Tumor cells express and maintain HMGB1 in the reduced isoform to enhance CXCR4-mediated migration

Abstract

During inflammation and tissue regeneration, the alarmin High Mobility Group Box 1 (HMGB1), in its reduced isoform, enhances the activity of the chemokine CXCL12, forming a heterocomplex that acts via the chemokine receptor CXCR4. Despite the established roles of both HMGB1 and CXCL12 in tumor progression and metastatic spread to distal sites, the role of the CXCL12/HMGB1 heterocomplex in cancer has never been investigated. By employing a newly established mass spectrometry protocol that allows an unambiguous distinction between reduced (red-HMGB1) and oxidized (ox-HMGB1) HMGB1 isoforms in cell lysates, we demonstrate that human epithelial cells derived from breast (MCF-7 and MDA-MB-231) and prostate (PC-3) cancer predominantly express red-HMGB1, while primary CD3⁺ T lymphocytes from peripheral blood express both HMGB1 isoforms. All these cancer cells release HMGB1 in the extracellular microenvironment together with varying concentrations of thioredoxin and thioredoxin reductase. The CXCL12/HMGB1 heterocomplex enhances, via CXCR4, the directional migration and invasiveness of cancer cells characterized by high metastatic potential that possess a fully active thioredoxin system, contributing to maintain red-HMGB1. On the contrary, cancer cells with low metastatic potential, lack thioredoxin reductase, promptly uptake CXCL12 and fail to respond to the heterocomplex. Our study demonstrates that the responsiveness of cancer cells to the CXCL12/HMGB1 heterocomplex, resulting in enhanced cell migration and invasiveness, depends on the maintenance of HMGB1 in its reduced isoform, and suggests disruption of the heterocomplex as a potential therapeutic target to inhibit invasion and metastatic spread in cancer therapies.

Keywords: CXCL12; HMGB1; breast cancer; heterocomplex; prostate cancer; thioredoxin.

Figure 1

CXCL12 and HMGB1 are expressed in MCF-7, MDA-MB-231, and PC-3 cancer cell lines. (A, B) qRT-PCR analysis (normalized on 18S) is shown for CXCL12 and HMGB1 mRNA in cancer cells. Data are shown as mean ± SEM of 5 independent experiments. Statistical analysis of the differences among the cell lines was performed using one way ANOVA, followed by Tukey’s multiple comparison test: **p < 0.01. (C) CXCL12 and HMGB1 expression assessed by confocal microscopy. In the merged images, CXCL12 is shown in red and HMGB1 in green. Nuclei and cell membranes were counterstained with DAPI (blue) and MemBrite Fix Dye (cyan), respectively. Representative images, of at least 3 independent experiments, are shown. Scale bar represents 20 µm. (D, E) CXCL12 (D) and HMGB1 (E) expression was evaluated in at least 40 cells and expressed as fluorescence intensity. Horizontal lines represent means. Statistical analysis of the differences among the cell lines was performed using one way ANOVA, followed by Tukey’s multiple comparison test: *p < 0.05, ****p < 0.0001. (F) Confocal microscope z-stack overlays showing the intracellular localization of CXCL12 in MCF-7 cells (scale bar 20 μm). CXCL12 is shown in red, while nuclei were counterstained with DAPI (blue) and cell membranes with the MemBrite Fix Dye (cyan).

Figure 2

Endogenous redox status of HMGB1 in cancer cells, and release of HMGB1, Trx and TrxR1 in the extracellular milieu. (A) Annotated MS/MS spectra of the most abundant peptide forms containing either the light (blue) or heavy (red) iodoacetamide derivative-modified form of each of the HMGB1 cysteines identified with the recombinant full length HMGB1. Green and orange lines represent the annotated fragment peaks for each MS/MS spectrum, where green marks the b-ion series and orange marks the y-ion series. The resulting fragmentation pattern is reported on the peptide sequence with the same color code to visualize the peptide spectrum match. In grey the common methionine oxidation modification is depicted. Andromeda identifications are visualized with PDV (http://pdv.zhang-lab.org). (B) Percentage of reduced and oxidized HMGB1 based on the redox status of C23 and C106 in MCF-7 (n=3), MDA-MB-231 (n=3), PC-3 (n=3), and in CD3⁺ T cells from healthy donors (n=2). Data are shown as mean + SEM of the independent experiments performed. (C) HMGB1 concentration detected by ELISA in the cell culture supernatants. Data present the mean ± SEM of 3 independent experiments. (D) qRT-PCR analysis (normalized on 18S) is shown for TRX and TRXR1 mRNA in cancer cells. Data are shown as mean ± SEM of 3 independent experiments. (E) Concentration of Trx and TrxR1 detected by ELISA in the cell culture supernatants. Data present the mean ± SEM of 3 independent experiments. Statistical analysis of the differences among the cell lines was performed using one way ANOVA, followed by Tukey’s multiple comparison test: *p < 0.05.

Figure 3

CXCL12 uptake by cancer cells. (A) qRT-PCR analysis (normalized on 18S) is shown for CXCR4 mRNA expression in cancer cells. Data are shown as mean ± SEM of 5 independent experiments. (B) CXCR4 expression analyzed by flow cytometry in MCF-7, MDA-MB-231, and PC-3 cells. Data are shown as relative MFI (rMFI, mean ± SEM) of 3 independent experiments. (C) Uptake of CXCL12-Atto647 by cancer cells analyzed by flow cytometry. Left: A representative histogram showing fluorescence intensity of unstimulated (blue) and CXCL12-Atto647 stimulated cells (red). Right: CXCL12-Atto647 uptake shown as rMFI (mean ± SEM) of 5 independent experiments. (D) Uptake of CXCL12-Atto674 by cancer cells treated with the CXCR4 antagonist AMD3100. Chemokine uptake shown as rMFI (mean ± SEM) of 6 independent experiments. Statistical analysis of the differences among the cell lines was performed using two-way ANOVA, followed by Šídák’s multiple comparisons test: ****p < 0.0001. (E) CXCL12-Atto647 uptake assessed by confocal microscopy in MCF-7, MDA-MB-231, and PC-3 cells. Left: In the merged images, CXCL12 is shown in red. Nuclei and cell membranes were counterstained with DAPI (blue) and MemBrite Fix Dye (cyan), respectively (scale bars: 20 μm). Right: CXCL12-Atto647 uptake was evaluated in at least 10 cells and expressed as fluorescence intensity. Horizontal lines represent the mean values. Statistical analysis of the differences among the cell lines was performed using one way ANOVA, followed by Tukey’s multiple comparison test: *p < 0.05, ****p < 0.0001.

Figure 4

The CXCL12/HMGB1 heterocomplex enhances cancer cells wound healing, migration and invasion. (A) Wound healing in response to CXCL12, HMGB1 or the CXCL12/HMGB1 heterocomplex in PC-3 cells. Top panel: data are shown as area covered by migrating cells over time and represent the mean + SEM of at least 3 independent experiments. Bottom panel: dot plot representing statistical differences in covered area over time by stimulated cells compared to control cells. Circle size represents statistics; colors represent the fold-change increase (red) or decrease (blue), shown as ratio of the covered area of stimulated over control cells. Statistical analysis was performed using two-way ANOVA, followed by Dunnett’s multiple comparison test. (B) Inhibition of HMGB1-induced wound healing by AMD3100, Bordetella pertussis toxin (PTX), glychyrrizin and a neutralizing antibody targeting CXCL12 (α-CXCL12). Data are shown as fold increase of the covered area after 3 h compared to unstimulated PC-3 cells. Statistical analysis was performed using unpaired t-test: *p < 0.05, **p < 0.01, ***p < 0.001. (C) Cell migration assay. MDA-MB-231 cells were allowed to migrate in response to the indicated concentrations of CXCL12 and/or HMGB1 for 18 h. Representative spider plots, showing the trajectories of at least 60 single tracked cells migrating following (red) or moving in the opposite direction (blue) of the gradient, are shown. Blue and red dots in the plots represent the final position of each single tracked cell. (D) Quantitative evaluation of the Forward Migration Index (FMI) is shown. Data represent the mean + SEM of at least 2 independent experiments. (E) Quantitative evaluation of the accumulated distance from the experiments shown in (D). Statistical analysis was performed using unpaired t-test: **p < 0.01. (F, G) Invasion assay. Percentage of MDA-MB-231 (F) or MCF-7 (G) cells invading the matrigel 48 h after stimulation with the indicated concentrations of CXCL12 and/or HMGB1 in at least 5 fields. Horizontal lines represent means ± SEM. Statistical analysis was performed using one-way ANOVA, followed by Tukey’s multiple comparison test: **p < 0.01, ***p < 0.001.