© 2000 Nature America Inc. • http://medicine.nature.com
ARTICLES
HSP70 stimulates cytokine production through a CD14-dependant pathway, demonstrating its dual role as a chaperone and cytokine ALEXZANDER ASEA1, STINE-KATHREIN KRAEFT2, EVELYN A. KURT-JONES1, MARY ANN STEVENSON1, LAN BO CHEN2, ROBERT W. FINBERG1, GLORIA C. KOO3 & STUART K. CALDERWOOD1
© 2000 Nature America Inc. • http://medicine.nature.com
1
Department of Adult Oncology, Dana-Farber Cancer Institute,Harvard Medical School, 44 Binney Street, Boston, Massachusetts 02115, USA 2 Department of Cancer Biology, Dana-Farber Cancer Institute,Harvard Medical School, 44 Binney Street, Boston, Massachusetts 02115, USA 3 Department of Immunology Research, Merck Research Laboratories, 126 East Lincoln Street, Rahway, New Jersey 07065, USA Correspondence should be addressed to S. K. C.;
[email protected] E.A.K.-J. & R.W.F. present address: Department of Medicine, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, Massachusetts 01655, USA
Here, we demonstrate a previously unknown function for the 70-kDa heat-shock protein (HSP70) as a cytokine. HSP70 bound with high affinity to the plasma membrane, elicited a rapid intracellular calcium flux, activated nuclear factor (NF)-κB and upregulated the expression of pro-inflammatory cytokines tumor necrosis factor (TNF)-α, interleukin (IL)-1β and IL-6 in human monocytes. Furthermore, two different signal transduction pathways were activated by exogenous HSP70: one dependent on CD14 and intracellular calcium, which resulted in increased IL1β, IL-6 and TNF-α; and the other independent of CD14 but dependent on intracellular calcium, which resulted in an increase in TNF-α but not IL-1β or IL-6. These findings indicate that CD14 is a co-receptor for HSP70-mediated signaling in human monocytes and are indicative of an previously unrecognized function for HSP70 as an extracellular protein with regulatory effects on human monocytes, having a dual role as chaperone and cytokine.
Heat-shock proteins are highly conserved proteins found in all prokaryotes and eukaryotes. In normal physiological conditions, heat-shock proteins are expressed at low levels1. However, a wide variety of stressful stimuli, including environmental (ultraviolet radiation, heat shock, heavy metals and amino acids), pathological (viral, bacterial or parasitic infections, or fever, inflammation, malignancy or autoimmunity) or physiological stimuli (growth factors, cell differentiation, hormonal stimulation or tissue development), induce a substantial increase in intracellular heat-shock-protein synthesis2, known as the stress response. The main functions ascribed to heat-shock proteins are as an intracellular molecular chaperone of naive, aberrantly folded or mutated proteins, as well as in cytoprotection after the stressful stimuli described above. Indeed, the sensitivity of cancer cells to the chemical stresses encountered in the tumor microenvironment, caused by inadequate tumor perfusion or imposed by chemotherapy, is limited by the expression of cellular stress responses3. Given reports indicating increased levels of antibodies against the inducible form of the 70-kDa form of heat-shock protein (HSP70; Genome DataBase designation, HSPA) in patients with autoimmune diseases4, we speculated that heat-shock proteins are found in the extracellular milieu and could exert chaperone and regulatory effects on various immunocompetent cells. Here, we demonstrate that exogenous HSP70 acts as a cytokine to human monocytes by stimulating a pro-inflammatory signal NATURE MEDICINE • VOLUME 6 • NUMBER 4 • APRIL 2000
transduction cascade that results in an upregulation in interleukin (IL)-1β, IL-6 and tumor necrosis factor (TNF)-α expression. Furthermore, we address the signal transduction cascade stimulated by exogenous HSP70 and show that there are at least two different signals: a pathway dependent on CD14 and intracellular calcium that when activated results in the production of IL-1β, IL-6 and TNF-α; and a pathway independent of CD14 but dependent on intracellular calcium that results in the production of TNF-α but not IL-1β or IL-6. Thus, HSP70 functions in a dual role as a cytokine, as shown here, and as a chaperone. Exogenous HSP70 specifically binds to human monocytes We initially investigated whether the inducible form of HSP70 binds specifically to the surface of human monocytes. Treatment of human monocytes with highly purified, biotinylated HSP70 on ice (to inhibit internalization or phagocytosis) resulted in specific surface binding, as shown by fluorescence microscopic analysis (Fig. 1a) and confirmed by laser confocal microscopy (Fig. 1b). As little as 7 nM HSP70 was sufficient to obtain positive staining, indicating that HSP70 binds with high affinity to the cell surface. The best resolution for immunohistochemistry was achieved with 70 nM HSP70 (Fig. 1). Cross-sectional analysis through the center of human monocytes showed that at 4 °C, HSP70 binding was localized to the surface membrane and that approximately 80–90% of the cells bound HSP70 (Fig. 1b). 435
© 2000 Nature America Inc. • http://medicine.nature.com
ARTICLES Fig. 1 HSP70 specifically binds to the surface membrane of human monocytes. a, Fluorescence microscopy of human monocytes stained for surface expression of HSP70. Top left, phase contrast; top right, FITC-conjugated HSP70; bottom left, propidium iodide staining; bottom right, overlay of FITC-conjugated HSP70 and propidium iodide (nuclear) staining. Data represent four independent experiments with similar results. b, Laser confocal microscopy of a section through the center of human monocytes incubated with biotinylated-HSP70 (left), biotinylated ovalbumin (middle) or FITCconjugated streptavidin only (right). Scale bar represents 10 µm. Data represent two independent experiments with similar results.
© 2000 Nature America Inc. • http://medicine.nature.com
Control cells treated with biotinylated ovalbumin or fluorescein isothiocyanate (FITC)-conjugated streptavidin showed only weak, nonspecific binding (Fig. 1b). When binding was accomplished at 37 °C for 20 minutes, intracellular staining was found (data not shown). Exogenous HSP70 induces cytokine production We next determined whether the high-affinity surface binding shown by HSP70 has a functional consequence in human monocytes. Treatment of human monocytes with 7 nM exogenous HSP70 at 37 °C for 2–4 hours resulted in the upregulation in expression of pro-inflammatory cytokines IL-1β (from 1.2% at 0 hours to 71.1% at 2 hours and 84.5% at 4 hours after treatment), TNF-α (from 6.4% at 0 hours to 41.8% at 2 hours and 76.5% at 4 hours after treatment) and IL-6 (from 16.9% at 0 hours to 39.7% at 2 hours and 42.0% at 4 hours after treatment), as judged by flow cytometry analysis of the percentage of cells expressing these intracellular cytokines (Fig. 2a). Treatment with control protein ovalbumin did not upregulate the expression of either IL-1β, TNF-α or IL-6 substantially above background values (Fig. 2a). To study the mechanism of HSP70-induced cytokine upregulation, we compared the effect of HSP70 with that of lipopolysaccharide (LPS), a component of the outer membrane of Gram-negative bacteria and a potent activator of many functions of human monocytes, including pro-inflammatory cytokine expression5. LPS upregulated the expression of TNF-α (Fig. 2b), IL-1β and IL-6 (data not shown) with a potency similar to that of exogenous HSP70 (Fig. 2b). Ovalbumin, used as a protein control, did not increase TNF-α expression (Fig. 2b). To confirm that the HSP70-induced cytokine expression was indeed due to a cell surface signaling event, we pre-treated monocytes with pronase (a compound that hydrolyzes cell surface proteins including membrane bound receptors). Pronase pre-treatment completely abrogated HSP70- and LPS-induced expression of TNF-α (Fig. 2b). We next examined potential signal transduction steps in HSP70 activation of cytokine expression using small-molecule inhibitors. Pre-treatment with BAPTA-AM (an intracellular calcium chelator) blocked the expression of TNF-α in response to HSP70, indicating involvement of intracellular free calcium in transduction of the HSP70 signal (Fig. 2b). In contrast, LPS-induced TNF-α expression was not significantly inhibited by pretreatment of monocytes with BAPTA-AM (Fig. 2b). We then used the non-steroidal anti-inflammatory drug sodium salicylate at concentrations that inhibit the transactivation of NF-κB (ref. 6). Sodium salicylate completely abrogated HSP70- and LPS-induced expression of TNF-α (Fig. 2b). Pre-treatment of monocytes with pronase, BAPTA-AM or sodium salicylate followed by ovalbumin treatment did not significantly alter baseline levels of intracellular expression of TNF-α in human monocytes (Fig. 2b). To rule out the possibility of bacterial contamination of HSP70, 436
a
b
we pre-incubated human monocytes with polymyxin B or lipid IVa (potent inhibitors of LPS)7,8. Pre-incubation of human monocytes with polymyxin B or lipid IVa completely abrogated LPSinduced but not HSP70-induced expression of TNF-α, IL-6 and IL-1β (data not shown). The endotoxin content of HSP70 was consistently less than 0.01 ng/ml, as measured by Limulus amebocyte lysate assay. Furthermore, heat treatment of HSP70 (100 °C for 20 minutes) before incubation with cells completely abrogated HSP70-induced but not LPS-induced cytokine production. Trypan blue exclusion analysis, done at the end of each experiment to test for cell viability, showed that at the concentrations used, none of the compounds had a toxic effect on the cells (data not shown). These data demonstrate that HSP70 is a powerful regulator of pro-inflammatory cytokine expression in human monocytes, with a potency similar to that of LPS, and that intracellular free calcium ions and NF-κB are potential signaling intermediates in HSP70 induced cytokine production. Role of NF-κB and intracellular calcium ions Activation of NF-κB is an essential step in the upregulation of many pro-inflammatory cytokines by human monocytes9. The activation of NF-κB is regulated by its cytoplasmic inhibitor, IκBα, through phosphorylation of serine at residues 32 (Ser32) and 36 (Ser36), which targets it for degradation by the proteosome and releases NF-κB to migrate to the nucleus and activate the promoter of target genes10. To further address the signal transduction cascade activated by exogenous HSP70 that resulted in pro-inflammatory cytokine expression in human monocytes, we next assessed the phosphorylation of I-κBα at Ser32 (ref. 10,11) using antibodies specific for Ser32, followed by flow cytometric analysis. Treatment of monocytes with 7 nM exogenous HSP70 resulted in the phosphorylation of I-κBα at Ser32 (Fig. 3a). Treatment with the control protein ovalbumin did not result in the phosphorylation of I-κBα (data not shown). LPS at a concentration of 100 µg/ml, used as a positive control, stimulated I-κBα phosphorylation with a potency similar to that of exogenous HSP70 (Fig. 3a). To eliminate the possibility of NATURE MEDICINE • VOLUME 6 • NUMBER 4 • APRIL 2000
© 2000 Nature America Inc. • http://medicine.nature.com
The high-affinity binding of HSP70 to the cell surface of monocytes indicates the possibility of a HSP70 receptor coupled to a transmembrane signaling pathway. To address this, we next analyzed calcium flux, an early event after receptor occupation in many systems9, after treatment of cells with HSP70. Human monocytes, isolated from healthy blood donors and enriched by magnetic bead separation technique for CD14+CD45+ monocytes, showed a rapid, dose-dependent calcium flux when treated with HSP70 (Fig. 3d). The potency of exogenous HSP70induced calcium flux was similar to that of monocyte chemotactic peptide-1 (Fig. 3d), a chemokine known to induce calcium flux through interaction with carbon–carbon chemokine receptor CCR2 expressed on the surface of human monocytes13. Ovalbumin, used at similar concentrations, did not elicit a calcium flux (Fig. 3d). The possibility that bacterial contaminants such as LPS might be responsible for this calcium flux was eliminated by prior experiments demonstrating that LPS does not induce a calcium flux (ref. 14 and data not shown). In addition, pre-treatment with the LPS antagonist polymyxin B failed to block calcium flux (data not shown).
contamination of purified HSP70 with LPS, we pre-treated cells with polymyxin B and lipid IVa. HSP70-induced phosphorylation of I-κBα was not substantially affected by pre-treatment with polymyxin B (Fig. 3a) or lipid IVa (data not shown). However, LPS-induced phosphorylation of I-κBα was completely abrogated by pre-treatment of monocytes with polymyxin B (Fig. 3a). Support for data showing the importance of intracellular free calcium in the regulation of cytokine expression in human monocytes (Fig. 2b) was demonstrated by experiments in which pre-treatment of human monocytes with the intracellular calcium chelator, BAPTA-AM abrogated HSP70-induced phosphorylation of I-κBα (Fig. 3a). The LPS-induced phosphorylation of I-κBα was not inhibited by pre-treatment with BAPTA-AM. This indicates different receptor-mediated signaling pathways that diverge at the receptor and converge at the phosphorylation of I-κBα and activation of NF-κB. We used BAPTA-AM at a concentration (10 µM) at which it prevented transient calcium flux but did not deplete basal intracellular free calcium levels in cells, an effect that requires higher BAPTA-AM concentrations. Neither polymyxin B alone nor BAPTA-AM alone had an effect on baseline levels of I-κBα phosphorylation (data not shown). We next used western blot analysis to confirm the involvement of intracellular calcium in HSP70-induced cytokine production. Pre-treatment of human monocytes with BAPTA-AM abrogated phosphorylation of I-κBα induced by HSP70 but not by LPS (Fig. 3b). Sodium salicylate inhibits the phosphorylation by the Ser/Thr protein kinase RSK2 of cyclic AMP response-elementbinding protein and I-κBα on residues essential for their transcriptional activity in vivo, and thus represses cyclic AMP response-element-binding protein and transcription dependent on NF-κB (ref. 12). Similarly, sodium salicylate inhibited phosphorylation of I-κBα induced by both HSP70 and LPS (Fig. 3b). This was further demonstrated by densitometry analysis (Fig. 3c) of the western blot in Fig. 3b. These results strongly indicate that the HSP70- and LPS-transduced signaling occurs through distinct receptor-mediated pathways that converge at NF-κB activation.
HSP70-induced cytokine production: involvement of CD14 In myeloid cells, the pro-inflammatory response induced by bacterial LPS is transduced through a glycosylphosphatidylinositolanchored membrane protein, CD14, and is increased by a plasma protein, LPS-binding protein15. To determine whether the putative surface bound HSP70 receptor is associated with CD14 in a similar manner, we used U373 ‘wild-type’ (U373) human astrocytoma cells that do not endogenously express the membrane bound CD14 receptor16, and U373 cells transfected with the human cDNA for CD14 (U373–CD14)(ref. 17). As expected, treatment of U373 ‘wild-type’ cells with 10 µg/ml HSP70 for 4 hours at 37 °C did not result in the upregulation of the expression of IL-1β or IL-6 (Fig. 4a). Unexpectedly, TNF-α expression was considerably increased in HSP70-treated U373 cells (Fig. 4a). As anticipated, HSP70 induced the expression of IL-1β and
a
b Relative cell number
© 2000 Nature America Inc. • http://medicine.nature.com
ARTICLES
IL-1β
TNF-α
Fig. 2 Exogenous HSP70 upregulates the expression of pro-inflammatory cytokines. a, Human monocytes were treated with ovalbumin (light lines) or HSP70 (dark lines) for 0 h (top row), 2 h (middle row) or 4 h (bottom row), then simultaneously fixed and permeabilized, and then stained with PE-conjugated IL-1β or TNF-α or FITC-conjugated IL-6 and analyzed by flow cytometry. Data represent the relative number of human monocytes (vertical axes) staining for intracellular cytokines (horizontal axes, intensity of NATURE MEDICINE • VOLUME 6 • NUMBER 4 • APRIL 2000
IL-6
staining, and represent two independent experiments with similar results. b, Human monocytes were pre-treated with culture medium only (), pronase ( ), BAPTA-AM () or sodium salicylate ( ), then washed and resuspended in ovalbumin (OVA), HSP70 or LPS. Data represent the percentage cells staining for intracellular TNF-α (mean ± s.e.m. of results from three independent experiments). *, P < 0.01, compared with respective controls (Student’s t-test). 437
© 2000 Nature America Inc. • http://medicine.nature.com
ARTICLES
a
b
d
© 2000 Nature America Inc. • http://medicine.nature.com
Ratio
c
Fig. 3 Exogenous HSP70 induces the phosphorylation of I-κBα and a rapid calcium flux. a, Human monocytes were pre-treated with the BAPTA-AM or polymyxin B, activated with HSP70 () or LPS (), fixed and permeabilized, treated with phospho-specific I-κBα, stained with FITC-conjugated antibody against rabbit, and analyzed by flow cytometry. Data represent percentage cells staining with phospho-specific-I-κBα in three independent experiments with similar results. b, Human monocytes were pre-treated with culture medium (control; lanes 2, 3 and 6), sodium salicylate (lanes 4 and 7) or BAPTA-AM (lanes 5 and 8), then treated with ovalbumin (lane 2), HSP70 (lanes 3–5) or 100 µg/ml LPS (lanes 6–8); cell extracts were assayed by western blot analysis using phospho-specific I-κBα (Ser32). The bound proteins were visualized by
chemiluminecence. Arrow, molecular weight marker (MW) for phospho-IκBα (Pi-I-κBα). Data represent three independent experiments with similar results. c, Densitometric analysis of western blot in b, of human monocytes pre-treated with culture medium (control; ), sodium salicylate () or BAPTA-AM () followed by treatment with ovalbumin, HSP70 or LPS. Data represent the relative density of phospho-specific I-κBα in three independent experiments. d, Exogenous HSP70 elicits a rapid calcium flux in human monocytes. Human monocytes enriched for CD14+CD45+ were ‘loaded’ with indo-1, then treated with MCP-1 (crosses), 7 nM (triangles) or 0.7 nM (circles) HSP70, or ovalbumin (rectangles; control), and calcium flux was measured. Vertical axis, indo-1 bound to calcium:free indo-1 (ratio).
IL-6 in stably transfected U373–CD14 astrocytoma cells treated with HSP70 (Fig. 4a). Confirmation of involvement of intracellular free calcium in HSP70-induced cytokine production was obtained in experiments in which pre-treatment of U373–CD14 astrocytoma cells with the intracellular calcium chelator BAPTAAM consistently and substantially abrogated HSP70-induced expression of IL-1β, IL-6 and TNF-α (Table 1). Treatment of U373 astrocytoma cells with HSP70 induced the expression of TNF-α, but not of IL-1β or IL-6. BAPTA-AM efficiently abrogated HSP70induced expression of TNF-α in U373 astrocytoma cells. However, ovalbumin, used as a protein control, did not affect cytokine expression in either U373 or U373–CD14 astrocytoma cells (Table 1). These data support a requirement for calcium ions in HSP70-induced cytokine production. A subsequent experiment (Fig. 4b) provided support for data indicative of an association between the putative HSP70 receptor and CD14 in HSP70-induced cytokine expression (Fig. 4a). In this experiment, pre-treatment of peripheral blood mononuclear cells (PBMCs) with antibody against CD14 for 30 minutes at 37 °C resulted in a significant inhibition of HSP70-induced production of IL-6 in supernatant obtained 24 hours after treatment and analyzed by enzyme-linked immunosorbent assay (Fig. 4b). Similar pre-treatment with polymyxin B had no significant effect on HSP70-induced IL-6 production (Fig. 4b). Control antibody alone did not induce IL-6 production in human PBMCs (Fig. 4b). These results strongly indicate an association of CD14 and the putative HSP70 receptor in extracellular heat-shock-protein-induced for cytokine production in human monocytes.
3d), an early event in receptor–ligand binding in many systems9. The finding that HSP70-induced cytokine production was abrogated by pronase treatment (Fig. 2b) formed part of the basis for the hypothesis that HSP70 transduces its signals through a receptor for HSP70. This is in agreement with results showing that incubation of the monocytic cell line P388D and the dendritic cell line D2SC/1 with gold-labeled HSC70, the constitutive form of HSP70, resulted in its appearance within clathrin-coated pits, strongly indicative of the presence of membrane-specific heatshock protein receptors18. Our results corresponded with those findings and further showed that the inducible form of HSP70 bound with high affinity to the surface of human monocytes. Moreover, the functional consequence of specific binding was the stimulation of a potent pro-inflammatory cytokine response and that HSP70 was as effective as LPS in inducing pro-inflammatory cytokine production (Fig. 2). The HSP70-induced flux in intracellular calcium resulted in
U373
OVA HSP70 HSP70 + BAPTA
IL-1β 15 17 19
IL-6 13 14 12
TNF-α 12 74 19
Discussion Our studies have demonstrated a previously unknown function of HSP70 as a cytokine, and addressed the mechanism by which extracellular HSP70 elicits a potent pro-inflammatory immune response in human monocytes. Exogenous HSP70 bound with high specificity to the surface of human monocytes (Fig. 1), eliciting a rapid, dose-dependant flux in intracellular calcium (Fig.
U373–CD14
OVA HSP70 HSP70 + BAPTA
13 53 21
10 48 18
15 62 20
438
Table 1
Involvement of intracellular calcium in HSP70-induced cytokine expression Percentage of cells staining for intracellular cytokinesb
Cells
Treatmenta
a
Cells were pre-treated with culture medium only or BAPTA-AM, then treated with ovalbumin (OVA) or HSP70, then fixed and permeabilized. bCells were stained with FITCconjugated IL-6 or PE-conjugated IL-1β or TNF-α, then analyzed for intracellular cytokine expression by flow cytometry. Data represent percentage cells staining for respective intracellular cytokines (two independent experiments).
NATURE MEDICINE • VOLUME 6 • NUMBER 4 • APRIL 2000
© 2000 Nature America Inc. • http://medicine.nature.com
ARTICLES Fig. 4 Involvement of the CD14 receptor in HSP70-induced cytokine expression. a, U373 human astrocytoma cells transfected with human cDNA for CD14 (U373–CD14) or wild-type U373 cells (U373) were treated with ovalbumin (OVA; control) or HSP70, then fixed and permeabilized, and treated with PE-conjugated antibody against human IL-1β (open bars) or TNF-α () or FITC-conjugated antibody against human IL-6 (), then analyzed by flow cytometry for intracellular cytokine expression. Data represent the mean percentage (± s.d.) of cells expressing cytokines. *, P < 0.01, compared with respective controls (Student’s t-test). b, PBMCs were pre-treated with control antibody (mouse IgG; ), culture medium only (), antibody against CD14 () or polymyxin B ( ), then stimulated with LPS or with HSP70 at a concentration of 0.1 µg/ml (HSP70 (1)) or 0.3 µg/ml (HSP70
(0.3)).The amount of IL-6 released by the cells was determined in supernatants by ELISA. Data represent the mean (± s.d.) IL-6 concentration of quadruplicate samples. *, P < 0.01, compared with respective controls (Student’s t-test).
the subsequent phosphorylation of the intracelluar inhibitory subunit I-κBα (Fig. 3b and c) and the activation of NF-κB and subsequent transcription of pro-inflammatory cytokines. The possibility that HSP70 may be important in various aspects of immune regulation has been addressed19. For example, there are many infectious diseases (such as tuberculosis, malaria and Lyme disease) in which HSP70 is the immunodominant antigen19. Moreover, recognition of a specific epitope of the 65-kDa mycobacterial heat-shock protein by γδ+ T cells20,21 and αβ+ T cells22 results in autoimmune diseases such as adjuvant arthritis23. The isolation from the synovial fluid of patients with rheumatoid arthritis of γδ+ T cells responding to this 65-kDa heat-shock protein indicates that these cells could be relevant to autoimmunity. Indeed, in a rat model of autoimmune arthritis24, T cells generated in rats immunized with Mycobacterium tuberculosis in complete Freund’s adjuvant were able to transfer arthritis to naive rats22. Subsequent analysis of the mycobacterial antigens showed that the 65-kDa heat-shock protein is the arthrogenic component that contains a nine-amino-acid sequence that shares homology with a cartilage-link protein, with four identical amino acids25. In cells expressing CD14 (as seen with ‘professional’ antigenpresenting cells), CD14 was a co-receptor of the putative HSP70 receptor (Table 1 and Fig. 4). This was in agreement with results showing that CD14 is an essential receptor for HSP60-induced IL-6 production in human mononuclear cells26. Here, in response to exogenous HSP70, both the HSP70 receptor and its coreceptor CD14 were activated, resulting in the transduction of at least two separate signals. One signal was transduced through the putative HSP70 receptor and was dependent on intracellular calcium. The other signal was transduced through the CD14 receptor and was also dependent on intracellular calcium. Both signals converged at the phosphorylation of I-κBα and the activation of NF-κB. This was followed by the transcription and upregulation of the pro-inflammatory cytokines TNF-α, IL-6 and IL-1β. In response to LPS stimulation, only the CD14 receptor signal independent of intracellular calcium was activated, resulting in the transcription and release of TNF-α, IL-6 and IL-1β. In the absence of the CD14 receptor (as in U373 ‘wild-type’ cells), stimulation of the putative HSP70 receptor by HSP70 resulted in an induction of TNF-α but not IL-6 or IL-1β, dependent on intra-
cellular calcium (Table 1 and Fig. 4). Future experiments to determine the extent to which CD14 contributes to the cell surface binding of HSP70 are essential. In particular, experiments in which monoclonal antibody against CD14 is pre-incubated with freshly recovered PBMCs and/or U373 (‘wild-type’ and transfected with CD14) before treatment with HSP70 will help answer this question. That experimental design will also provide information on the extent to which CD14 contributes to HSP70-mediated signaling and cytokine production. CD14 is a glycosylphosphatidylinositol-anchored membrane protein, lacking transmembrane and intracellular domains, yet it functions as a signaling receptor for LPS. Although it is not apparent how intracellular signals are generated through CD14, studies indicate involvement of Toll-like receptors 2 and 4 in this primitive activation pathway27–29. Studies are now underway to determine if Toll-like proteins participate in the response of monocytes to heat-shock proteins. Recent findings have shown that LPS transduces pro-inflammatory signal in a manner similar to that of IL-1 receptor signaling30. In those studies, dominant negative mutants of IL-1 signaling pathway intermediates myeloid differentiation marker 88, interleukin-1 receptor-associated kinase (IRAK), IRAK2 and tumor necrosis factor receptor-associated factor (TRAF)6 abrogated LPS- and IL-1-induced activation of NF-κB in human dermal microvessel endothelial cells and THP-1 monocytic cells. Dominant negative mutants of TRAF2 involved in TNF-signaling do not inhibit LPS- or IL-1-induced activation of NF-κB activation, but strongly inhibit TNF-α induced activation of NF-κB (ref. 30). Our results indicated that in the absence of CD14 receptor, as seen in U373 ‘wild-type’ astrocytoma cells, treatment with HSP70 resulted in a signal transduction cascade that involved only the putative HSP70 receptor. This pathway was dependent on intracellular calcium, and possibly activated TRAF2 (bypassing activation of myeloid differentiation marker 88, IRAK, IRAK2 and TRAF6) and converged at an as-yet-unknown transcription factor (possibly inducing weak NF-κB activation) that resulted in the selective transcription and upregulation of TNF-α but not IL1β or IL-6 (Table 1 and Fig. 4). The physiological relevance of this differential cytokine induction is unknown at present. However, it is possible that the effect of HSP70 on CD14– cells, like natural killer cells, cytotoxic T lymphocytes and T-helper cells might be
© 2000 Nature America Inc. • http://medicine.nature.com
NATURE MEDICINE • VOLUME 6 • NUMBER 4 • APRIL 2000
b IL-6 concentration (pg/ml × 10–3)
a
439
© 2000 Nature America Inc. • http://medicine.nature.com
© 2000 Nature America Inc. • http://medicine.nature.com
ARTICLES to propel these cells to produce TNF-α, a powerful apoptogenic cytokine against tumor cells, instead of IL-1β or IL-6, which might enhance proliferation instead. Efforts are now underway test this. Recent work supports a physiological relevant function for exogenous HSP70 (ref. 31). Those results showed that tumor lysate containing HSP70 stimulates immature dendritic cells to increased phagocytosis in vitro, and that expression of HSP70 in B16 tumors results in the accumulation of T lymphocytes, macrophages and dendritic cells into the tumor microenvironment31. Attempts to address this mechanism have shown that tumor-derived HSP70 was able to elicit a potent pro-inflammatory cytokine response in murine mononuclear cells (A.A. and S.K.C., manuscript submitted). Those findings, with the results herein, provide a clue to the potent effect seen for heat-shockprotein-based immunotherapy against certain cancers32. Heatshock protein preparations are effective at immunizing against tumors and inducing potent cytotoxic T lymphocytes in vivo in nanogram rather than microgram amounts33–36. Previous work has clearly and consistently shown that the induction HSP70 gene transcription represses pro-inflammatory gene transcription6,37–39. Heat-shock-factor-1-induced downregulation of the pro-inflammatory process in the cell is mediated by stressful stimuli that would otherwise result in the death of the cells. Our results here did not contradict those findings. Although exogenous HSP70 admixed with human monocytes for 20 minutes at 37 °C resulted in intracellular staining, our results indicate that the increase in intracelluar HSP70 was due to receptor-mediated endocytosis18, because longer incubation of HSP70 at 37 °C resulted in less intense, diffuse staining, indicating that HSP70 was being broken down within the endocytic vesicles. Therefore, unlike the HSP70 generated by stressful stimuli that results in the augmentation of intracellular HSP70 gene expression and subsequent inhibition of pro-inflammatory cellular functions, exogenously added HSP70 bound specifically to a putative HSP70 receptor associated with CD14 receptor complex, transduced a pro-inflammatory signal cascade that resulted in cytokine production, and was internalized and broken down within endocytic vacuoles. By definition, cytokines are proteins secreted by cells with regulatory effects on other cells40. Treatment of human monocytes with a sub-lethal dose of hyperthermia (42 °C for 1 hour) resulted in the release of HSP70 into the extracellular milieu 8–16 hours after treatment (data not shown), indicating the possibility that in certain conditions HSP70 can be secreted from viable immunocompetent cells. Therefore, in addition to its function as an intracellular molecular chaperone, HSP70 in the extracellular milieu acts as a powerful cytokine, affecting the functional properties of immunocompetent cells. This dual role, as both a chaperone and cytokine, helps elucidate recent findings indicating that heat-shock proteins can be potent adjuvants for eliciting immune responses and are powerful inducers of anti-tumor immunity. Methods Isolation and enrichment of human monocytes. PBMCs were isolated from freshly drawn peripheral venous blood (Kraft Family Blood Center, Dana-Farber Cancer Institute) using Ficoll-Paque separation technique as described in detail elsewhere41. Monocytes were enriched from the PBMC fraction by negative depletion using magnetic beads coated with appropriate monoclonal antibodies. With a monocyte isolation kit (Miltenyi Biotec, Aubum, California), PBMCs were treated with Fc receptor (FcR) blocking agent, followed incubation for 5 min at 4 °C with a hapten–antibody ‘cock440
tail’ containing monoclonal hapten-conjugated antibodies against CD3 (mouse Ig2A), CD7 (mouse Ig2A), CD19 (mouse IgG1), CD45RA (mouse IgG1) CD56 (mouse IgG2b) and IgE (mouse IgG2A). Cells were washed twice in phosphate-buffered saline (PBS) and incubated in Fc receptor (FcR) blocking buffer and magnetic cell sorting anti-hapten microbeads. After 5 min of incubation at 4 °C, cells were washed and passed through a column attached to a magnet. Unlabeled monocytes (CD45+CD14+) eluted from the column were more than 98% pure, as judged by flow cytometry. Cell viability was measured by Trypan blue exclusion. HSP70 purity. The endotoxin content of HSP70 was routinely measured by measured by Limulus ambeocyte lysate assay kit (BioWhittaker, Walkersville, Maryland). Confocal microscopy. HSP70 (StressGen, Victoria, British Columbia, Canada) and ovalbumin (Sigma) were biotinylated using an Immunoprobe Biotinylation Kit from Sigma, according to the manufacturer’s directions. Cells were then stained with 70 nM biotinylated HSP70 or 70 nM biotinylated ovalbumin or 70 nM FITC-conjugated streptavidin (Sigma) for 15 min at 4 °C (surface staining) or 37 °C (intracellular staining). For surface staining, cells were washed twice and stained with FITC-conjugated streptavidin, then fixed using 2% paraformaldehyde in PBS for 30 min, then washed twice in PBS alone. For intracellular staining, cells were simultaneously fixed and permeabilized using PermeaFix buffer (OrthoDiagnostics) for 40 min at room temperature in the dark, as described42,43. Cells were then washed twice in PBS and stained with FITC-conjugated streptavidin for 15 min at room temperature in the dark. After both treatment regimes, 5 µl of the cell pellet was mixed with an equal volume of SlowFade (Molecular Probes, Eugene, Oregon) and placed onto a glass slide, and the cover slip was sealed with nail polish. Fluorescence distribution was analyzed using a Zeiss model LSM410 confocal laser scanning microscope (Zeiss, New York, New York) equipped with an external argon-krypton laser (488 and 568 nm). Optical sections of 512 × 512 pixels were digitally recorded within 2 s. Images were printed with a Fujix Pictography 3000 printer (Fuji Film, Fuji, Japan) using Adobe Photoshop software (Adobe Systems, Mountain View, California) on an Apple Macintosh computer. Flow cytometric analysis. Human monocytes (2 × 106 cells) were treated with 100 µg/ml ovalbumin or 7 nM HSP70 for 0, 2 or 4 h, then simultaneously fixed and permeabilized using 2 ml PermeaFix (OrthoDiagnostics, Raritan, New Jersey) for 40 min at room temperature in the dark, as described42,43. Cells were then washed three times in PBS. Nonspecific binding was inhibited by treating cells with 5.5% normal goat serum in PBS for an additional 1 h at room temperature with gentle rocking. For the measurement of intracellular cytokine expression, cells were treated for 40 min at room temperature in the dark with antibodies against human TNF-α (conjugated to phycoerythrin (PE)), IL-6 (conjugated to FITC) or IL-1β (conjugated to PE) (1 µl/1 × 106 cells; Becton Dickinson, Mountain View, California). Cells were then washed twice in PBS and analyzed by flow cytometry. For TNF-α expression analysis, human monocytes were pre-treated with culture medium only, 10 mM pronase (Sigma), 10 mM BAPTA-AM (Calbiochem, La Jolla, California) or 20 mM sodium salicylate for 30 min, then washed and resuspended in culture medium supplemented with 100 µg/ml ovalbumin, 7 nM HSP70 or 100 µg/ml LPS for a further 4 h. The percentage of cells staining for intracellular TNF-α was assessed by flow cytometry. For the measurement of the phosphorylation state of I-κBα, human monocytes were pre-treated with the intracellular calcium chelator BAPTAAM (10 µM) or the LPS inhibitor polymyxin B (0.1 µg/ml), then activated for 30 min with either 7 nM HSP70 or 100 µg/ml LPS. Then, cells were immediately fixed and permeabilized using PermeaFix (OrthoDiagnostics, Raritan, New Jersey), and phospho-specific I-κBα (NEB) at a 1:500 dilution was admixed with the cells for 1 h at room temperature. Cells were then washed three times with PBS (5.5% normal goat serum) and treated with FITC-conjugated antibody against rabbit for 30 minutes. The cells were then washed twice in PBS and analyzed by flow cytometry. For assessing involvement of the CD14 receptor in HSP70-induced cytokine expression, U373 human astrocytoma cells transfected with human cDNA for CD14 or wild-type U373 cells were treated for 4 h with 100 µg/ml ovalbumin (protein control) or 10 µg/ml HSP70. For assessing involvement NATURE MEDICINE • VOLUME 6 • NUMBER 4 • APRIL 2000
© 2000 Nature America Inc. • http://medicine.nature.com
ARTICLES
© 2000 Nature America Inc. • http://medicine.nature.com
of intracellular calcium in HSP70-induced cytokine expression, U373 human astrocytoma cells were pre-treated with culture medium only or 10 µM BAPTA-AM (BAPTA) followed by treatment for 4 h with either 100 µg/ml ovalbumin or 7 nM HSP70. After treatment, cells were simultaneously fixed and permeabilized using PermeaFix buffer (OrthoDiagnostics, Raritan, New Jersey) and treated with PE-conjugated antibody against human IL-1β or TNF-α, or FITC-conjugated antibody against human IL-6 (Becton Dickinson, Mountain View, California), then analyzed by flow cytometry for intracellular cytokine expression. Flow cytometric analysis used a FACScan with a Lysis II software program (Becton Dickinson, Mountain View, California). Individual cells were gated on the basis of forward scatter and orthogonal (side) scatter. The photomultiplier for FITC (FL1-height) or PE (FL2-height) was set on a logarithmic scale. Cell debris was excluded by raising the forward-scatter-height photomultiplier threshold. The flow rate was adjusted to less than 200 cells/s, and at least 10,000 cells were analyzed for each sample. Western blot analysis. Human monocytes were pre-treated with culture medium, 20 mM sodium salicylate or 10 µM BAPTA-AM, then treated with 100 µg/ml ovalbumin, 7 nM HSP70 or 100 µg/ml LPS. Cell extracts were assayed by western blot analysis using phospho-specific I-κBα (Ser32) as described in the manufacturer’s handbook (NEB). Proteins were extracted from cells after treatment with ice-cold RIPA buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% NP-40, 2.5% deoxycholate, 2 mM EGTA, 1 mM leupeptin, 1 mM aprotinin, 10 mM NaF and 1 mM PMSF), and samples were cleared by centrifugation at 15,000g for 20 min at 4 °C. For immunoprecipitation experiments, supernatant was carefully removed and incubated with primary antibody for 1–2 h on ice, and immunoprecipitated samples were collected for 30 min at 4 °C with protein A–Sepharose beads (Pharmacia Biotech, Piscataway, New Jersey). The precipitate was washed three times with phosphate-buffered saline (PBS) and boiled in SDS–PAGE sample buffer before being separated by 12% SDS–PAGE and transferred to Immobilon PVDF membranes (Millipore, Bedford, Massachusetts). The membrane was blocked by incubation in Tris-buffered saline buffer supplemented with 5% nonfat dry milk (Bio-Rad Laboratories, Hercules, California) and 0.1% Tween-20 for 1 h at room temperature, washed three times in Tris-buffered saline buffer and incubated overnight with the appropriate primary antibody at 4 °C. Membranes were washed three times with Tris-buffered saline buffer and incubated with the appropriate alkaline phosphate-conjugated secondary antibody for 1 h at room temperature. Detection of proteins was achieved by the ECL system. Densitometric analysis of western blot was also done, where indicated. Calcium flux measurements. Calcium flux measurements were made by a method described in detail elsewhere44. Monocytes were isolated from freshly drawn peripheral venous blood and enriched for CD14+CD45+ monocytes by a magnetic bead separation technique. They were then incubated for 45 min at 37 °C in the dark at a concentration of 2 × 106 cells/ml in 40 ml RPMI supplemented with 1 mM indo-1 acetoxymethyl dye (Molecular Probes, Eugene, Oregon). ‘Indo-1-loaded’ cells (1 × 106) were treated with 100 µM MCP-1, 7 nM HSP70, 0.7 nM HSP70 or 100 µg/ml ovalbumin (control). Cells were then washed and resuspended in fresh culture medium for an additional 45 min at 37 °C in the dark. An increase in intracellular calcium was detected with indo-1 as an increase in the ratio of lower-emission wavelength to higher-emission wavelength (405-nm emission/485-nm emission), which is proportional to the internal calcium level in the cell. As no standard curve was determined to correlate with the indo-1 ratio, the data are expressed as relative intracellular calcium. Calcium flux was analyzed over 3–5 min using a FACS analyzer Becton Dickinson, Mountain View, California) and collected as a continuous file into Consort 40/µ VAX computer system through FACS Data Lister (Becton Dickinson, Mountain View, California). Mean intracellular calcium time course data was showed using KINPRO Software (Becton Dickinson, Mountain View, California). ELISA. Peripheral blood mononuclear cells (1 × 107 cells/well) were pretreated with either 5 µg/ml control antibody (mouse IgG), culture medium only, 5 µg/ml antibody against CD14 (Coulter, Hialeah, Florida) or 0.1 µg/ml polymyxin B. Cells were then stimulated with either 10 ng/ml LPS or 0.1 µg/ml or 0.3 µg/ml HSP70. After 24 h, supernatant from each well was NATURE MEDICINE • VOLUME 6 • NUMBER 4 • APRIL 2000
recovered and the amount of IL-6 released by the cells determined by ELISA according to the manufacturer instructions (Endogen, Woburn, Massachusetts). Acknowledgments We thank P.E. Auron (Harvard Institute of Medicine) for review of the manuscript; T. Blake and P. Fischer (Merck Research Laboratories) and L. Popova (Dana-Farber Cancer Institute) for technical assistance; and The Kraft Family Blood Center (Dana-Farber Cancer Institute) for the freshly drawn peripheral venous blood. This work was supported by National Institutes of Health Grants CA47407, CA31303, CA50642, CA77465 (to S.K.C.) and in part by UNCF/Merck Science Initiative (to A.A.)
RECEIVED 25 OCTOBER 1999; ACCEPTED 1 FEBRUARY 2000 1. Craig, E.A. & Gross, C.A. Is hsp70 the cellular thermometer? Trends Biochem. Sci. 16, 135–40 (1991). 2. Lindquist, S. & Craig, E.A. The heat-shock proteins. Annu. Rev. Genet. 22, 631–77 (1988). 3. Calderwood, S.K. in Proceedings of the 86th Annual Meeting of the American Association for Cancer Research 682, (American Association for Cancer Research, Philadelphia, Pennsylvania, 1995). 4. Minota, S., Cameron, B., Welch, W. J. & Winfield, J. B. Autoantibodies to the constitutive 73-kD member of the hsp70 family of heat shock proteins in systemic lupus erythematosus. J. Exp. Med. 168, 1475–1480 (1988). 5. Schletter, J., Heine, H., Ulmer, A.J. & Rietschel, E. T. Molecular mechanisms of endotoxin activity. Arch. Microbiol. 164, 383–389 (1995). 6. Housby, J.N. et al. Non-steroidal anti-inflammatory drugs inhibit the expression of cytokines and induce hsp70 in human monocytes. Cytokine 11, 347–358 (1999). 7. Golenbock, D.T., Hampton, R.Y., Qureshi, N., Takayama, K. & Raetz, C.R. Lipid Alike molecules that antagonize the effects of endotoxins on human monocytes. J. Biol. Chem. 266, 19490–19498 (1991). 8. Duff, G. W. & Atkins, E. The inhibitory effect of polymyxin B on endotoxin-induced endogenous pyrogen production. J. Immunol. Methods 52, 333–340 (1982). 9. Ghosh, S., May, M.J. & Kopp, E.B. NF-kappa B and Rel proteins: evolutionarily conserved mediators of immune responses. Annu. Rev. Immunol. 16, 225–260 (1998). 10. Baeuerle, P.A. & Baltimore, D. I kappa B: a specific inhibitor of the NF-kappa B transcription factor. Science 242, 540–546 (1988). 11. Beg, A.A. & Baldwin, A.S.Jr. The I kappa B proteins: multifunctional regulators of Rel/NF-kappa B transcription factors. Genes Dev. 7, 2064–2070 (1993). 12. Stevenson, M.A., Zhao, M.-J., Asea, A., Coleman, N.C. & Calderwood, S.K. Salicylic acid and asprin inhibit the activity of RSK2 kinase and repress RSK2-dependent transcription of CREB and NF-κB responsive genes. J. Immunol. 163, 5608–5616 (1999). 13. Rollins, B.J., Walz, A. & Baggiolini, M. Recombinant human MCP-1/JE induces chemotaxis, calcium flux, and the respiratory burst in human monocytes. Blood 78, 1112–1116 (1991). 14. McLeish, K.R., Dean, W.L., Wellhausen, S.R. & Stelzer, G.T. Role of intracellular calcium in priming of human peripheral blood monocytes by bacterial lipopolysaccharide. Inflammation 13, 681–692 (1989). 15. Ulevitch, R.J. & Tobias, P.S. Recognition of endotoxin by cells leading to transmembrane signaling. Curr. Opin. Immunol. 6, 125–130 (1994). 16. Tapping, R.I., Orr, S.L., Lawson, E.M., Soldau, K. & Tobias, P.S. Membrane-anchored forms of lipopolysaccharide (LPS)-binding protein do not mediate cellular responses to LPS independently of CD14. J. Immunol. 162, 5483–5489 (1999). 17. Solomon, K.R. et al. Heterotrimeric G proteins physically associated with the lipopolysaccharide receptor CD14 modulate both in vivo and in vitro responses to lipopolysaccharide. J. Clin. Invest. 102, 2019–2027 (1998). 18. Arnold-Schild, D. et al. Receptor-mediated endocytosis of heat shock proteins by professional antigen-presenting cells. J. Immunol. 162, 3757–3760 (1999). 19. Kaufmann, S.H.E. & Schoel, B. in The Biology of Heat Shock Proteins and Molecular Chaperones (eds. Morimoto, R. I., Tissieres, A. & Georgopoulos, C.) 495–531 (Cold Spring Harbor Laboratory, Plainview, New York, 1994). 20. Haregewoin, A., Soman, G., Hom, R.C. & Finberg, R.W. Human gamma delta+ T cells respond to mycobacterial heat-shock protein. Nature 340, 309–312 (1989). 21. Haregewoin, A., Singh, B., Gupta, R.S. & Finberg, R.W. A mycobacterial heat-shock protein-responsive gamma delta T cell clone also responds to the homologous human heat-shock protein: a possible link between infection and autoimmunity. J. Infect. Dis. 163, 156-160 (1991). 22. van Eden, W. et al. Cloning of the mycobacterial epitope recognized by T lymphocytes in adjuvant arthritis. Nature 331, 171–173 (1988). 23. Holoshitz, J., Koning, F., Coligan, J.E., De Bruyn, J. & Strober, S. Isolation of CD4CD8- mycobacteria-reactive T lymphocyte clones from rheumatoid arthritis synovial fluid. Nature 339, 226–229 (1989). 24. Holoshitz, J., Matitiau, A. & Cohen, I.R. Arthritis induced in rats by cloned T lymphocytes responsive to mycobacteria but not to collagen type II. J. Clin. Invest. 73, 211–215 (1984). 25. van Eden, W. et al. Heat-shock protein T-cell epitopes trigger a spreading regulatory control in a diversified arthritogenic T-cell response. Immunol. Rev. 164, 169–174 (1998). 26. Kol, A., Lichtman, A.H., Finberg, R.W., Libby, P. & Kurt-Jones, E.A. Cutting edge:
441
© 2000 Nature America Inc. • http://medicine.nature.com
ARTICLES 35. Srivastava, P.K. & Udono, H. Heat shock protein-peptide complexes in cancer immunotherapy. Curr. Opin. Immunol. 6, 728–732 (1994). 36. Tamura, Y., Peng, P., Liu, K., Daou, M. & Srivastava, P.K. Immunotherapy of tumors with autologous tumor-derived heat shock protein preparations. Science 278, 117–120 (1997). 37. Soncin, F. & Calderwood, S.K. Reciprocal effects of pro-inflammatory stimuli and anti-inflammatory drugs on the activity of heat shock factor-1 in human monocytes. Biochem. Biophys. Res. Commun. 229, 479–484 (1996). 38. Sistonen, L., Sarge, K.D. & Morimoto, R.I. Human heat shock factors 1 and 2 are differentially activated and can synergistically induce hsp70 gene transcription. Mol. Cell Biol. 14, 2087–2099 (1994). 39. Baler, R., Zou, J. & Voellmy, R. Evidence for a role of Hsp70 in the regulation of the heat shock response in mammalian cells. Cell Stress Chaperones 1, 33–39 (1996). 40. Janeway, C.A. & Travers, P. Immunobiology: The Immune System in Health and Disease (eds. Janeway, C.A. & Travers, P.) (Garland Publishing, New York, 1997). 41. Asea, A. Role of Histamine in the Regulation of Natural Killer Cells. Doctoral dissertation, Univ. Göteborg (Göteborg, Sweden, 1995). 42. Asea, A. et al. Histaminergic regulation of interferon-gamma (IFN-gamma) production by human natural killer (NK) cells. Clin. Exp. Immunol. 105, 376–382 (1996). 43. Hansson, M., Asea, A., Ersson, U., Hermodsson, S. & Hellstrand, K. Induction of apoptosis in NK cells by monocyte-derived reactive oxygen metabolites. J. Immunol. 156, 42–47 (1996). 44. Koo, G.C. et al. Association of serine protease with the rise of intracellular calcium in cytotoxic T lymphocytes. Cell. Immunol. 174, 107–115 (1996).
© 2000 Nature America Inc. • http://medicine.nature.com
Heat shock protein (HSP) 60 activates the innate immune response: CD14 is an essential receptor for HSP60 activation of mononuclear cells. J. Immunol. 164, 13–17 (2000). 27. Yang, R. B. et al. Toll-like receptor-2 mediates lipopolysaccharide-induced cellular signalling. Nature 395, 284–248 (1998). 28. Poltorak, A. et al. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 282, 2085–2088 (1998). 29. Hoshino, K. et al. Cutting edge: Toll-like receptor 4 (TLR4)-deficient mice Are hyporesponsive to lipopolysaccharide: evidence for TLR4 as the LPS gene oroduct. J. Immunol. 162, 3749–3752 (1999). 30. Zhang, F.X. et al. Bacterial lipopolysaccharide activates nuclear factor-kappaB through interleukin-1 signaling mediators in cultured human dermal endothelial cells and mononuclear phagocytes. J. Biol. Chem. 274, 7611–7614 (1999). 31. Todryk, S. et al. Heat shock protein 70 induced during tumor cell killing induces Th1 cytokines and targets immature dendritic cell precursors to enhance antigen uptake. J. Immunol. 163, 1398–1408 (1999). 32. Suto, R. & Srivastava, P.K. A mechanism for the specific immunogenicity of heat shock protein- chaperoned peptides. Science 269, 1585–1588 (1995). 33. Srivastava, P.K., Udono, H., Blachere, N.E. & Li, Z. Heat shock proteins transfer peptides during antigen processing and CTL priming. Immunogenetics 39, 93–98 (1994). 34. Srivastava, P.K., Menoret, A., Basu, S., Binder, R.J. & McQuade, K.L. Heat shock proteins come of age: primitive functions acquire new roles in an adaptive world. Immunity 8, 657–665 (1998).
442
NATURE MEDICINE • VOLUME 6 • NUMBER 4 • APRIL 2000