Heat shock protein A2 is a novel extracellular vesicle-associated protein

Abstract

70-kDa Heat Shock Proteins (HSPA/HSP70) are chaperones playing a central role in the proteostasis control mechanisms. Their basal expression can be highly elevated as an adaptive response to environmental and pathophysiological stress conditions. HSPA2, one of poorly characterised chaperones of the HSPA/HSP70 family, has recently emerged as epithelial cells differentiation-related factor. It is also commonly expressed in cancer cells, where its functional significance remains unclear. Previously, we have found that proteotoxic stress provokes a decrease in HSPA2 levels in cancer cells. In the present study we found that proteasome inhibition-related loss of HSPA2 from cancer cells neither is related to a block in the gene transcription nor does it relate to increased autophagy-mediated disposals of the protein. Proteotoxic stress stimulated extracellular release of HSPA2 in extracellular vesicles (EVs). Interestingly, EVs containing HSPA2 are also released by non-stressed cancer and normal cells. In human urinary EVs levels of HSPA2 were correlated with the levels of TSG101, one of the main EVs markers. We conclude that HSPA2 may constitute basic components of EVs. Nevertheless, its specific role in EVs and cell-to-cell communication requires further investigation.

Figure 1

Changes in HSPA1 and HSPA2 expression in cells exposed to proteasome inhibitors. (a–c) Levels of HSPA1 and HSPA2 proteins after treatment with MG132 (a), Bortezomib (BTZ) (b), or Manumycin A (MA) (c). Representative immunoblots are provided (n ≥ 3). β-actin was used as a protein loading control. Prior to incubation with primary antibody membranes were cut according to 55 kDa molecular ladder band. Chemiluminescent signal was detected on X-ray film. Original autoradiograms/immunoblots are presented in Fig. S2. (d–g) Effects of proteasome inhibitors (24 h) on the HSPA2 (d–f) and HSPA1 (g) mRNA levels. Results of RT‐qPCR analysis showed as mean ± SD from at least three independent experiments, each in three technical replicas. RPL13A, TMEM43, and B2M were used as the reference genes (Table S1). Statistical significance was determined by the two-tailed t-test performed in regard to cells exposed to DMSO solvent only. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 2

Autophagy inhibition does not prevent proteasome inhibition-related reduction in intracellular levels of HSPA2. (a) Effects of the single or combined treatment with bafilomycin A (BAF) and bortezomib (BTZ) on the HSPA1, HSPA2, and p62 protein levels. (b) Effects of a single or combined treatment (16 h) with VER-155008 (VER), BTZ, or chloroquine (ChQ) on the protein levels of HSPA1, HSPA2 and p62. In either experiment, HSPA1 and p62 were used as a BTZ or ChQ/BAF treatment control, respectively. β-actin was used as a protein loading control. Prior to incubation with primary antibodies membranes were cut according to the molecular ladder band (55 kDa). For chemiluminescent signal detection X-ray film was used. Original autoradiograms/immunoblots are presented in Fig. S3. Graphs on the right side show results of densitometric analysis of immunoblots representative for HSPA2 expression (mean ± SD; n ≥ 4). The protein level is presented relative to β-actin. Statistical significance was determined by the two-tailed t-test performed in regard to cells exposed to DMSO solvent only. *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 3

Proteasome inhibition stimulates extracellular release of the HSPA2 and HSPA1 proteins. (a, d) Intracellular and (b, e) extracellular levels of HSPA1 and HSPA2 in cells exposed to bortezomib (BTZ) or manumycin A (MA). Extracellular HSPAs were detected in concentrated conditioned FBS-free media (Opti-MEM™). Samples were prepared in relation to the sample volume (µl) or protein concentration (µg; in (B) only); concentrated Opti-MEM™ only sample was tested as a negative control. Prior to incubation with primary antibody membranes were cut according to the 55 kDa molecular ladder band. X-ray film was used for chemiluminescent signal detection. Original autoradiograms/immunoblots are presented in Fig. S4. Graphs in (a, b, d, e) show results of densitometric analysis of immunoblots for HSPA1 and HSPA2 detection (mean ± SD; n ≥ 3). In (b) ‘volume’ group was quantified. (c) Results of lactate dehydrogenase (LDH) release detection assay are expressed in relation to the DMSO-treated control cells (mean values ± SD; n = 3, each in duplicate). Statistical significance was determined using the two-tailed t-test. *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 4

HSPA2 is present in SEC-purified small extracellular vesicles (EVs) from NCI-H1299 cells. (a) Levels of HSPA2, HSPA1, and EVs marker in EVs derived from non-treated and BTZ-treated (32 nM, 8 h) NCI-H1299 cells. Cells were exposed to BTZ in FBS-free medium (Opti-MEM™); GM130 was used as a cell lysate-contamination control. Membranes were cut into fragments according to the proteins’ molecular weight. For chemiluminescent signal detection X-ray film was used. Original autoradiograms/immunoblots are presented in Fig. S5. (b) Number of particles isolated by SEC from conditioned cell culture media from non-treated (DMSO) or BTZ-treated cells, measured by nanoparticle tracking analysis (NTA). EVs numbers were calculated in relations to the number of cells. (c) Representative histogram of particle size distribution in EVs sample obtained from non-treated (DMSO) cells, calculated by NTA measurement.

Figure 5

Specificity of HSPA2 detection in SEC-purified EVs derived from HSPA2-knockout and HSPA2-overproducing NCI-H1299 cells. (a) Analysis of HSPAs in HSPA2-knockout cells modified by CRISPR/Cas9 system. WT, wild-type cells, CTR I, CTR II, modified control cells; MIX I, MIX II; pools of isogenic clones with partial and full knockout of the HSPA2 gene expression, respectively. (b) Levels of HSPA2; positive (CD63, CD81, TSG101) and negative (GM130) protein markers in EVs produced by wild-type (WT), modified control (CTR I, CTR II) and HSPA2-knockout cells (MIX I, MIX II). (c) Expression of HSPAs and GFP-HSPA2 fusion protein in WT, control GFP tag-overexpressing (p-GFP); GFP-HSPA2-overexpressing (p-GFP-A2) cells. Stable cell lines were established by lentiviral transduction. (d) Detection of HSPA2 in EVs-enriched SEC fraction derived from WT, control p-GFP and p-GFP-A2 cells. WT HSPA2, 70 kDa; GFP-HSPA2 fusion protein, 100 kDa. Experiments were repeated twice for each HSPA2 model (HSPA2-knockout or GFP-tagged). Representative immunoblots are presented (n = 2). β-actin was used as a protein loading control. Membranes were cut into two (or three) fragments according to the proteins’ molecular weight. For chemiluminescent signal detection X-ray film was used. Original autoradiograms/immunoblots are presented in Fig. S6.

Figure 6

HSPA2 is present in small extracellular vesicles (EVs) derived from different biological sources. (a) Western blot detection of HSPA2 and HSPA1 in EVs released into cell culture media by non-cancerous (HaCaT) and cancer (NCI-H1299, FaDu) cells; CD63, CD9, CD81, the EVs protein markers; GM130, negative EVs marker. (b) Intracellular protein levels of HSPA2 and HSPA1. β-actin was used as a protein loading control. (c) Detection of HSPA1, HSPA2 and HSPA8 in SEC-purified EVs derived from urine of female healthy donors (A-C). (d) Detection of HSPA2 and EVs’ markers in SEC-purified EVs derived from urine of two independent groups of males, each compromising four healthy donors (samples 1–4; 5–8) and eight patients with prostate cancer (samples 9–16; 17–24). Membranes were cut into two (or more) fragments according to the proteins’ molecular weight. For chemiluminescent signal detection X-ray film was used. Original autoradiograms/immunoblots are presented in Fig. S7. (e) HSPA2-TSG101 densitometry correlation analysis; results obtained using ImageJ software ;. Each dot represents intensity values of HSPA2 and TSG101 bands calculated from immunoblots shown in (d).

Figure 7

Immunodetection of HSPA2, EVs markers, and prostate-related marker (PSMA) in urinary EVs isolated using antibody-conjugated beads. (a) Beads were coated with the mixture of anti-CD63, anti-CD81 and anti-CD9 antibodies in equal proportions or (b) with anti-PSMA antibody. Urine samples from one male healthy donor (HD) and two prostate cancer patients (PC1; PC2) were tested. A sample with PBS and coated beads was used as a control for non-specific antibody interaction during Western blot (WB) procedure. Antibody fragments detached from the coated beads during samples preparation for WB were visualized with Ponceau S staining as a control of the sample loading and electrotransfer performance. Membranes were cut into fragments according to the proteins’ molecular weight. For chemiluminescent signal detection X-ray film was used. Original autoradiograms/immunoblots are presented in Fig. S8.