Intercellular transmission of the unfolded protein response promotes survival and drug resistance in cancer cells

Abstract

Increased protein translation in cells and various factors in the tumor microenvironment can induce endoplasmic reticulum (ER) stress, which initiates the unfolded protein response (UPR). We have previously reported that factors released from cancer cells mounting a UPR induce a de novo UPR in bone marrow-derived myeloid cells, macrophages, and dendritic cells that facilitates protumorigenic characteristics in culture and tumor growth in vivo. We investigated whether this intercellular signaling, which we have termed transmissible ER stress (TERS), also operates between cancer cells and what its functional consequences were within the tumor. We found that TERS signaling induced a UPR in recipient human prostate cancer cells that included the cell surface expression of the chaperone GRP78. TERS also activated Wnt signaling in recipient cancer cells and enhanced resistance to nutrient starvation and common chemotherapies such as the proteasome inhibitor bortezomib and the microtubule inhibitor paclitaxel. TERS-induced activation of Wnt signaling required the UPR kinase and endonuclease IRE1. However, TERS-induced enhancement of cell survival was predominantly mediated by the UPR kinase PERK and a reduction in the abundance of the transcription factor ATF4, which prevented the activation of the transcription factor CHOP and, consequently, the induction of apoptosis. When implanted in mice, TERS-primed cancer cells gave rise to faster growing tumors than did vehicle-primed cancer cells. Collectively, our data demonstrate that TERS is a mechanism of intercellular communication through which tumor cells can adapt to stressful environments.

Fig. 1. Prostate cancer cells undergoing ER stress can transmit an ER stress response to recipient cells

(A) Expression of the indicated mRNA (by RT-qPCR) in PC3 cells cultured for 1, 3, or 5 days in Veh CM or TERS CM (n = 2 per condition). Gene expression was normalized to Veh CM day 1. RQ, relative quantification. Inset shows gel banding for unspliced (XBP-1u) and spliced (XBP-1s) XBP-1. (B) Western blot analysis for GRP78 abundance in whole-cell lysates from PC3 cells cultured as described in (A). V, Veh CM; T, TERS CM. (C) RT-qPCR in DU145 cells as described in (A) treated with PC3 generated Veh CM or TERS CM (n = 2 per condition). Gene expression was normalized to Veh CM day 1 condition to determine relative quantification. (D) RT-qPCR analysis for IL-6 expression in PC3 cells cultured with Veh CM or TERS CM as described in (A). Values are normalized to Veh CM day 1 (n = 2 per condition). (E) Confocal microscopy for GRP78 in Veh CM– or TERS CM–treated PC3 cells for 48 hours. Scale bars, 25 µm. (F) Flow cytometry analysis of surface abundance of GRP78 (sGRP78) in Veh CM– or TERS CM–cultured, unpermeabilized PC3 cells. Data are means ± SEM; *P < 0.05, **P < 0.01, ***P < 0.001, paired two-tailed Student’s t tests. Data in (C) to (E) are representative of two experiments; data in (A), (B), and (F) are from three independent experiments.

Fig. 2. TERS-primed cancer cells display a unique UPR and are protected against nutrient deprivation

(A) Treatment design of TERS priming: 2-day culture in Veh CM or TERS CM followed by 2-day rest period. Cells were then challenged and analyzed as indicated. (B) Flow cytometry analysis for surface abundance of GRP78 in vehicle- or TERS-primed PC3 cells grown in standard growth medium. (C) Western blot analysis of GRP78 in vehicle (V)– or TERS (T)–primed PC3 cells after 48-hour culture in standard growth medium (cDMEM) or in nutrient-deprived condition [-Glu/FBS (fetal bovine serum)]. (D) Western blot analysis of proteins of the PERK pathway in vehicle- or TERS-primed PC3 cells after 48-hour culture in cDMEM or in -Glu/FBS. (E) Apoptosis analysis by flow cytometry detection of annexin V in vehicle- or TERS-primed PC3 cells after 48-hour culture in cDMEM or in -Glu/FBS (each plot represents at least 10,000 events per condition). Data in (D) are representative of two experiments; data in (B), (C), and (E) are from three or more independent experiments.

Fig. 3. Proteasome inhibition–mediated cytotoxicity is less effective in TERS-primed cells

(A) Western blot analysis of GRP78 in vehicle (V)– or TERS (T)–primed PC3 cells after 24-hour culture with various concentrations of bortezomib (Bz). (B) Flow cytometry analysis for the abundance of GRP78 (sGRP78) on the surface of vehicle- or TERS-primed PC3 cells 24 hours after EtOH control solution or bortezomib (100 nM) treatment. Data are shown individually and as an overlay for comparison. (C) Western blot analysis of PERK pathway proteins of vehicle- or TERS-primed PC3 cells 24 hours after addition of bortezomib. (D) Apoptosis analysis by annexin V/propidium iodide (PI) staining in vehicle- or TERS-primed PC3 cells 24 hours after addition of bortezomib (each plot represents at least 10,000 events per condition). (E) Western blot of GRP78 expression in vehicle- or TERS-primed PC3 cells cultured in cDMEM and harvested at the specified postpriming day. (F) Percent live cells determined by flow cytometry analysis of annexin V/PI apoptosis staining in vehicle- or TERS-primed PC3 cells treated with EtOH control solution or bortezomib (100 nM). Data in (A), (C), and (E) are representative of two experiments; data in (B) and (F) are from three experiments; data in (D) are from five independent experiments.

Fig. 4. TERS-primed cells are protected against genotoxic insults in the absence of ER stress

(A) Vehicle- or TERS-primed PC3 cells treated with increasing concentrations of paclitaxel and analyzed after 24 hours by RT-qPCR for relative gene expression of UPR genes. Samples were normalized to 0 µM vehicle–primed gene expression (n = 2 per condition). (B) Western blot analysis of GRP78 in PC3-primed cells treated with paclitaxel for 24 hours. (C) Western blot analysis of PERK signaling in PC3 vehicle- or TERS-primed cells treated with paclitaxel for 24 hours. PC3 cells treated with Tg (300 nM) serve as positive control. (D) Annexin V apoptosis assay of primed PC3 cells untreated or treated with paclitaxel for 48 hours (each plot represents at least 10,000 events per condition). (E) Cell cycle analysis as determined by BrdU incorporation in vehicle- or TERS-primed PC3 cells (each plot represents at least 10,000 events per condition). (F) DNA double-stranded breaks visualized through γ-H2AX staining (pink) and imaged by confocal microscopy in vehicle- or TERS-primed PC3 cells after 24-hour treatment with paclitaxel (1 µM). Scale bars, 100 µm. Data in (B), (C), and (F) are representative of two experiments; data in (A), (D), and (E) are from three or more independent experiments.

Fig. 5. TERS induces WNT signaling and cytoplasmic export of TERT

(A) RT-qPCR analysis of the transcriptional activation of CTNNB1 and AXIN2 of Veh CM– or TERS CM–treated PC3 cells throughout 5 days of culture with CM resupplementation every other day. Relative quantification of gene expression of samples is normalized to Veh CM day 1 condition (n = 2 per group). (B) TOP-GFP expression of PC3 cells during TERS priming, determined by flow cytometry (at least 10,000 events were analyzed per condition). PC3.TOP cells treated with Veh CM or TERS CM for 48 hours in the absence or presence of (C) the IRE1α inhibitor 4μ8C or (D) the PERK inhibitor GSK2656157 and measured for mean fluorescence intensity (MFI). MFI expression was then normalized to Veh CM uninhibited value (n = 2; at least 10,000 events were analyzed per condition). (E) Normalized MFI expression of PC3.TOP cells treated with tunicamycin (5 µg/ml) for 48 hours during IRE1α or PERK inhibition (n = 2; at least 10,000 events were analyzed per condition). (F) Normalized MFI expression of PC3.TOP stimulated for 48 hours with recombinant WNT3a (rWNT3a) (20 ng/ml) during IRE1α or PERK inhibition (n = 2; at least 10,000 events were analyzed per condition). (G) hTERT RT-qPCR analysis of 48-hour Veh CM– or TERS CM–treated PC3 cells (n = 3 per condition). (H) Relative firefly TERT promoter-luciferase or ATF6 promoter-luciferase of dually transfected LNCaP cells. Cells were treated with LNCaP Veh CM or TERS CM for 48 hours and normalized for expression by Renilla-luciferase (n = 3 per condition). **P < 0.01, Student’s t test (paired two-tailed). (I) Immunofluorescence staining for TERT in PC3 cells treated for 48 hours. Scale bars, 25 µm. Error bars represent SEM. Data in (C), (D), and (H) are representative of two experiments; data in (A), (B), (F), (G), and (I) are from at least three independent experiments.

Fig. 6. PERK signaling is necessary for TERS-mediated cytoprotection

(A and B) Survival of (A) wild-type (WT) or (B) PERK KO MEFs primed by TC1 Veh CM or TERS CM and challenged for 48 hours as specified (Tx, paclitaxel). Survival determined by flow cytometry analysis via 7AAD exclusion. Percent (%) survival calculated by normalizing the percent live (7AAD⁻) population of the unstimulated condition for each line (n = 2 per group; at least 10,000 events analyzed per condition). (C) CRISPR/Cas9 design of guide targets within the ATF4 gene. (D) PCR detection of ATF4 WT and ATF4 CRISPR 293XT cells. WT 293XT (E) and ATF4 CRISPR 293XT (F) cells were primed with PC3 Veh CM or TERS CM, and survival was measured by 7AAD exclusion after 48 hours of treatment, as specified (n = 2 per condition; at least 10,000 events were analyzed per condition). Error bars represent SEM. Data in (A), (B), and (D) are representative of two experiments; data in (E) and (F) are from three independent experiments.

Fig. 7. TERS-primed murine prostate cancer cells have improved cellular fitness in vitro and in vivo

(A) Cartoon of coculture experimental design. Briefly, RFP-tagged TC1 cells (TC1-RFP) are primed with homologous Veh CM, whereas untagged TC1 cells are primed with TERS CM. After priming, cell populations are cocultured overnight and subsequently challenged. (B) Flow cytometry analysis to determine the percent RFP⁺ (vehicle-primed) and RFP⁻ (TERS-primed) TC1 cells, 7AAD excluded, after 24-hour treatment with Tg, 2-deoxyglucose (2DG), bortezomib, or paclitaxel (n = 2 per coculture; at least 10,000 events were analyzed per condition). (C) Growth kinetics of vehicle- or TERS-primed TC1 cells subcutaneously injected into immunocompetent C57BL/6 mice (n = 5 per group). (D) Weight of vehicle- or TERS-primed tumors after 30 days. (E) Gross visualization of excised tumors. **P < 0.01, Student’s t test (paired two-tailed). Error bars are SEM. Data are representative of two independent experiments.

Fig. 8. Model of TERS-mediated signaling in cancer cells

We propose a model in which cancer cells exposed to TERS undergo an adaptive UPR that involves diverse signaling events. One effect is the activation of Wnt signaling. TERS drives Wnt signaling through the activation of the TCF. This effect appears to be IRE1α-dependent. The other relevant event is cytoprotection. In this case, TERS engages PERK but also leads to reduced ATF4 activation. Reduced levels of ATF4 are insufficient to drive full activation of apoptosis through the downstream CHOP target (red strikethrough). In this respect, ATF4 serves as a rheostat for cell survival. TERS also increases the amounts of GRP78, both intracellularly and at the cell surface. Finally, TERS induces the export of TERT to the cytoplasm. These effects, possibly in combination, promote cytoprotection and, ultimately, cell fitness to endogenous (nutrient) and exogenous (chemotherapeutic) stress.