mediating granzyme A-induced cell death and in cystic fibrosis arthritis, as well as glaucoma (36, 37). Although hsp27 is capable of acting as a classic thermoprotective protein, its thermoprotective and protein chaperone activities are independent of its phosphorylation (38). Hsp27 can play other physiologic roles when phosphorylated (15, 20, 39). There are a number of suggestions that hsp27 might be exogenously active on monocytes. Hsp27 is both more constitutively expressed and more cytokine inducible in monocytes than in T cells (40). Hsp27 is phosphorylated in monocytes as a substrate of MAPKAPK-2 after LPS or TNF-α activation of the p38 MAPK pathway and is seen as a circulating protein that can eventually induce Abs in breast cancer and cystic fibrosis patients with arthritis (23, 36). Increased hsp27 expression has also been suggested as protecting from TNF-α-mediated apoptosis (15). Induction of hsp27 has been reported to result in protection from lethal endotoxin shock and to be highly induced in monocytes from patients with systemic inflammatory shock syndrome, supporting a possible antiinflammatory role for hsp27 (16, 17). Hsp60, one of the large hsp, can induce monocyte TNF-α, and TNF-α induction augments monocyte IL-10 levels (7, 32). These data, combined with those emphasizing the pivotal role of MAPKAPK-2, the kinase that phosphorylates hsp27, led us to investigate exogenous hsp27 as a stimuli for monocyte IL-10 or TNF-α.

Hsp27 addition to primary human monocyte cultures induced exaggerated IL-10 production (≈3400 pg/ml). In contrast, LPS-stimulated IL-10 levels reported for primary human monocytes range from 500-1500 pg/ml, similar to our SEB + MDP- and zymosan A-induced monocyte IL-10 levels of ≈800–1100 pg/ml (30, 31, 41). However, LPS simultaneously stimulates the same human monocytes to produce from 6,000 to 15,000 pg of TNF-α (30, 31, 41). In our polymyxin B-containing system, zymosan A similarly induced high levels of TNF-α (≈13,000 pg/ml), but only ≈800–900 pg of IL-10. Hsp27 induced ≈300–500 pg/ml of monocyte TNF-α, somewhat similar to the reported 750 pg/ml induced by human hsp60 in the human Mono Mac-6 cell line (7). Hsp27’s exaggerated stimulation of human monocyte IL-10 vs its modest induction of TNF-α is further illustrated by our data showing that combining SEB + MDP + hsp27 induced no further increase in IL-10 levels, but augmented monocyte TNF-α levels more than 3-fold. Similarly, TNF-α addition did not further increase hsp27-induced IL-10 levels. These data imply that hsp27 alone induced maximal monocyte IL-10, but only minimal TNF-α levels, and that hsp27 may preferentially activate human monocyte IL-10 production. Increased production of IL-10 by APC has been suggested to polarize the immune responses to Th2 (42). The paradigm that bacterial hsp may pivotally act as stimulating danger signals whose immune recognition is highly conserved, while autologous Hsp could have preferential immune regulatory activity has been previously postulated (6, 7, 43, 44). Our data showing recombinant human hsp27 to preferentially induce antiinflammatory IL-10 rather than proinflammatory TNF-α in human monocytes are consistent with this paradigm.

The mechanism(s) behind hsp27 preferential induction of monocyte IL-10 may involve preferential activation of the MAPK p38 pathway. Previous reports demonstrated that the p38 but not the ERK pathway was required for LPS induction of IL-10 (10). However, LPS induction of maximal monocyte IL-10 appears also to be pivotally dependent on its induction of TNF-α and/or IL-1 (10, 32). LPS-stimulated maximal IL-10 production by monocytes is consequently delayed, peaking at 24 h after LPS stimuli and continuing for 96 h (27). In contrast, hsp27-induced monocyte IL-10 levels were maximal at 18 h, and anti-TNF-α mAb only inhibited 40% of hsp27-induced IL-10 levels in our experiments vs a reported 70% inhibition of LPS-induced IL-10 (32). These data suggest that exogenous human hsp27 directly stimulates human monocytes to produce IL-10. Inhibition of the p38 MAPK pathway with SB203580 reduced our hsp27-induced monocyte IL-10 levels by 80%, while the ERK inhibitor, PD98059, which inhibited TNF-α production almost 70%, had no effect on maximal IL-10 production. The significant level of hsp27-induced IL-10 generated even in the face of MAPKAPK-2 inhibition by SB203580 might result from other pathways of hsp27 phosphorylation since activated PKC-δ has been also demonstrated to directly phosphorylate hsp27 (20). Alternatively, the SB203580 may only be capable of inhibiting 80% of MAPKAPK-2 activity because of prior adherence-induced activation. The exaggerated monocyte IL-10 levels induced by hsp27 could result from the persistence of phosphorylated p38 (P-p38) in the hsp27-induced monocytes. Similar to LPS-induced human monocyte activation data, levels of phosphorylated ERK p42/p44, JNK p46/p54, and p38 peaked at 40-min post-hsp27 stimulation, then both ERK p42/p44 and JNK p46/p54 declined (33). In contrast to the published data for LPS induction of human monocyte MAPK pathways, hsp27-induced P-p38 persisted even at 180 min (33). Since P-p38 activation of MAPKAPK-2 is required for monocyte IL-10 induction, this increased persistence of P-p38 may partially explain both the higher levels of MAPKAPK-2 induced by hsp27 (2-fold) vs SEB + MDP (1.3-fold) and hsp27’s exaggerated induction of IL-10. Phosphorylation of endogenous hsp27 was also prolonged in hsp27-induced monocytes. Alternatively or additionally, the rapid disappearance of JNK p46/p54 after hsp27 stimulation might be allowing continued monocyte IL-10 production. A critical down-regulatory role for activation of the JNK pathway has been implicated for IL-10 production, since JNK 1⁻/⁻ Th cells produced exaggerated IL-10 in response to CD3 induction (45). Our data suggested that JNK was only minimally activated by hsp27. The importance of this minimal JNK pathway activation in hsp27 induction of exaggerated monocyte IL-10 needs to be further explored, utilizing JNK inhibitors. However, JNK inhibitors are not yet commercially available.

The ability of exogenous hsp27 to differentially stimulate MAPK pathway activation and distinctive cytokine production might imply that it is binding to a monocyte surface receptor rather than nonspecifically internalized by pinocytosis. The different stimulatory capacity of autologous vs microbial hsp, despite their sequence homology, along with their surface binding of peptides, has led several investigators to propose a putative hsp receptor (7, 8, 35). The receptor for advanced glycation end products (RAGE) has been proposed as a likely candidate for an hsp receptor (8). The RAGE levels on monocytes are up-regulated in diabetics (46), a disease in which large hsp have been shown to have immunoregulatory activity (47) and RAGE activation induces MAPK signaling pathways (48). Most of the hypotheses developed and the immune regulation investigations performed involved the large hsp. As far as we could determine, our study represents the first described data of cytokine-inducing activities for exogenous hsp27 and is consistent with differential monocyte activation through an hsp27 receptor.

In summary, we have demonstrated a novel role for exogenous human hsp27 in preferentially inducing human monocytes to produce IL-10 by activating the p38 MAPK, MAPKAPK-2, and endogenous hsp27 in a prolonged fashion. Exogenous hsp27 was able to induce monocyte IL-10 mRNA increases as well as protein increases. Increased IL-10 induction may be responsible for the reported ability of hsp27 to down-regulate TNF-α-mediated apoptosis and ROI induction (3, 15). Whether our demonstration of hsp27-mediated antiinflammatory function is limited to human monocytes or is a unique property of autologous hsp27 vs large hsp remains to be investigated. The ability of hsp27 to induce a predominant in vitro antiinflammatory monokine profile could imply that some autologous small hsp serve as antiinflammatory agents in vivo.

We thank Laura Orphin for her excellent technical assistance and Claire Lavallee for manuscript preparation.