Increased proteolysis of diphtheria toxin by human monocytes after heat shock: a subsidiary role for heat-shock protein 70 in antigen processing

Abstract

The expression of heat-shock proteins (hsp) increases after exposure to various stresses including elevated temperatures, oxidative injury, infection and inflammation. As molecular chaperones, hsp have been shown to participate in antigen processing and presentation, in part through increasing the stability and expression of major histocompatibility complex molecules. Heat shock selectively increases human T-cell responses to processed antigen, but does not affect T-cell proliferation induced by non-processed antigens. Here, we have analysed the mechanisms by which stress such as heat shock, and the ensuing hsp over-expression affect the processing of diphtheria toxin (DT) in peripheral blood monocytes. We found that heat shock increased DT proteolysis in endosomes and lysosomes while the activities of the cathepsins B and D, classically involved in DT proteolysis, were decreased. These effects correlated with the heat-shock-mediated increase in hsp 70 expression observed in endosomes and lysosomes. Actinomycin D or blocking anti-hsp 70 antibodies abolished the heat-shock-mediated increase in DT proteolysis. These data indicate that the increased expression of hsp 70 constitutes a subsidiary mechanism that facilitates antigen proteolysis in stressed cells. Confirming these data, presentation by formaldehyde-fixed cells of DT proteolysates that were obtained with endosomes and lysosomes from heat-shocked peripheral blood monocytes showed higher stimulation of T cells than those generated with endosomes and lysosomes from control peripheral blood monocytes.

Introduction

The heat-shock proteins (hsp) belong to a highly conserved group of molecular chaperones^1,2 that are both constitutively expressed (hsc: heat-shock cognates) and inducible. The hsp are classified into families according to their apparent molecular weight. Under stress conditions, such as elevated temperature, oxidative stress, infection or inflammation, the increased expression of hsp has been associated with cell protection against the deleterious effects generated by these conditions.³ Inflammation results from infections by pathogens (e.g. bacterium, virus) or is associated with the production of reactive oxygen species, the lipid mediators of inflammation and cytokines, all of which may modulate the expression of hsp.⁴

Processing and presentation of antigens for T-cell recognition requires unfolding of polypeptides, partial degradation, association with major histocompatibility complex (MHC) molecules and intracellular traffic of MHC–peptide complexes, which ultimately lead to membrane expression.^5,6 The hsp that function as molecular chaperones⁷ contribute to the correct subcellular targeting of proteins. They bind specific polypeptides such as clathrin-uncoating ATPase (hsc 70) or pigeon cytochrome c (PBP 72-74) (‘private chaperones’) or any unfolded or misfolded proteins (‘public chaperones’, constitutive and inducible hsp, to which group hsp 70 belongs).^8,9 As such, the possibility that members of the hsp 70 family could bind antigen and contribute to its presentation has been raised and examined by a number of groups, including ours.^10–12 Among the members of the hsp 70 family described as playing a role in antigen presentation, prp 73,^13,14 PBP 72-74,^15,16 the inducible hsp 70¹⁷ and the glucose-related protein of 78 kDa (grp 78)^18–21 must be cited. Antibodies to hsp 70 were found to block the presentation of antigenic peptides to class II-restricted T cells in both murine and human models.^9,22 The specificity of these various chaperones is not only restricted to the antigen but also to the intracellular compartment: indeed prp 73, PBP 72-74 and grp 78 are, respectively, specific chaperones into lysosomes, at the cell membrane and within the endoplasmic reticulum.

The precise step of antigen processing and presentation influenced by hsp 70 following heat shock (HS) is not clearly determined and might vary according to the temperature, the type of antigen-presenting cells (APC) and the antigen under consideration. One hypothesis is that HS modulates the expression or the stability of MHC molecules at the cell surface.^9,17,22 However, we observed that presentation of diphtheria toxin (DT) by human peripheral blood monocytes (PBM) pre-exposed to HS increased T-cell proliferation in the absence of any modification of MHC class II expression.¹⁷ In these APC, HS did not affect T-cell responses to non-processed antigens, such as superantigens, or in mixed lymphocyte reactions. Thus another hypothesis is that the HS-induced increase in T-cell activation in response to DT presented by PBM is the result of a HS-mediated modulation of antigen processing. Processing involves efficient antigen endocytosis, adequate proteolysis and stability of the resulting peptide–MHC class II complex, all of which are crucial for optimal antigen presentation and subsequent T-cell activation.^23,24 For exogenous antigens such as DT, proteolysis takes place in acidic compartments such as endosomes and/or lysosomes, where the thiol protease cathepsin B and the aspartic protease cathepsin D play a key role.^25–28

In the current study, we developed three different approaches to analyse the mechanisms by which HS and the ensuing hsp 70 overexpression affect the processing pathway of DT in human PBM. First, we investigated the expression of hsp 70 and DT proteolysis in endosomes and lysosomes. Second, we tested the effect of the transcriptional inhibitor actinomycin D under conditions that specifically abolished hsp synthesis; on HS mediated increase in DT proteolysis and presentation. Third, we estimated the direct involvement of hsp 70 in DT proteolysis using antibodies directed against hsp 70. We observed an increase in DT proteolysis in endosomes and lysosomes of heat-shocked PBM whereas the activities of cathepsin were diminished. The HS-mediated increase in DT processing induced greater stimulation of T-cell responses and correlated with hsp 70 overexpression. We propose that hsp 70 may constitute a subsidiary mechanism to promote adequate antigen proteolysis, and to amplify antigen presentation in stressed cells.

Materials and methods

Antibodies and reagents

Anti-hsc/hsp 70 monoclonal antibodies (mAbs) were purchased from StressGen Biotech Corp., Victoria, Canada (SPA810 and SPA820, respectively). Rabbit polyclonal antibody against tetanus toxin was provided by Dr M.G. Colomb (CEA, Grenoble, France). Iodogen was from Pierce Chemical Co (Brebières, France) and iodoacetamide, pepstatin A, actinomycin D, cathepsin B and cathepsin D were from Sigma-Aldrich (L'Isle d'Abeau, France). Substrates, benzyloxycarbonyl-Arg-Arg-2-naphthylamide and benzoyl-Arg-Gly-Phe-Phe-Pro-4-methoxy-2-naphthylamide were purchased from Calbiochem (Fontenay sous bois, France). Na¹²⁵I in solution (100 (m)Ci/ml) and [³H]thymidine (specific activity 20 Ci/mmol) were obtained from Life-Science (Saclay, France).

Cell preparations

Peripheral blood mononuclear cells were isolated from normal volunteers by Ficoll–Hypaque gradient centrifugation and monocytes were purified by adherence as previously described.²⁹ The freshly isolated cells were cultured at 37° in antibiotic-free RPMI-1640 (InVitroGen, Cergy Pontoise, France) supplemented with 10% fetal calf serum and 1% glutamine, at 37°, in a humidified atmosphere (95% air, 5% CO₂). Human T lymphocytes (provided by the Etablissement du Sang Français, La Tronche, France) were prepared as described by Mariéthoz et al.¹⁷ Briefly, after adherence of peripheral blood mononuclear cells for 45 min, non-adherent cells were applied for 1 hr to a nylon wool column (Biotest Diagnostics, Dreieich, Germany) previously made wet with 3 ml complete medium (RPMI-1640–1% glutamine supplemented with 20% pooled human male AB serum, provided by the Etablissement du Sang Français). Lymphocytes were then slowly eluted with complete medium, washed and resuspended in complete medium at a final concentration of 1 × 10⁶ cells/ml (the suspension contained 95% T lymphocytes).

Analysis of protein synthesis

In these experiments, RPMI-1640 medium was replaced by RPMI-1640 medium without methionine (InVitroGen). Human PBM were exposed to the indicated temperatures for 20 min, and then labelled with [³⁵S]methionine (6 μCi/ml; specific activity > 1000 Ci/mmol; Amersham) for 90 min at 37° after the 2-hr recovery. After labelling, samples corresponding to equal cell numbers were resolved by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE; 9% polyacrylamide) according to Laemmli³⁰ and were revealed by autoradiography.

Exposure to HS and actinomycin D

For HS, PBM were incubated in a water-bath at 44° for 20 min, followed by a recovery period of 3·5 hr at 37°. To analyse the involvement of hsp synthesis in our experiments, we used the transcriptional inhibitor actinomycin D (5 μg/ml): it was added to the cells 10 min before HS and removed at the end of HS according to a previously described protocol.^17,31 Lymphocytes were not exposed to HS.

Proliferation assays

T lymphocytes (1 × 10⁵) and autologous PBM (3 × 10⁴) exposed or not to HS, were cocultured in 96-well flat-bottom tissue culture plates. DT (10 μg/ml) or phytohaemagglutinin (0·75 μg/ml) were added to the cells. All conditions were treated in quintuplicate. On day 5, [³H]thymidine was added for 16 hr. Cells were harvested and lysed (Skatron automatic cells harvester, Skatron Instruments, Lier, Norway), and labelled DNA was counted in scintillation vials using a Beckman Liquid Scintillation Counter (Beckman LS 7500, Gagny, France).

In DT proteolysate presentation to T cells, autologous PBM were fixed according to Casten and Pierce.³² Then, 3 × 10⁴ fixed cells were cocultured in 96-well flat-bottom tissue culture plates (Costar) in the presence of 50 μl DT digested by endosome and lysosome fractions obtained from PBM that had been exposed or not to HS as previously described. The plates were centrifuged before the addition of 1 × 10⁵ T cells. Proliferation of these cells was analysed as described before.

Subcellular fractionation

Cytosol-, lysosome- and endosome-enriched fractions (thereafter called cytosol, lysosomes and endosomes) were obtained by differential centrifugation of PBM lysates in homogenization buffer (pH 7·4) containing 0·25 m sucrose, 1 mm ethyleneglycoltetraacetic acid and 3 mm imidazole, as described elsewhere.³³ Briefly, the cells were disrupted by 45 runs in a cell-cracker (8·02 EMBL: 8·006-mm diameter ball, European Molecular Biology, Heidelberg, Germany), leading to more than 90% lysis as monitored by phase-contrast microscopy. Nuclei were removed by centrifugation at 600 g (15 min at 4°) and the supernatant was fractionated by centrifugation on a Percoll gradient according to a modified protocol of Stoorvogel et al.³⁴ Briefly, 11 ml Percoll (17%, v/v) (Life Sciences) in homogenization buffer was layered on a 1-ml 60% (w/v) sucrose cushion in 14 × 76-mm Quick-Seal tubes (Beckman, Gagny, France). The postnuclear supernatant was layered on top. After centrifugation for 90 min at 22 000 g in a 70.1 TI rotor (Beckman) at 4°, gradients were recovered in 0·6-ml fractions from the bottom of the tubes, using an Auto Densi Flow IIc apparatus (HBI; Roucaire, Courtaboeuf, France). In some experiments, the postnuclear supernatant was centrifuged for 40 min at 100 000 g at 4°. The supernatant corresponded to cytosol. Fractions were stored at − 20° until use.

After subcellular fractionation on Percoll gradients, endosome and lysosome fractions were localized using transferrin incorporation and measurement of galactosaminidase activity, respectively. After incorporation of radiolabelled transferrin, a peak of radioactivity was observed at the top of the gradient and the corresponding tubes were assessed to be endosome-containing fractions.³⁵ Galactosaminidase activity was determined using paranitrophenyl-N-acetylgalactosaminidine (5 mm in 10% dimethylsulphoxide, Sigma) as substrate. Galactosaminidase activity was seen in a single peak at the bottom of the gradient and the corresponding tubes were assessed to be lysosome-containing fractions.³⁶

Analysis of hsc/hsp 70 subcellular localization

Subcellular fractions prepared from PBM were resolved by SDS–PAGE (7% polyacrylamide) then transferred to nitrocellulose. Samples were probed with the SPA810 and SPA820 antibodies and bound antibodies were revealed with an anti-mouse immunoglobulin G–peroxidase conjugate (Sigma) in the presence of H₂O₂ and 4-chloro-1-naphthol (Sigma). The SPA820 antibody recognized the constitutive hsc 70 (anti-hsc 70) and cross-reacted with the inducible hsp 70 whereas the SPA810 antibody recognized only the inducible hsp 70.

In vitro proteolysis of DT by PBM subcellular fractions

Native monochain DT was purchased from Calbiochem and labelled with Na¹²⁵I (Amersham, France) using the Iodogen (Pierce) method.³⁷ Free iodine was eliminated by filtration through a Sephadex G50-fine column. The average activity of preparations was 6 × 10⁶ counts per minute (c.p.m.)/μg protein. Antigen proteolysis by PBM cytosol, lysosome or endosome fractions was analysed as follows: 1 μg ¹²⁵I-labelled DT was incubated for 4 hr at 37° in the presence of 30 μl (corresponding to the same number of cells) of subcellular fractions prepared from 10 × 10⁶ PBM exposed or not to HS in 50 mm sodium acetate buffer, pH 5·5 (final volume 50 μl). The proteolysis was stopped by cooling samples to 0° before analysis of proteolysates by SDS–PAGE according to Laemmli (12% acrylamide gels) under reducing conditions.³⁰ Alternatively, 30 μl of the subcellular fractions prepared from 10 × 10⁶ PBM exposed or not to HS were preincubated with 10 μl anti-hsc/hsp 70 (1 μg/μl) or purified polyclonal anti-tetanus toxin antibody (1 μg/μl, used as an irrelevant antibody) for 45 min at 4°. Then, ¹²⁵I-labelled DT was added and the experiment was performed as described before.

Cathepsins

Activities of cathepsins towards synthetic substrates:

Activities of cathepsins B and D were determined using benzyloxycarbonyl-Arg-Arg-2-naphthylamide and benzoyl-Arg-Gly-Phe-Phe-Pro-4-methoxy-2-naphthylamide (20 mm in 50 mm sodium acetate buffer, pH 5), respectively, as substrates. Twenty microlitres of each fraction was mixed with 5 μl substrate, 75 μl 50 mm sodium acetate buffer (pH 5) containing 0·01% (w/v) Triton X-100 and 6 mm cysteine for 1 hr at 37°; the reaction was then stopped with 1 mm iodoacetamide. Release of 2-naphthylamide was monitored by its fluorescence at 410 nm (excitation = 335 nm) using an F-2000 Fluorescence Spectrophotometer (Hitachi, Tokyo, Japan). To confirm the specificity of this assay, each fraction was preincubated for 30 min at room temperature with 10 mm iodoacetamide (cathepsin B) or 10 mm pepstatin A (cathepsin D) before the above described procedure.

DT proteolysis by purified cathepsins:

¹²⁵I-labelled DT (1 μg) was incubated at 37° for different times in the presence of 5 μg cathepsin B or D (Calbiochem) in 50 mm sodium acetate buffer, pH 5·5 (final volume 50 μl). For cathepsin B, the sodium acetate buffer was 8 mm in l-cysteine to optimize thiol proteolysis. The reactions were stopped by adding 10 mm iodoacetamide (cathepsin B) or 10 mm pepstatin A (cathepsin D) followed by a further incubation of 30 min at 37° before cooling to 0°. Proteolysates were analysed by SDS–PAGE according to Laemmli (12% acrylamide gels) under reducing conditions.

Results

DT processing and presentation by PBM pre-exposed to HS increase T-cell proliferation

In PBM, maximal hsp70 synthesis without inhibition of normal protein synthesis was obtained after exposure to 44°, as shown by metabolic labelling with [³⁵S]methionine (Fig. 1a). This temperature was therefore selected for most subsequent studies. The transcriptional inhibitor actinomycin D, added 10 min before and during HS, selectively prevented hsp induction by HS, without altering normal protein synthesis (Fig. 1b).

Figure 1.

HS-induced expression of hsp in human PBM. (a) SDS–PAGE analysis of protein synthesis by PBM after exposure to the indicated temperature. While the expression of inducible hsp 65, hsp 70, hsp 90 and hsp 110 was observed after exposure to 44° and 45°, the decrease in normal protein synthesis only occurred after exposure of PBM to 45°. (b) SDS–PAGE analysis of protein synthesis by PBM after treatment with actinomycin D (AD; 5 μg/ml) added 10 min before HS to parallel cultures and removed at the end of HS. The hsp 65, hsp 70, hsp 90 and hsp 110 induced after exposure to 44° were no longer detectable when actinomycin D had been added, whereas total protein synthesis was unaltered.

Levels of T-cell proliferation in response to antigens processed by PBM that had been pre-exposed to 44° or maintained at 37° were compared (Table 1). HS altered neither basal levels of [³H]thymidine incorporation into T cells (negative control, 2 × 10³ c.p.m. in both cases) nor T-cell proliferation in response to phytohaemagglutinin (positive control, 35 × 10³ versus 32 × 10³ c.p.m.). In contrast, T-cell proliferation in response to DT was significantly increased when the APC were pre-exposed to HS (16 × 10³ versus 29 × 10³ c.p.m.). This increase was also blocked by actinomycin D, which had no effect on lymphocyte proliferation when APC were maintained at 37°, suggesting that the HS-mediated stimulation of T cells involved hsp synthesis (Table 1). The efficiency of heat-shocked PBM to present DT was further confirmed using various concentrations of antigen (Fig. 2). Similar T-cell proliferation was observed with 50-fold less DT when the APC were pre-exposed to HS. The HS-mediated increase in DT presentation and T-cell proliferation was still abolished in the presence of actinomycin D.

Table 1.

Effect of HS on T-cell proliferation: actinomycin D inhibits the increase in T-cell proliferation induced by HS in the DT-stimulated PBM

		[3H]TdR incorporation
	n	37°	44°	P-value
Control	40	2 ± 0·3	2 ± 0·4	NS
PHA	40	35 ± 6·0	32 ± 4·0	NS
DT	38	16 ± 3·0	29 ± 4·0	≤0·01
DT + AD	16	14 ± 2·0	15 ± 4·0	≤0·01

T lymphocytes (1 × 10⁵) were cocultured in the presence of 3 × 10⁴ autologous PBM exposed or not to heat shock and [³H]TdR incorporation was determined in proliferation assays performed with or without actinomycin D. The results are expressed as c.p.m. × 10³ ± SEM. The heat-shock-induced increase in lymphocyte proliferation (P = 0·0009) was significantly decreased in the presence of actinomycin D (P < 0·05).

n, number of experiments; PHA, phytohaemagglutinin; DT, diphtheria toxin; AD, actinomycin D.

Figure 2.

Comparative efficiency to present DT of PBM that have been exposed or not to HS. T lymphocytes (1 × 10⁵) and autologous PBM (3 × 10⁴) exposed or not at 44° for 20 min (in the presence or not of actinomycin D) were cocultured in the presence of increasing concentrations of DT. T-cell proliferation was estimated, as described in the Materials and methods. DT presentation by PBM pre-exposed to HS was more efficient for T-cell stimulation. This effect was abolished in the presence of actinomycin D.

Effects of HS on DT proteolysis

T-cell activation can be modulated according to the number of MHC class II–peptide complexes expressed at the membrane of APC or the processing of the antigenic material after its capture and endocytosis. We and others have previously reported that no increase in class II molecule expression was detectable in PBM after HS. To test the possibility that the HS-mediated increase in DT presentation by PBM resulted from a HS-induced modulation of antigen processing, we used SDS–PAGE analysis to compare ¹²⁵I-labelled DT proteolysis in endosome, lysosome and cytosolic fractions prepared from human PBM pre-exposed, or not, to HS (Fig. 3a). In control PBM, DT proteolysis took place primarily in endosomes (lane 2) and lysosomes (lane 3): in those subcellular fractions intact DT was metabolized into its two major proteolytic fragments with molecular weights (MW) of 39 000 (Da) and 24 000 (Da), respectively (corresponding to the heavy and light chains of DT) and minor peptides which migrated at 29 000 (Da) or 15 000 (Da) MW, or accumulated at the front of the gel. In PBM pre-exposed to HS before fractionation, DT proteolysis was increased in both endosomes and lysosomes (Fig. 3a, compare lanes 5 and 6 to lanes 2 and 3), but not in cytosol (compare lane 4 to lane 7); there was no more intact DT in endosomes and lysosomes of heat-shocked PBM. The effect of HS was prevented by actinomycin D (compare lanes 8 and 9 to lanes 5 and 6), suggesting that hsp themselves could be involved in DT proteolysis.

Figure 3.

Effect of HS on DT proteolysis by human monocytes. (a) Endosomes (E), lysosomes (L) and cytosol (C) (corresponding to equal cell number) prepared from PBM exposed or not to HS (20 min at 44°) with or without actinomycin D (AD) were incubated for 4 hr at 37° with ¹²⁵I-labelled DT before being analysed in SDS–PAGE under reducing conditions. DT proteolysis which took place in endosomes and lysosomes (lanes 2 and 3) was increased following HS (lanes 5 and 6). This effect was partly inhibited by actinomycin D (lanes 8 and 9). Lane 1: control DT. (b) When endosomes and lysosomes were exposed to HS after subcellular fractionation (excluding hsp synthesis), elevation of temperature had no effect on DT proteolysis (compare lanes 4 and 5 to lanes 6 and 7). Column markers: H, heavy chain; L, light chain.

To test whether the observed HS-mediated increase in DT proteolysis was not solely the result of a direct physical effect of HS on proteolysis, endosome and lysosome fractions isolated from control PBM were exposed to HS after fractionation (excluding hsp synthesis), then DT proteolysis was analysed by SDS–PAGE (Fig. 3b). Under these conditions, proteolysis was similar to that of control fractions (lanes 4 and 5 compared to lanes 2 and 3). Only when HS was performed on PBM before fractionation (including hsp synthesis) did an increase in DT proteolysis occur (lanes 6 and 7 compared to lanes 4 and 5).

Effects of HS on cathepsin activities

Cathepsins D and B are major endosomal and lysosomal proteases involved in intracellular antigen cleavage. The proteolytic activities of purified cathepsin B and cathepsin D were tested on ¹²⁵I-labelled DT. The purity and specificity of the two proteases were established by addition of leupeptin or iodoacetamide (cathepsin B) and pepstatin A (cathepsin D) in control tests (data not shown). As shown in Fig. 4(a), both cathepsins were able to proteolyse DT and the major proteolytic products were similar: 39 000 (Da) (DT heavy chain), 29 000 (Da), 24 000 (Da) (DT light chain), 15 000 (Da) MW fragments and peptides migrating to the front. However, their kinetics were different: proteolysis by cathepsin B was more rapid than the proteolysis by cathepsin D because after 1 hr there was no more intact DT with cathepsin B while there was a significant amount with cathepsin D (Fig. 4a). In both cases DT proteolysis started by the cleavage of the monochain into its heavy and light chains which were then degraded.

Figure 4.

Kinetics of DT proteolysis by purified cathepsins B and D. (a) SDS–PAGE analysis in reducing conditions of DT digested by cathepsins B and D, at 15 min, 30 min, and 1, 2, 4 and 6 hr at 37°. Both cathepsins are able to digest DT in a manner similar to subcellular endosomes and lysosomes. (b) Effect of HS on the activity of cathepsins B and D. The activities were estimated in endosomes (E) and lysosomes (L) from cells exposed or not to HS using specific substrates as described in the Materials and methods. (IAA, iodoacetamide; pepA, pepstatine A; H, heavy chain; L, light chain). HS down-regulated both cathepsin's activities in endosomes as well as in lysosomes.

Using specific substrates for these two cathepsins we then compared their activities in endosomes and lysosomes prepared from PBM pre-exposed or not to HS (Fig. 4b). The specificity of cathepsin activity in the various fractions was confirmed by inhibition experiments with iodoacetamide (cathepsin B) or pepstatin A (cathepsin D). The activity of both cathepsins was higher in endosomes than in lysosomes, while in both endosomes and lysosomes cathepsin B activity was 4·40-fold greater than cathepsin D activity. In endosomes of heat-shocked PBM, cathepsin B and D activities were diminished, respectively, by 79% and 64% and by 81% and 46% in lysosomes.

hsp 70 is involved in the HS-mediated increase of DT proteolysis

To test the possibility that hsp 70 synthesis modulates the effect of HS on DT proteolysis, we investigated the intracellular localization of hsp 70 in PBM pre-exposed, or not, to HS (Fig. 5). Using two different antibodies against the 70 000 (Da) MW hsp, we confirmed that hsp 70 expression increased after HS (Fig. 5a). The SPA820 antibody directed against the constitutive hsc 70 cross-reacted with the inducible hsp 70 and recognized both hsp 70 and hsc 70 (anti-hsc/hsp 70, lane 2 compared to lane 1, upper panel), whereas the SPA810 antibody only recognized the inducible hsp 70 (lane 2 compared to lane 1, lower panel).

Figure 5.

Subcellular localization of hsp following HS. (a) Western blot analysis of hsc and hsp 70 expression following HS (lanes 2); controls are shown in lanes 1. The SPA820 antibody directed against the constitutive hsc 70 cross-reacts with the inducible hsp 70 and recognizes both hsp 70 and hsc 70 (anti-hsc/hsp 70, lane 2 upper panel), whereas the SPA810 antibody only recognizes the inducible hsp 70 (lane 2, lower panel). (b) Western blot analysis of subcellular localization of hsp 70 following HS. HS induced hsp 70 expression in endosomes and lysosomes (lanes 2 and 5). This effect was inhibited by actinomycin D (lanes 3 and 6).

As DT proteolysis took place in endosomes and lysosomes, we further analysed the expression of hsc 70 and hsp 70 in both these compartments using the SPA820 antibodies. hsc 70 was detected in endosomes and lysosomes of unstressed cells (Fig. 5b, lanes 1 and 4). After HS, there was an increase in the expression of hsp 70 in both subcellular fractions (lanes 2 and 5), which was abolished by actinomycin D (lanes 3 and 6).

The link between the HS-mediated up-regulation of hsp 70 expression in endosomes/lysosomes and the HS-mediated increase in DT proteolysis was investigated by comparing DT proteolysis in the presence or absence of SPA820 antibodies. We chose this antibody because it recognizes an epitope that is present in the peptide-binding site and so blocks the action of hsp 70 while SPA810 recognizes an epitope located at the C-terminal region immediately after the ATPase domain.³⁸ As shown in Fig. 6, the presence of SPA820 antibodies abolished the HS-mediated increase in DT proteolysis by endosomes (lane 4 compared to lane 3) and lysosomes (lane 8 compared to lane 7), while, as expected, SPA810 had no such blocking effect (data not shown). Inhibition of DT proteolysis by SPA820 antibody was specific for hsp 70: irrelevant antibodies directed for example against tetanus toxin had no effect (lanes 5 and 9). It appeared that under altered proteasic activity, such as following exposure to HS, hsp 70 constitutes a subsidiary mechanism ensuring antigen degradation.

Figure 6.

Inhibition of HS increases DT proteolysis by antibodies against hsc/hsp 70. SDS–PAGE analysis of DT proteolysis by endosomes and lysosomes from control or heat-shocked cells, performed for 4 hr at 37° in the presence or absence of antibodies against hsc/hsp 70 (SPA820 Ab). Antibodies for hsc/hsp 70 block the increased proteolysis of DT in lysosomes and endosomes from heat-shocked cells (lanes 4 and 8 compared to lanes 3 and 7). Irrelevant antibodies (Irrel. Ab) had no effect on DT proteolysis (lanes 5 and 9). Lane 1: control DT.

The HS-mediated increase of DT proteolysis leads to enhanced specific T-cell proliferation

The link between the hsp 70-mediated increase of DT proteolysis and improved T-cell proliferation was investigated as follows: DT was incubated for different periods, in the presence of endosome/lysosome-enriched fractions from PBM that had been pre-exposed or not to HS. The resulting proteolysates were then presented to autologous T cells by formaldehyde-fixed PBM (Fig. 7). T-cell activation and proliferation were maximal at 8 hr with proteolysates of DT obtained from unstressed cells, and then declined. In contrast, T-cell activation and proliferation were more efficient by 4 hr and maximal at 8 hr with proteolysates of DT obtained from heat-shocked cells. The proteolysis of DT in the presence of hsp 70 optimized antigen presentation, as reflected by [³H]thymidine uptake in T cells.

Figure 7.

DT proteolysis by endosomes/lysosomes from heat-shocked PBM leads to enhanced T-cell proliferation. DT was incubated for 4, 8 or 24 hr in the presence of endosome/lysosome-enriched fractions from PBM that had been exposed or not to HS. The resulting proteolysates were then presented to autologous T cells by formaldehyde-fixed PBM.

Discussion

Here we report that HS enhances antigen-driven T-cell responses (where DT is the antigen) by increasing DT proteolysis while simultaneously inhibiting the activity of the two proteases, cathepsins B and D, which degrade DT. This HS-induced modulation of the immune response is mediated by hsp 70 over-expression, which thereby constitutes a subsidiary mechanism for antigen processing that amplifies antigen presentation in stressed immune cells.

The effects of HS on antigen presentation have been attributed by several groups to a modulation of the expression and/or the stability of MHC molecules.^16,22 Consistent with this analysis, we found that in human B cells, HS-induced stimulation of T-cell activation results from an up-regulation of MHC class II molecules, whereas in murine B cells, HS did not modulate class II expression, but rather accelerated the formation of compact αβ class II dimers. Detection of proteins related to the hsp 70 family in processing compartments suggests that these chaperones may ensure the translocation of the antigen in these compartments and/or the appropriate assembly of processed antigen–MHC complexes.^11,13,39,40 This hypothesis has been emphasized by the demonstration, using co-immunoprecipitation experiments, of a transient association into lysosomes of constitutive hsc 70 with immature or mature arthritis-specific human leucocyte antigen (HLA) DR molecules.⁴¹ However, in PBM, HS increased neither the expression of class II molecules nor the stability of MHC–peptide complexes. Despite inducible hsp 70 being able to associate with HLA-DR molecules in endosomes and lysosomes, we did not detect any changes in the formation of stable compact αβ-chain dimers identified by their resistance to SDS denaturation (refs ^{17, 42} and our unpublished data).

In the present investigation, we further analysed how HS could modulate DT processing. We found that HS-exposed PBM process and present antigen more rapidly, as compared to PBM maintained at 37°. This accelerated DT proteolysis correlates with enhanced activation of T cells, suggesting that more antigenic epitopes are available earlier in heat-shocked PBM. In contrast, Pépin et al. described that while HS increased tetanus toxin (TT) proteolysis in human Epstein–Barr virus-transformed B lymphocytes, they were less efficient in TT presentation and in activation of TT-specific T cells.⁴³ This may be explained by the observation that limited proteolysis is required for efficient activation of TT-specific T cells and that following HS, the increase in TT proteolysis is incompatible with the production of immunogenic peptides.²⁸ However, HS may not only result in a change of the quantity of peptides but also in the type of peptides that result from processing, leading to a determinant with less affinity for MHC molecules or that is unrecognized by T-cell receptor. The suppression of the activation of two TT-specific T-cell clones upon presentation of TT by heat-shocked APC supports this hypothesis (personal communication). So the HS-mediated alteration in antigen processing appears to be a general process that could modulate the immune response differently (boost or suppress the presentation of determinants) according to the nature of the antigen.

We first considered the possibility that the HS-mediated increase in antigen proteolysis was associated with an increase in the activities of cathepsin B and D, two proteases involved in DT degradation and more generally in antigen proteolysis.^28,44,45 However, cathepsin's activities were actually inhibited in heat-shocked PBM. We then developed three different approaches to test the hypothesis that the effects of HS on DT proteolysis were mediated by members of the hsp 70 family and constituted a subsidiary mechanism for antigen processing and presentation in stressed immune cells. First, we investigated the expression of hsp 70 in endosomes and lysosomes of PBM exposed or not to HS. Second, we tested the effect of actinomycin D on the HS-mediated increases in DT proteolysis and presentation to T cells. Third, we estimated the direct involvement of hsp70 in the HS-mediated effects on DT proteolysis using antibodies that recognized the peptide binding site common to hsc 70 and hsp 70.

Following exposure to HS, over-expression of hsp 70 occurred in parallel with the increased DT proteolysis in both endosomes and lysosomes. Furthermore, under conditions which specifically inhibit hsp expression, we abolished the increase in DT proteolysis in endosomes and lysosomes and its presentation by PBM cells. These observations indicated that HS-induced hsp 70 expression in these processing compartments was involved in the HS-induced modulation of DT processing and presentation.

Since HS diminished cathepsin's activities and actinomycin D suppressed hsp 70 synthesis, one would expect a lower activation of T cells under these conditions when compared to controls, but T-cell activation was similar in both cases, suggesting that the residual activity of cathepsins was sufficient for DT degradation. Another explanation may be thermotolerance: loss of enzymatic activity following protein denaturation has been described in lysates from HS-exposed cells, while molecular chaperones contribute to protein renaturation according to their folding activity.^46–48 In our tests, T-cell proliferation was estimated after 5 days, a delay sufficient for restoration of protease activities and APC functions.

Preincubation of endosomes and lysosomes from heat-shocked PBM with SPA820 antibody (but not with antibodies directed against proteins irrelevant for these tests) before addition of ¹²⁵I-labelled DT, prevented the degradation of DT in both processing compartments. However, this effect was more important in lysosomes where cathepsin activity was weaker. While DT proteolysis increased after HS, suggesting the contribution of hsp 70, we also noticed an inhibition in control lysosomes, suggesting that hsc 70 might contribute to antigen degradation in the absence of stress (data not shown). The constitutive hsc 73 has been described as facilitating the lysosomal import and degradation of proteins containing a KFERQ sequence.^12,49–51 Though DT does not contain this motif, we cannot exclude the contribution of hsc 73 or another ‘private’ chaperone (recognizing an as yet unidentified sequence) to DT proteolysis in lysosomes of control PBM.

In any case, our observations suggest that the HS response does not only affect processed antigen–class II complex assembly, as previously described, but also modulates the cellular machinery required for processing. Heat-shock may enhance ‘normal processing’ (leading to availability of more antigenic epitopes) or it may favour a ‘differential processing’ (leading to the generation of neo-epitopes). We propose that HS-mediated modulation of antigen processing may be the result of the dramatic increase in denatured or improperly folded proteins in stressed cells as well as the induction of new cellular machinery, namely the hsp 70 family members, to cope with the damaged proteins. When cathepsin activities are elevated, as is the case in control endosomes, hsp 70 are not required for antigen proteolysis, which relies on classical proteases. In contrast, when the proteolytic potential of cathepsins is low or impaired, such as in lysosomes or following HS, hsc 70, and upon HS, hsp 70, constitute a physiologically relevant mechanism aimed at facilitating antigen proteolysis (Fig. 8). On one hand, association of hsp with antigenic peptides might serve to shield the antigens from their ultimate degradation, (i) by maintaining or favouring unfolding of antigen, so leading to exposure of previously inaccessible proteolytic sites, or (ii) by reducing antigen, a step which has been described to be sufficient to induce T-cell proliferation.^52,53 Indeed, reductive activity of hsp 70 has been evoked by Simpkins et al.⁵⁴ according to the presence in its N-terminal region of a cysteine-containing hydrophobic sequence (YSCVGVF). On the other hand, we could not exclude the possibility that, depending upon which portion of the antigenic molecules was bound by the hsp, different T-cell determinants on the molecules would be preferentially generated and presented in association with MHC molecules. A similar process has already been described for the immunoglobulin receptor on the B cell, which could display different determinants following an altered antigen-processing activity.^55–59 This latter possibility would present an interest in the inflammatory environment where hsp are up-regulated. In particular, hsp 70 association with antigen may be involved in the phenomenon of determinant spreading in autoimmune diseases, as described by Sercaz's group (for a review see ref. 60). Dominant targets of self-antigen that remain cryptic under normal conditions because of insufficient processing may be produced after association of antigen with hsp, leading to the activation of pathogenic autoreactive T-cell clones.

Figure 8.

Model for antigen processing under stress conditions. See the text.

Finally, one should mention that the effects of hsp are not limited to class II-restricted antigenic peptides: HS has been associated with increasing endogenous antigen presentation by both class I and II molecules.^61–64 Suitable expression of molecular chaperones might overcome the defective antigen processing and presentation often observed in virus-infected and malignant cells.^65–67

In conclusion, increased antigen proteolysis after HS provides additional evidence for an active role of molecular chaperones in the immune response and a mechanism for immune cells to develop an adaptive response to stress conditions. On the one hand, hsp are immunodominant antigen of many bacteria and parasites^68,69 and they also contribute to antitumour immunity.^20,70–74 On the other, hsp as ‘private’ chaperones for antigens can amplify the presentation of such antigens to T cells, specifically under stress conditions where processing and presentation pathways are defective. This latter effect might be particularly relevant at inflammatory sites where hsp are up-regulated by cytokines and reactive oxygen species.⁷⁵

Abbreviations

APC

antigen-presenting cell

c.p.m.

counts per minute

DT

diphtheria toxin

HLA

human leucocyte antigen

HS

heat shock

hsp

heat-shock protein

mAb

monoclonal antibody

MHC

major histocompatibility complex

PBM

peripheral blood monocytes

SDS–PAGE

sodium dodecyl sulphate–polyacrylamide gel electrophoresis