THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 287, NO. 15, pp. 12387–12394, April 6, 2012 © 2012 by The American Society for Biochemistry and Molecular Biology, Inc. Published in the U.S.A.

Inducible Heat Shock Protein 70 Reduces T Cell Responses and Stimulatory Capacity of Monocyte-derived Dendritic Cells*□ S

Received for publication, September 25, 2011, and in revised form, January 29, 2012 Published, JBC Papers in Press, February 13, 2012, DOI 10.1074/jbc.M111.307579

Pawel Stocki1, Xiao N. Wang2, and Anne M. Dickinson2 From the Department of Haematological Sciences, Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne NE2 4HH, United Kingdom Background: An increased presence of Hsp70 has been observed in many clinical conditions; however, its role remains elusive. Results: Hsp70 reduces the stimulatory capacity of dendritic cells and T cell responses in an in vitro experimental setup. Conclusion: Hsp70 might moderate inflammatory reactions mediated by the immune system. Significance: Provides additional insight into the role of Hsp70 in autoimmune diseases, transplantation, and cancer. Heat shock protein 70 (Hsp70) has gained a lot of attention in the past decade due to its potential immunoregulatory functions. Some of the described proinflammatory functions of Hsp70 became controversial as they were based on recombinant Hsp70 proteins specimens, which were later shown to be endotoxin-contaminated. In this study we used low endotoxin inducible Hsp70 (also known as Hsp72, HSPA1A), and we observed that after a 24-h incubation of monocyte-derived immature dendritic cells (mo-iDCs) with 20 ␮g/ml of low endotoxin Hsp70, their ability to stimulate allogenic T cells was reduced. Interestingly, low endotoxin Hsp70 also significantly reduced T cell responses when they were simulated with either IL-2 or phytohemagglutinin, therefore showing that Hsp70 could alter T cell responses independently from its effect on mo-iDCs. We also reported a greater response of Hsp70 treatment when activated versus nonactivated T cells were used. This effect of Hsp70 was similar for all tested populations of T cells that included CD3ⴙ, CD4ⴙ, or CD8ⴙ. Taken together, our observations strongly suggest that Hsp70 might dampen, rather than provoke, T cell-mediated inflammatory reactions in many clinical conditions where up-regulation of Hsp70 is observed.

Heat shock proteins (HSPs)3 have been studied as potential antigenic molecules that can induce a specific antitumor immune response. In 1993 Srivastava and colleagues showed that the two most extensively studied HSPs, Hsp70 and gp96,

* This work was supported by Leukaemia Research Fund, United Kingdom, Grant 04070; a grant from the Tyneside Leukaemia Research Fund, United Kingdom; and European Commission “Trans-Net” Grant MRTN-CT-2004-512253. □ S This article contains supplemental Figs. 1 and 2. 1 To whom correspondence should be addressed: Dept. of Biochemistry, Medical Sciences Bldg., University of Toronto, 1 King’s College Circle, Toronto, ON M5S 1A8, Canada. Tel.: 1-416-9786034; E-mail: pawel.stocki@ utoronto.ca. 2 Both authors contributed equally to this work. 3 The abbreviations used are: HSP, heat shock protein; APC, antigen-presenting cell; DC, dendritic cell; EU, enzyme unit(s); IU, international unit(s); mo-DC, monocyte-derived DC; mo-iDC, monocyte-derived immature dendritic cell; PBMC, peripheral blood mononuclear cell; PHA, phytohemagglutinin; TLR, Toll-like receptor.

APRIL 6, 2012 • VOLUME 287 • NUMBER 15

both acted as antigenic peptide carriers but were not antigenic themselves (1, 2). These HSPs were shown to be potent antigenic peptide carriers (3, 4), which when isolated from a tumor can deliver transported antigenic peptides to professional antigen-presenting cells (APCs) and prime CD8⫹ T cells against the same tumor from which the HSPs were purified (5). The peptide delivery by Hsp70 to APCs was described as being mediated by CD91 or/and LOX-1 receptors (6 – 8). Controversially, in addition to its role in the antigen crosspresentation process, Hsp70 at a concentration as low as 0.5 ␮g/ml has been described to be a potent proinflammatory cytokine and to be involved in the activation of innate immunity (9). The Hsp70 receptors responsible for this activation have been proposed to be CD14 (9) and Toll-like receptors (TLRs) (10, 11), which also mediate activation of innate immunity caused by microbial endotoxins. The shared activation pathway as well as the similar activation profile for endotoxins and Hsp70 led to speculation that the observed cytokine-like property of Hsp70 may have been incorrectly identified. Accordingly, it was shown that the endotoxin contamination in recombinant Hsp70 is common and exclusively responsible for the described proinflammatory properties, including activation of APCs (12–16). It was also shown that endotoxin-free Hsp70 did not induce APC activation while still effectively delivered antigenic peptides for cross-presentation (17). In a different study, Hsp70 failed to bind to null cells overexpressing either CD14, TLR2, or TLR4 which had previously been associated with the Hsp70-mediated activation of innate immunity (8). Hsp70 was also shown to be involved in lipopolysaccharide (LPS) signal transduction (18, 19), suggesting that even a negligible amount of microbial endotoxin in combination with Hsp70 might potentially be amplified and subsequently responsible for the activation of innate immunity. Conversely, in some studies where Hsp70 was purified from a nonbacterial origin, excluding the possibility of microbial contamination, it was shown that Hsp70 could actually stimulate macrophages to secrete nitric oxide (20, 21). Hsp70 purified from a nonbacterial origin at concentrations as high as 400 ␮g/ml was shown to have very moderate and only partial stimJOURNAL OF BIOLOGICAL CHEMISTRY

12387

Hsp70 Abrogates T Cell Responses ulatory capability on dendritic cells (DCs), which did not resemble that of endotoxins (20). Recently low endotoxin Hsp70 was shown to protect neuroblastoma cells against heat stress, stimulated intracellular calcium flux in monocytes, and moderately induced mouse splenocyte cytokine secretion and the selective proliferation of particular subpopulations of CD4 and CD11, but not CD8 cells (22). In a different study Hsp70 was shown to induce immune responses of CD4 and CD8 cells upon extended incubation with Hsp70 (23). In this study, we examined the properties of low endotoxin Hsp70 on monocyte-derived immature DCs (mo-iDCs) as well as purified T cell populations. Interestingly, we observed that in our in vitro experimental model low endotoxin specimen of Hsp70 reduced the stimulatory capacity of mo-iDCs and abrogated T cell proliferation independently from used stimuli, which included either mo-iDCs, phytohemagglutinin (PHA), or IL-2. This novel property of Hsp70 may advance our understanding of immunological mechanisms associated with inflammation and tumor development, which are directly associated with the up-regulation of Hsp70.

EXPERIMENTAL PROCEDURES Cell Culture Conditions—Cells were cultured in RPMI 1640 medium (Invitrogen) containing 100 IU/ml penicillin, 100 ␮g/ml streptomycin (Invitrogen), and 2 mM L-glutamine (Invitrogen) supplemented with 10% human AB serum (PAA Laboratories) unless otherwise indicated. All cultures were incubated at 37 °C in a humidified 5% CO2 in air incubator (IR 1500; Flow Laboratories). Before collecting human blood samples, informed consent was obtained, and the local ethics committee approved the project. Generation of Monocyte-derived Immature DCs—Peripheral Blood Mononuclear Cells (PBMCs) were prepared from the blood of healthy donors by density gradient centrifugation on Lymphoprep (Axis-Shield). Monocyte-derived DCs (mo-DCs) were prepared as previously described (24). MACS technology (Miltenyi Biotec) was used to isolate monocytes from PBMCs. Purified monocytes were then grown in 24-well plates (Greiner Bio-One) in the serum-free medium, CellGro DC (CellGenix), at a concentration of 5 ⫻ 105/well (1 ml). To differentiate the monocytes into mo-DCs, 50 ng/ml granulocyte macrophage colony-stimulating factor (GM-CSF) (Immunotools) and 50 ng/ml IL-4 (Immunotools) were added to the culture. After 3 days the medium and cytokines were refreshed and cultured for additional 3 days; these cells were then referred to as immature mo-DCs (mo-iDCs). The level of endotoxin contamination in the cytokines according to the manufacturer: GM-CSF ⬍0.1 EU/␮g; IL-4 ⬍ 0.1 EU/␮g. mo-iDC Phenotype and Morphology—Hsp70 from three different sources was tested for its influence on maturation of mo-iDCs. Two Hsp70 samples produced in Escherichia coli were acquired from Assay Design (previously StressGen, currently Enzo Life Sciences) which according to the manufacturer had an endotoxin content of either ⬎500 EU/mg or ⬍50 EU/mg, catalogue number NSP-555 or ESP-555, respectively. Hsp70 with a low endotoxin contamination was purified and kindly provided by Britta Eiz-Vesper (23). As described previously, a low endotoxin Hsp70 sample was purified from the

12388 JOURNAL OF BIOLOGICAL CHEMISTRY

HEK293 human cell line and without any further endotoxin removal tested for endotoxin contamination using Limulus amebocyte lysate assay (Lonza) indicating endotoxin content of ⬃0.32 EU/mg (23). These Hsp70 samples were incubated at different concentrations with 2 ⫻ 105 mo-iDCs/well in 96-well, round-bottomed plates for 24 h. The cells were in the serumfree medium, CellGro DC (200 ␮l/well). As a positive control, LPS (Sigma-Aldrich) at a concentration of 1 ␮g/ml was used. After 24-h incubation the cell morphology was acquired using light microscopy as well as flow cytometry, FACSCalibur (BD Biosciences). Nonviable cells were represented by the characteristic change in their size and granularity visible using forward and side scatter plots. Flow cytometry was also used to test the expression of CD14 (BD Biosciences, clone rmC5–3), CD1a (Dako, clone NA1/34), CD80 (BD Biosciences, clone L307.4), CD86 (BD Biosciences, clone 2331:FUN-1), CD83 (BD Biosciences, clone HB15e), and HLA-DR (BD Biosciences, clone L243). Isotype control was mouse IgG FITC- or phycoerythrinconjugated (BD Biosciences, clone X40). For apoptosis detection the TACS Annexin-V FITC detection kit (R&D Systems) was used according to the manufacturer’s instructions. For each test 10,000 events were acquired, and data were analyzed using FlowJo software (Tree Star). Generation of Activated T Cells—Healthy donors were serologically typed for HLA-A*03 (Abcam, clone 4i153); T cells from HLA-A*03-positive donors were isolated using a T cell enrichment mixture (Stemcell Technologies). For stimulation of T cells, either an influenza antigen (ILRGSVAHK) or Pmel17/gp100 (ALLAVGATK) specific to HLA-A*03 was used and commercially synthesized at a purity of ⬎95% (GenScript). Autologous mo-iDCs were mixed with the T cells at a 1:10 ratio in the presence of 10 ␮M peptide. The cells were plated at a concentration of 0.7 ⫻ 106/ml in 6-well plates with addition of 0.5 IU/ml IL-2 (Roche Applied Science). After 2 days of incubation 20 IU/ml IL-2 and 5 ␮M peptide were added to the culture and further incubated for 8 days. The cells were then restimulated every 7 days with feeder cells (autologous PBMCs primed with 10 ␮M peptide overnight and then irradiated with 20 grays). The T cells were mixed with feeder cells at a 1:1 ratio in the presence of 10 ␮M peptide and 20 IU/ml IL-2. After 7 weeks the cells were tested using flow cytometry for the expression of CD25 (BD Biosciences, clone 2A3) in either CD8 (BD Biosciences, clone SK1) or CD4 (BD Biosciences, clone SK3) population. The activated T cell expression profile was compared with a freshly enriched T cell population, and for each test 100,000 events were acquired. T Cell Proliferation and IFN␥ Production—Either freshly isolated T cells (nonactivated T cells) or activated T cells were stimulated with mo-iDCs at a ratio of 10:1 (0.5 ⫻ 105 T cells versus 0.5 ⫻ 104 mo-iDCs). CD3, CD8, or CD4 cells were enriched from whole blood using CD3, CD8, or CD4 enrichment cocktails by negative selection (Stemcell Technologies). Where indicated mo-iDCs were preincubated with the low endotoxin Hsp70 at different concentrations for 24 h before being used in the proliferation assay. The preincubation of 2 ⫻ 105 mo-iDC/well in 96-well, round-bottomed plates was performed in the serum-free medium, CellGro DCs (200 ␮l/well), after preincubation the cells were washed with an excess of PBS. VOLUME 287 • NUMBER 15 • APRIL 6, 2012

Hsp70 Abrogates T Cell Responses

FIGURE 1. Adjuvant properties of Hsp70 depend on endotoxin contamination level. Hsp70 samples with different endotoxin contents of 0.32, ⬍50, or ⬎500 EU/mg were tested for their activatory potential. mo-iDCs were incubated with the different samples of Hsp70 for 24 h and then monitored for their expression of CD80 (A), CD86 (B), HLA-DR (C), and CD83 (D) molecules, which were used as maturation markers. LPS was used as a positive control at a concentration of 1 ␮g/ml. Data are presented as fold increase in mean fluorescence compared with untreated control ⫾S.D. (error bars). n ⫽ 3 for 0, 5, 10, and 20 ␮g/ml concentrations of Hsp70 samples, and n ⫽ 1 for LPS and the remaining concentrations of Hsp70.

Where no preincubation was performed, Hsp70 at different concentrations was added upon setup of the proliferation assay. Alternatively, T cell stimulation was performed with 5 ␮g/ml PHA (Remel) or 40 IU/ml IL-2. The proliferation assay was performed on 96-well round-bottomed plates (200 ␮l/well). When mo-iDCs were used for stimulation, the assay lasted for either 1 or 4 days, for activated or nonactivated T cells, respectively; when PHA or IL-2 was used for stimulation, the incubation for either activated or nonactivated T cells was 1 day only (in these assays the activated T cells used were 7 days after their last restimulation). Then 50 ␮l of the medium was collected for IFN␥ analysis before being pulsed with [3H]thymidine (TRA310; GE Healthcare) at a concentration of 0.037 MBq/ml. The cells were further incubated for 16 h, and [3H]thymidine incorporation of proliferating cells was determined using a ␤-counter (Matrix 9600; Packard Instrument). Each experiment was performed in triplicate. IFN␥ secretion was analyzed using a cytometric bead array (BD Biosciences) and analyzed using the BD-FCAP Array software (BD Biosciences). Statistical Analysis—Statistical significance was assessed with the two-tailed, paired Student t test. At least three repeats of each experiment were performed, and a p value ⬍ 0.05 was considered significant. For the statistical analysis SPSS version 14.0 was used (SPSS).

RESULTS Endotoxin Contamination in Hsp70 Samples Is Responsible for Maturation of mo-iDCs—The mo-iDC generation protocol used in this study produced a cell population that was shown to be ⬎90% both positive for CD1a and negative for CD14 (supplemental Fig. 1A). mo-iDCs are known to respond to endoAPRIL 6, 2012 • VOLUME 287 • NUMBER 15

toxin contaminants inducing their maturation. Their maturation can be monitored by measuring the expression of certain molecules such as CD80, CD86, CD83, or HLA class II. Accordingly, mo-iDCs can be used as an indicative tool for the detection of endotoxins in protein samples. Three Hsp70 samples with different endotoxin content were incubated with mo-iDCs for 24 h and tested for their expression of CD80, CD86, HLADR, and CD83 (Fig. 1). Nonspecific antibody binding to moiDCs was assessed using isotype control staining and confirmed the specificity of the antibodies used (supplemental Fig. 1B). The concentration of Hsp70 ranged from 1 to 80 ␮g/ml whereas LPS, which was used as a positive control, was at a concentration of 1 ␮g/ml (Fig. 1A). Commercially available Hsp70 samples that contained either high (⬎500 EU/mg) or medium (⬍50 EU/mg) levels of endotoxin induced the maturation of mo-iDCs. The Hsp70 sample with high endotoxin content was ⬃10-fold more potent than the Hsp70 with a medium level of endotoxin (Fig. 1). These two commercially available Hsp70 samples were shown to have adjuvant properties on mo-iDCs; however the low endotoxin Hsp70 of ⬃0.32 EU/mg did not induce the maturation even at a concentration as high as 80 ␮g/ml (Fig. 1). Interestingly, it was observed that the cells incubated with low endotoxin Hsp70 had modestly reduced expression of CD80 and CD86 compared with the untreated control (Fig. 1, A and B). Only the low endotoxin sample of Hsp70 was included in further investigation because it is known that the endotoxin contamination in Hsp70 samples can independently stimulate T cell responses, e.g. when induced with PHA (15), and consequently produce misleading results. JOURNAL OF BIOLOGICAL CHEMISTRY

12389

Hsp70 Abrogates T Cell Responses

FIGURE 2. Morphological changes of mo-iDCs after treatment with low endotoxin Hsp70. A, either Hsp70 at a concentration of 20 ␮g/ml or LPS at 1 ␮g/ml was incubated with mo-iDCs for 24 h and compared with untreated control for their morphological changes using FACS, forward (FSC) and side (SSC) scatter plots. The observed changes were associated with the cell viability. Accordingly, percentages of apoptotic and live cells are presented in lower left and lower right sections of the plots, respectively. B, combined mean percentage results ⫾S.D. (error bars) for mo-iDC viability as assessed using FSC and SSC analysis method. C, morphology of untreated mo-iDCs compared with mo-iDCs incubated for 24 h with 20 ␮g/ml Hsp70 (0.32 EU/ml) using light microscopy. *, statistical significance of the results (n ⫽ 5, p ⬍ 0.05, t test).

Low Endotoxin Hsp70 Influences Morphology of mo-iDCs— Although the expression of tested molecules of mo-iDCs after low endotoxin Hsp70 treatment was only marginally abrogated, the morphology of the mo-iDCs incubated with the low endotoxin Hsp70 was visibly affected (Fig. 2). An incubation of Hsp70 with the cells resulted in a visibly lower number of clusters in the culture compared with the untreated control (Fig. 2C). The Hsp70-treated mo-iDCs were distributed evenly at the bottom of the U-shaped well whereas the spread of the control cells was distinctly different and created cell clusters. We observed that mo-DCs clustering was related to the maturation level; the more mature the cells become the more clusters were observed in the culture. Newly generated mo-iDCs, before being used for the experiments, showed a small amount of clusters which were increasing in number with time, even when unstimulated (data not shown). Addition of Hsp70 to the culture seemed to reduce this spontaneous maturation process consequently affecting the cell morphology, by decreased cluster formation, as well as the expression profile of maturation markers (Figs. 1 and 2). However, mo-iDCs incubated with the commercially available Hsp70, containing medium or high endotoxin levels, had an opposite effect on the maturation marker expression (Fig. 1) and the morphology of the cells (data not shown). Low endotoxin Hsp70-treated mo-iDCs were observed to be more viable than the untreated cells as assessed using flow cytometry (Fig. 2, A and B). Forward and side scatter were used to determine the morphology of the cells, and the observed changes in size and granularity were characteristic of nonviable cells as assessed with Annexin-V staining (supplemental Fig. 1C). The survival advan-

12390 JOURNAL OF BIOLOGICAL CHEMISTRY

tage of the Hsp70-treated cells was ⬃20% (Fig. 2, A and B). In contrast, LPS treated mo-iDCs showed a decreased number of viable cells by ⬎15% compared with the untreated control and by ⬎25% compared with Hsp70-treated cells (Fig. 2, A and B). Low Endotoxin Hsp70 Reduces Stimulatory Capacity of mo-iDCs—To assess the effect of low endotoxin Hsp70 on moiDC functional characteristics we investigated whether Hsp70treated mo-iDCs would show altered stimulatory capacity of either activated or freshly isolated nonactivated T cells. Activated T cells were generated using an enriched population of CD3⫹ cells, which were primed against an antigen. The flow cytometric analysis showed that the activated T cells increased in size and in cytoplasmic granularity compared with nonactivated T cells (supplemental Fig. 2). The population of either activated or nonactivated T cells contained CD8⫹ and CD4⫹ cells which were tested for their expression of the activation marker, CD25. Activated T cell populations contained 97% of CD8⫹ cells positive for CD25, whereas nonactivated T cells contained only CD8⫹ cells negative for CD25. When the population of CD4⫹ cells was analyzed for the expression of CD25, it revealed that 69.5% of nonactivated T cells were negative for CD25 and 30.3% were positive. The activated T cell population of CD4⫹ cells was 98.2% positive for CD25 (supplemental Fig. 2). In further experiments either activated or nonactivated T cells were used and tested for their response against untreated or Hsp70 preincubated mo-iDCs. After preincubation with Hsp70, mo-iDCs were washed with PBS to remove the Hsp70 remaining in the medium; the control cells were treated the VOLUME 287 • NUMBER 15 • APRIL 6, 2012

Hsp70 Abrogates T Cell Responses

FIGURE 3. Preincubation with low endotoxin Hsp70 reduces the stimulatory capacity of allogenic mo-iDCs. Allogenic mo-iDCs were preincubated with Hsp70 at concentrations of 1, 10, or 20 ␮g/ml for 24 h. The cells were then harvested and washed with PBS to remove Hsp70 before being used as stimulatory cells for activated and nonactivated T cells. A and B, proliferation rate (A) and IFN␥ secretion (B) by activated T cells upon stimulation with Hsp70pretreated allogenic mo-iDCs. C and D, proliferation rate (C) and IFN␥ secretion (D) by nonactivated CD3, CD8, or CD4 T cells upon stimulation with Hsp70-pretreated mo-iDCs. Data are presented as mean percentage change relative to the untreated control ⫾S.D. (error bars).*, statistical significance of the results (n ⫽ 5, p ⬍ 0.05, t test).

FIGURE 4. Low endotoxin Hsp70 reduces T cell response when stimulated with allogenic mo-iDCs. Either activated or nonactivated T cells were stimulated with allogenic mo-iDCs that were cultured without the presence of Hsp70. A and B, proliferation rate (A) and IFN␥ secretion (B) by activated T cells upon allogenic mo-iDCs stimulation. Hsp70 at concentrations of 1, 10, or 20 ␮g/ml was added to the cell mixture upon setup of the assays. C and D, proliferation rate (C) and IFN␥ secretion (D) by nonactivated CD3, CD8, or CD4 T cells upon stimulation with allogenic mo-iDCs in the absence or presence of Hsp70 at a concentration of 20 ␮g/ml which was added to the cell mixture upon the setup of the assays. Data are presented as mean percentage change relative to the untreated control ⫾S.D. (error bars). *, statistical significance of the results (n ⫽ 4, p ⬍ 0.05, t test).

same way as those preincubated with Hsp70. Preincubation of mo-iDCs with Hsp70 for 24 h altered their stimulatory capacity as observed by analyzing the activated T cell proliferative response and IFN␥ secretion (Fig. 3, A and B). The effect was shown to be concentration-dependent; proliferation and IFN␥ secretion were decreased by ⬃20% at a concentration of 20 ␮g/ml Hsp70 (Fig. 3, A and B). However, nonactivated T cells (CD3⫹) or purified subpopulations, either CD8⫹ or CD4⫹, did not show altered proliferative responses to mo-iDCs preincubated with 20 ␮g/ml Hsp70 (Fig. 3C); however, a significant down-regulatory effect was observed for CD3⫹, CD8⫹, and CD4⫹ cells in the IFN␥ secretion assay (Fig. 3D). Low Endotoxin Hsp70 Abrogates T Cell Responses—Alternatively designed experiments, in which Hsp70 was added to the

culture medium and remained present during the T cell response, were shown to have a cumulative down-regulatory effect. The altered activated T cell response was dependent on Hsp70 concentration and showed that Hsp70 at a concentration as low as 1 ␮g/ml effectively reduced proliferation and IFN␥ secretion in the activated T cell population. Furthermore, activated T cell proliferation and IFN␥ secretion decreased by ⬎50% at a protein concentration of 20 ␮g/ml (Fig. 4, A and B). Nonactivated CD3⫹, CD8⫹, or CD4⫹ cells showed ⬃20% decrease of proliferation and ⬃50% decrease of IFN␥ secretion when Hsp70, at a concentration of 20 ␮g/ml, was present in the culture medium (Fig. 4, C and D). In the light of the obtained results we hypothesized that Hsp70 may not only have an effect on mo-iDC stimulatory capacity but also modulate T cell responses independently, in

APRIL 6, 2012 • VOLUME 287 • NUMBER 15

JOURNAL OF BIOLOGICAL CHEMISTRY

12391

Hsp70 Abrogates T Cell Responses

FIGURE 5. Low endotoxin Hsp70 reduces T cell response upon IL-2 or PHA stimulation. A and B, activated T cell (A) or nonactivated T cell proliferation (B), either with or without IL-2 at a concentration of 40 IU/ml; Hsp70 was added to the cells at concentrations of 1, 10, or 20 ␮g/ml. C and D, nonactivated CD3, CD8, or CD4 T cell proliferation (C) and IFN␥ secretion (D) upon stimulation with 5 ␮g/ml PHA in the absence or presence of Hsp70 at a concentration of 20 ␮g/ml. Data are presented as mean percentage change relative to the untreated control ⫾S.D. (error bars). *, statistical significance of the results (n ⫽ 5, p ⬍ 0.05, t test).

the absence of mo-iDCs. Activated T cells were shown to proliferate upon addition of 40 IU/ml IL-2; the proliferation was down-regulated in the presence of Hsp70 by ⬃25% at the protein concentration of 20 ␮g/ml (Fig. 5A). As observed, nonactivated CD3⫹, CD8⫹, or CD4⫹ cells did not respond to IL-2 (Fig. 5B); however, they did respond to the addition of PHA (Fig. 5, C and D). Nonactivated CD3⫹, CD8⫹, or CD4⫹ upon stimulation with 5 ␮g/ml PHA was shown to proliferate and to secrete IFN␥; these responses were significantly down-regulated with 20 ␮g/ml Hsp70 (Fig. 5, C and D). The proliferative response was altered by ⬃20% whereas IFN␥ secretion was altered by ⬃60% and remained on the same level for CD3⫹, CD8⫹, and CD4⫹ cells (Fig. 5, C and D).

DISCUSSION The function of Hsp70 as a potential proinflammatory cytokine capable of inducing innate immunity (9) has long been

12392 JOURNAL OF BIOLOGICAL CHEMISTRY

considered controversial because it has been shown that the presence of endotoxin contaminants was most likely responsible for the observed effects (25). However, Hsp70 purified from nonbacterial origins was shown to be able to moderately stimulate macrophages as well as monocytes (21, 22). Hsp70 purified from a nonbacterial origin had only a modest effect on the maturation of iDCs, moderately inducing CD80 but not CD86, MHC class II, or CD40 at a concentration as high as 400 ␮g/ml (20). Although we were able to recreate artifactual proinflammatory properties of recombinant Hsp70 with two commercially available specimens of the protein, we failed to obtain similar results with low endotoxin Hsp70. The stimulatory capacity of Hsp70 has previously been shown to dependent solely on the endotoxin contamination and therefore considered an artifact (12). In this study we used Hsp70 containing very low level of endotoxin that failed to activate maturation of mo-iDCs within the tested concentration range. We also observed very reproducible changes in morphology of mo-iDCs that occurred after incubation with Hsp70. However, it has to be emphasized that these effects were only observed when DC serum-free medium was used. We noticed that mo-iDCs were very sensitive to human and fetal calf serum, which induced their spontaneous maturation that was reduced when mo-iDCs were incubated in the serum-free medium (data not shown). Therefore, we speculate that the addition of Hsp70 to mo-iDCs might have decreased the spontaneous maturation process that was further translated to their reduced stimulatory capacity toward T cells. We noticed that when Hsp70 was present in the culture medium during the T cell response assays it was able to reduce activated and nonactivated T cell response upon stimulation with allogenic mo-iDCs. The decrease in T cell response was higher than when Hsp70 was washed out after preincubation with mo-iDCs, thus suggesting that Hsp70 may have an influence on proliferating T cells independently from its effect on mo-iDCs. To test this hypothesis, T cell responses were assessed when stimulated by either IL-2 or PHA. Only activated T cells responded to the addition of IL-2, and that response was down-regulated by the presence of Hsp70. PHA stimulated activated and nonactivated T cells, and that activity was reduced by Hsp70. Therefore, Hsp70 was able to reduce both activated and nonactivated T cell responses independently from its modulatory effect on mo-iDCs. We also did not observe any difference in response to Hsp70 when either CD4 or CD8 subpopulations were used; however, this experiment was only performed with nonactivated T cells because the majority of the activated T cells co-expressed both T cell markers. Controversially, an extended incubation of T cells with low endotoxin Hsp70 has recently been reported to increase T cell proliferation and cytotoxic responses against target cells (23). Because Figueiredo et al. and we used the same specimen of Hsp70, leading to contradictory conclusions, it is important to critically appraise used methodology. We speculate that in the Figueiredo et al. study the extended (7 days) incubation of T cells with Hsp70 would have improved the viability of cells, thus improving their proliferation and cytokine secretion but as a function of cell viability rather than stimulation. Accordingly, VOLUME 287 • NUMBER 15 • APRIL 6, 2012

Hsp70 Abrogates T Cell Responses endotoxin-free Hsp70 when added to the culture medium at a concentration comparable with the one used by Figueiredo et al. was shown before to increase cell viability significantly (22). In our experiments we carefully monitored the response kinetics of T cells, which were adjusted for different stimuli and never exceeded 4 days, which in our opinion might have skewed the results. Consequently, we believe that our straightforward experimental setup provided true representation of Hsp70 function. In addition, our results are in agreement with the study by Motta et al. where the PHA-induced proliferative response of murine splenocytes was down-regulated by endotoxin-free Mycobacterium tuberculosis Hsp70 (15). Although it has been shown repeatedly that Hsp70 expression generally correlated with the grade of inflammation, it remains unclear whether the up-regulation of Hsp70 during the immune response is potentiating, antagonizing, or remains a bystander in this process. However, it was observed that Hsp70 was up-regulated in the late stages, rather than during the onset of an immune reaction, and by the affected tissue and not immune cells (26 –29). Consequently, one might speculate that the overexpression of Hsp70 is a protective mechanism that is activated by the targeted cells to counteract immune cytotoxicity. Furthermore, active secretion as well as an abundant surface expression of Hsp70 is common to tumor cells (30, 31). Surface expression of Hsp70 was shown to be specific to tumor cells and associated with poorer survival in cancer patients (32, 33). In the light of our results, it can therefore be speculated that Hsp70 provides a general mechanism that allows affected cells to counteract the immune response and was adapted by cancer cells as an immune evasive mechanism. Our results exhibit a novel role for Hsp70 in the regulation of the immune response. Hsp70 decreased T cell proliferation and cytokine secretion regardless of stimuli used and affected the stimulatory capacity of mo-iDCs. These observations might explain the role of Hsp70 as an immune modulator in several clinical scenarios where overexpression of Hsp70 is observed. Acknowledgment—We thank Prof. Britta Eiz-Vesper for providing low endotoxin Hsp70.

REFERENCES 1. Udono, H., and Srivastava, P. K. (1993) Heat shock protein 70-associated peptides elicit specific cancer immunity. J. Exp. Med. 178, 1391–1396 2. Li, Z., and Srivastava, P. K. (1993) Tumor rejection antigen gp96/grp94 is an ATPase: implications for protein folding and antigen presentation. EMBO J. 12, 3143–3151 3. Stocki, P., Morris, N. J., Preisinger, C., Wang, X. N., Kolch, W., Multhoff, G., and Dickinson, A. M. (2010) Identification of potential HLA class I and class II epitope precursors associated with heat shock protein 70 (HSPA). Cell Stress Chaperones 15, 729 –741 4. Stocki, P., Wang, X. N., Morris, N. J., and Dickinson, A. M. (2011) HSP70 natively and specifically associates with an N-terminal dermcidin-derived peptide that contains an HLA-A*03 antigenic epitope. J. Biol. Chem. 286, 12803–12811 5. Srivastava, P. (2002) Roles of heat-shock proteins in innate and adaptive immunity. Nat. Rev. Immunol. 2, 185–194 6. Binder, R. J., Harris, M. L., Ménoret, A., and Srivastava, P. K. (2000) Saturation, competition, and specificity in interaction of heat shock proteins (Hsp) gp96, Hsp90, and Hsp70 with CD11b⫹ cells. J. Immunol. 165, 2582–2587

APRIL 6, 2012 • VOLUME 287 • NUMBER 15

7. Delneste, Y., Magistrelli, G., Gauchat, J., Haeuw, J., Aubry, J., Nakamura, K., Kawakami-Honda, N., Goetsch, L., Sawamura, T., Bonnefoy, J., and Jeannin, P. (2002) Involvement of LOX-1 in dendritic cell-mediated antigen cross-presentation. Immunity 17, 353–362 8. Thériault, J. R., Mambula, S. S., Sawamura, T., Stevenson, M. A., and Calderwood, S. K. (2005) Extracellular HSP70 binding to surface receptors present on antigen presenting cells and endothelial/epithelial cells. FEBS Lett. 579, 1951–1960 9. Asea, A., Kraeft, S. K., Kurt-Jones, E. A., Stevenson, M. A., Chen, L. B., Finberg, R. W., Koo, G. C., and Calderwood, S. K. (2000) HSP70 stimulates cytokine production through a CD14-dependent pathway, demonstrating its dual role as a chaperone and cytokine. Nat. Med. 6, 435– 442 10. Vabulas, R. M., Ahmad-Nejad, P., Ghose, S., Kirschning, C. J., Issels, R. D., and Wagner, H. (2002) HSP70 as endogenous stimulus of the Toll/interleukin-1 receptor signal pathway. J. Biol. Chem. 277, 15107–15112 11. Asea, A., Rehli, M., Kabingu, E., Boch, J. A., Bare, O., Auron, P. E., Stevenson, M. A., and Calderwood, S. K. (2002) Novel signal transduction pathway utilized by extracellular HSP70: role of toll-like receptor (TLR) 2 and TLR4. J. Biol. Chem. 277, 15028 –15034 12. Gao, B., and Tsan, M. F. (2003) Endotoxin contamination in recombinant human heat shock protein 70 (Hsp70) preparation is responsible for the induction of tumor necrosis factor ␣ release by murine macrophages. J. Biol. Chem. 278, 174 –179 13. Bausinger, H., Lipsker, D., Ziylan, U., Manié, S., Briand, J. P., Cazenave, J. P., Muller, S., Haeuw, J. F., Ravanat, C., de la Salle, H., and Hanau, D. (2002) Endotoxin-free heat-shock protein 70 fails to induce APC activation. Eur. J. Immunol. 32, 3708 –3713 14. Ye, Z., and Gan, Y. H. (2007) Flagellin contamination of recombinant heat shock protein 70 is responsible for its activity on T cells. J. Biol. Chem. 282, 4479 – 4484 15. Motta, A., Schmitz, C., Rodrigues, L., Ribeiro, F., Teixeira, C., Detanico, T., Bonan, C., Zwickey, H., and Bonorino, C. (2007) Mycobacterium tuberculosis heat-shock protein 70 impairs maturation of dendritic cells from bone marrow precursors, induces interleukin-10 production, and inhibits T-cell proliferation in vitro. Immunology 121, 462– 472 16. Bendz, H., Marincek, B. C., Momburg, F., Ellwart, J. W., Issels, R. D., Nelson, P. J., and Noessner, E. (2008) Calcium signaling in dendritic cells by human or mycobacterial Hsp70 is caused by contamination and is not required for Hsp70-mediated enhancement of cross-presentation. J. Biol. Chem. 283, 26477–26483 17. Bendz, H., Ruhland, S. C., Pandya, M. J., Hainzl, O., Riegelsberger, S., Braüchle, C., Mayer, M. P., Buchner, J., Issels, R. D., and Noessner, E. (2007) Human heat shock protein 70 enhances tumor antigen presentation through complex formation and intracellular antigen delivery without innate immune signaling. J. Biol. Chem. 282, 31688 –31702 18. Triantafilou, K., Triantafilou, M., and Dedrick, R. L. (2001) A CD14-independent LPS receptor cluster. Nat. Immunol. 2, 338 –345 19. Triantafilou, M., and Triantafilou, K. (2004) Heat-shock protein 70 and heat-shock protein 90 associate with Toll-like receptor 4 in response to bacterial lipopolysaccharide. Biochem. Soc. Trans. 32, 636 – 639 20. Basu, S., Binder, R. J., Suto, R., Anderson, K. M., and Srivastava, P. K. (2000) Necrotic but not apoptotic cell death releases heat shock proteins, which deliver a partial maturation signal to dendritic cells and activate the NF-␬B pathway. Int. Immunol. 12, 1539 –1546 21. Panjwani, N. N., Popova, L., and Srivastava, P. K. (2002) Heat shock proteins gp96 and hsp70 activate the release of nitric oxide by APCs. J. Immunol. 168, 2997–3003 22. Zheng, H., Nagaraja, G. M., Kaur, P., Asea, E. E., and Asea, A. (20109) Chaperokine function of recombinant Hsp72 produced in insect cells using a baculovirus expression system is retained. J. Biol. Chem. 285, 349 –356 23. Figueiredo, C., Wittmann, M., Wang, D., Dressel, R., Seltsam, A., Blasczyk, R., and Eiz-Vesper, B. (2009) Heat shock protein 70 (HSP70) induces cytotoxicity of T-helper cells. Blood 113, 3008 –3016 24. Sallusto, F., and Lanzavecchia, A. (1994) Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/ macrophage colony-stimulating factor plus interleukin 4 and down-regulated by tumor necrosis factor ␣. J. Exp. Med. 179, 1109 –1118

JOURNAL OF BIOLOGICAL CHEMISTRY

12393

Hsp70 Abrogates T Cell Responses 25. Henderson, B., Calderwood, S. K., Coates, A. R., Cohen, I., van Eden, W., Lehner, T., and Pockley, A. G. (2010) Caught with their PAMPs down? The extracellular signalling actions of molecular chaperones are not due to microbial contaminants. Cell Stress Chaperones 15, 123–141 26. Jarvis, M., Marzolini, M., Wang, X. N., Jackson, G., Sviland, L., and Dickinson, A. M. (2003) Heat shock protein 70: correlation of expression with degree of graft-versus-host response and clinical graft-versus-host disease. Transplantation 76, 849 – 853 27. Baba, H. A., Schmid, C., Wilhelm, M. J., Blasius, S., Scheld, H. H., Böcker, W., and Dockhorn-Dworniczak, B. (1997) Inducible heat shock protein 70 in rat cardiac allograft and its immunohistochemical localization in cardiac myocytes. Transplantation 64, 1035–1040 28. Ludwig, D., Stahl, M., Ibrahim, E. T., Wenzel, B. E., Drabicki, D., Wecke, A., Fellermann, K., and Stange, E. F. (1999) Enhanced intestinal expression of heat shock protein 70 in patients with inflammatory bowel diseases. Dig. Dis. Sci. 44, 1440 –1447 29. Njemini, R., Lambert, M., Demanet, C., and Mets, T. (2003) Elevated se-

12394 JOURNAL OF BIOLOGICAL CHEMISTRY

30.

31.

32.

33.

rum heat-shock protein 70 levels in patients with acute infection: use of an optimized enzyme-linked immunosorbent assay. Scand J. Immunol. 58, 664 – 669 Multhoff, G., Botzler, C., Wiesnet, M., Müller, E., Meier, T., Wilmanns, W., and Issels, R. D. (1995) A stress-inducible 72-kDa heat-shock protein (HSP72) is expressed on the surface of human tumor cells, but not on normal cells. Int. J. Cancer 61, 272–279 Mambula, S. S., and Calderwood, S. K. (2006) Heat shock protein 70 is secreted from tumor cells by a nonclassical pathway involving lysosomal endosomes. J. Immunol. 177, 7849 –7857 Hantschel, M., Pfister, K., Jordan, A., Scholz, R., Andreesen, R., Schmitz, G., Schmetzer, H., Hiddemann, W., and Multhoff, G. (2000) Hsp70 plasma membrane expression on primary tumor biopsy material and bone marrow of leukemic patients. Cell Stress Chaperones 5, 438 – 442 Pfister, K., Radons, J., Busch, R., Tidball, J. G., Pfeifer, M., Freitag, L., Feldmann, H. J., Milani, V., Issels, R., and Multhoff, G. (2007) Patient survival by Hsp70 membrane phenotype: association with different routes of metastasis. Cancer 110, 926 –935

VOLUME 287 • NUMBER 15 • APRIL 6, 2012