Resistance to paclitxel in breast carcinoma cells requires a quality control of mitochondrial antiapoptotic proteins by TRAP1

Abstract

TRAP1 is a mitochondrial antiapoptotic protein up‐regulated in several human malignancies. However, recent evidences suggest that TRAP1 is also localized in the endoplasmic reticulum (ER) where it is involved in ER stress protection and protein quality control of tumor cells. Based on the mechanistic link between ER stress, protection from apoptosis and drug resistance, we questioned whether these novel roles of TRAP1 are relevant for its antiapoptotic function. Here, we show for the first time that: i) TRAP1 expression is increased in about 50% of human breast carcinomas (BC), and ii) the ER stress protecting activity of TRAP1 is conserved in human tumors since TRAP1 is co‐upregulated with the ER stress marker, BiP/Grp78. Notably, ER‐associated TRAP1 modulates mitochondrial apoptosis by exerting a quality control on 18 kDa Sorcin, a TRAP1 mitochondrial client protein involved in TRAP1 cytoprotective pathway. Furthermore, this TRAP1 function is relevant in favoring resistance to paclitaxel, a microtubule stabilizing/ER stress inducer agent widely used in BC therapy. Indeed, the transfection of a TRAP1 deletion mutant, whose localization is restricted to the ER, in shTRAP1 cells enhances the expression of mitochondrial Sorcin and protects from apoptosis induced by ER stress agents and paclitaxel. Furthermore, BC cells adapted to paclitaxel or ER stress inducers share common resistance mechanisms: both cell models exhibit cross‐resistance to single agents and the inhibition of TRAP1 by siRNAs or gamitrinib, a mitochondria‐directed HSP90 family inhibitor, in paclitaxel‐resistant cells rescues the sensitivity to paclitaxel. These results support the hypothesis that ER‐associated TRAP1 is responsible for an extramitochondrial control of apoptosis and, therefore, an interference of ER stress adaptation through TRAP1 inhibition outside of mitochondria may be considered a further compartment‐specific molecular approach to rescue drug‐resistance.

Highlights

• TRAP1 and BiP/Grp78 are co‐upregulated in human breast carcinomas.

• TRAP1 protein quality control in the ER is relevant for its antiapoptotic function.

• TRAP1 quality control on mitochondrial Sorcin is involved in apoptosis prevention.

• TRAP1 mediates resistance to paclitaxel.

• Pharmacological inhibition of TRAP1 reverts resistance to paclitaxel.

1. Introduction

TRAP1 is a molecular chaperone, a component of the HSP90 family proteins, with prevalent mitochondrial localization and antiapoptotic and antioxidant proprieties (Matassa et al., 2012). Its best‐characterized function is protection against mitochondrial apoptosis by regulating the opening of the mitochondrial transition pore (MTP), through interaction with HSP90 and cyclophillin D (Kang et al., 2007; Chae et al., 2012). Since TRAP1 is selectively up‐regulated in several human malignancies (i.e., prostate, colorectal, nasopharyngeal, and ovarian carcinomas) (Fang et al., 2008; Costantino et al., 2009; Leav et al., 2010; Landriscina et al., 2010a), its antiapoptotic function is relevant in protecting against the cytotoxic activity of antiblastic agents as well as favoring a multidrug resistant phenotype, as demonstrated by our group in human colorectal carcinoma (Costantino et al., 2009; Landriscina et al., 2010b). In this context, as TRAP1 is regarded as a molecular target for cancer therapy, mitochondria‐directed agents have been designed to selectively inhibit HSP90 chaperones to induce apoptosis and revert drug resistance (Kang and Altieri, 2009; Altieri, 2011).

In order to deal with the constant challenge of protein misfolding in the endoplasmic reticulum (ER), eukaryotic cells have evolved an ER protein quality control mechanism which is part of an adaptive stress response to maintain protein homeostasis or proteostasis, since aberrant proteostasis can trigger cellular apoptosis (Liu and Ye, 2011). Seen in this light, chaperones and other protein homeostasis factors interact with newly translated polypeptides to facilitate their folding and correct localization (del Alamo et al., 2011). Dysregulation of this process is associated with the development and progression of cancer (Ozcan and Tabas, 2012), while these adaptive stress response systems represent attractive targets for cancer therapy (Healy et al., 2009).

Recent findings have pointed to the putative role of TRAP1 in protecting cancer cells from the accumulation of unfolded proteins in the ER (Chae et al., 2012; Takemoto et al., 2011; Siegelin et al., 2011; Amoroso et al., 2012). Initial studies suggested that the direct inhibition of HSP90 chaperones in mitochondria is associated to an organelle unfolded protein response (UPR) triggering compensatory autophagy and secondarily ER stress (Siegelin et al., 2011). For the first time, our group recently described the presence of TRAP1 at the interface between ER and mitochondria, where this chaperone is functionally involved in ER stress protection through its interaction with TBP7, an AAA‐ATPase of the 19S proteasomal subunit. These studies suggested that the TRAP1 network is responsible for quality control of specific TRAP1‐client proteins (i.e., 18 kDa Sorcin and F1ATPase), destined to mitochondria and involved in protection from apoptosis, Ca²⁺ homeostasis and metabolism. Indeed, we demonstrated that expression of both proteins is decreased upon TRAP1 and/or TBP7 interference as a consequence of their increased ubiquitination (Amoroso et al., 2012). Interestingly, the mechanisms of protein quality control driven by the TRAP1/TBP7 network are conserved in human malignancies, since TRAP1, TBP7, 18 kDa Sorcin and F1ATPase are co‐expressed in colorectal tumors, thereby supporting the hypothesis that such a pathway may be relevant in tumor progression and drug resistance (Amoroso et al., 2012).

While these findings reveal the complexity of the mechanisms responsible for protein homeostasis in organelles, the relationship between TRAP1 protein quality control in the ER and its antiapoptotic function in mitochondria is still open to debate. This study was designed to address the hypothesis that TRAP1 function in ER stress protection and quality control of specific client proteins is relevant for its antiapoptotic activity and may contribute to induction of drug resistant phenotypes. This hypothesis was investigated in breast carcinoma (BC) cell models in relation to paclitaxel cytotoxicity. Paclitaxel is actually one of the most active agents in the treatment of human BC (Murray et al., 2012) and several reports suggest that its apoptotic activity relies on the induction of ER stress (Liao et al., 2008; Wang et al., 2009; Mhaidat et al., 2009). We demonstrated that ER‐associated TRAP1 regulates mitochondrial apoptosis by regulating the quality of specific client proteins and, among others, Sorcin, thus suggesting that TRAP1 may be a key player in coupling organelle proteostasis, adaptation to stress and cell survival, through the regulation of the mitochondrial apoptotic pathway.

2. Patients and methods

2.1. Tumor specimens, chemicals, cell cultures, constructs and siRNAs

Specimens from both 28 BCs and the corresponding normal, non‐infiltrated glands were obtained from the IRCCS‐CROB Basilicata Biobank. Following surgical removal of the tumor, samples were divided into small pieces 5–10 mm in diameter, with one piece stored in a cryomold, covered with cryogel OCT (Histolab, Goteborg, Sweden), snap frozen in vapor of liquid nitrogen and transferred to a low‐temperature freezer (−80 °C); and another stored in a cryovial, snap frozen, stored at −196 °C in liquid nitrogen vapor and used for Real Time PCR analysis. Another part of the sample was formaline‐fixed before further processing and paraffin‐embedding. Express written informed consent to use biological specimens for investigational procedures was obtained from all patients. Patient characteristics are reported in Supplementary Table 1.

Table Table 1.

TRAP1 protein levels and BiP/Gpr78 mRNA levels in 28 human breast carcinomas expressed as time increase in tumor compared to correspondent non‐infiltrated peritumoral gland.

Case N°	TRAP1 protein levels	BiP/Grp78 mRNA levels
1.	0.3	0.5
2.	2.8	10.3
3.	0.9	5.1
4.	12.9	8.0
5.	8.5	0.1
6.	3.6	2.7
7.	2.7	0.1
8.	4.4	66.7
9.	0.8	1.2
10.	1.0	0.3
11.	0.5	0.9
12.	0.6	2.9
13.	3.6	15.1
14.	1.8	0.1
15.	7.1	2.1
16.	1.0	0.5
17.	3.3	3.4
18.	0.1	n.d.
19.	12.5	3.2
20.	0.1	2.4
21.	5.3	2.8
22.	0.5	0.1
23.	2.7	n.d.
24.	2.6	n.d.
25.	0.4	1.1
26.	11.1	n.d.
27	2.5	n.d.
28.	2.9	2.4

n.d., not done.

Unless otherwise specified, reagents were purchased from Sigma–Aldrich (Milan, Italy). Gamitrinib was kindly provided by Dr. Altieri (The Wistar Institute, Philadelphia, PA, USA). Human MCF7 and MDA‐MB231 BC cells were purchased from ATCC (Manassas, VA, USA) and cultured in DMEM containing 10% (v/v) fetal bovine serum, 1.5 mM glutamine, and 100 units/ml penicillin and streptomycin. Cell lines were routinely monitored in our laboratory by microscopic morphology and by DNA profiling and immunocytochemistry for estrogen receptor (MCF7 cells) and immunocytochemistry for EGF receptor (MDA‐MB231 cells), in line with ATCC product description. MCF7 cells resistant to paclitaxel (Paclit‐R MCF7), bortezomib (Bort‐R MCF7) or thapsigargin (Tg‐R MCF7) were selected as previously reported (Barone et al., 2007). Cells were finally stabilized in the presence of 300 nM paclitaxel, 20 nM bortezomib, and 600 nM thapsigargin.

TRAP1 stable interference was achieved by transfecting MCF7 cells with TRAP1 (TGCTGTTGACAGTGAGCGACCCGGTCCCTGTACTCAGAAATAGTGAAGCCACAGATGTATTTCTGAGTACAGGGACCGGGCTGCCTACTGCCTCGGA) or scrambled (sequence containing no homology to known mammalian genes) shRNA (Open Biosystem, Huntsville, AL, USA) (Landriscina et al., 2010b). siRNAs of TRAP1 were purchased from Qiagen (Milan, Italy; Cat. No. SI00115150 for TRAP1, SI03650318 for negative control). For knock‐down experiments, siRNAs were diluted to a final concentration of 40 nM and transfected using HiPerFect Transfection Reagent according to manufacturer protocol (Qiagen, Milan, Italy). Constructs encoding for wild type TRAP1 and TRAP1 mutants Δ1‐59TRAP1‐Myc (Amoroso et al., 2012) were transiently transfected with Polyfect Transfection reagent (Qiagen, Milan, Italy).

2.2. Immunoblot analysis

Total cell lysates were obtained by homogenization of cell pellets and tissue samples in a cold lysis buffer (20 mM Tris, pH 7.5 containing 300 mM sucrose, 60 mM KCl, 15 mM NaCl, 5% (v/v) glycerol, 2 mM EDTA, 1% (v/v) Triton X‐100, 1 mM PMSF, 2 mg/ml aprotinin, 2 mg/ml leupeptin and 0.2% (w/v) deoxycholate) for 2 min at 4 °C and further sonication for 30 s at 4 °C. Mitochondria were purified by Qproteome Mitochondria Isolation kit (Qiagen) according to manufacturer protocol. Immunoblot analysis was performed as previously reported (Maddalena et al., 2011). Specific proteins were detected by using a rabbit polyclonal anti‐Sorcin antibody kindly provided by Prof. E. Chiancone (University of Rome “La Sapienza”), a mouse monoclonal anti‐GAPDH (sc‐47724, Santa Cruz Biotechnology, Segrate, Italy), a mouse monoclonal anti‐βactin (sc‐69879, Santa Cruz Biotechnology, Segrate, Italy), a mouse monoclonal anti‐TRAP1 (sc‐13557, Santa Cruz Biotechnology, Segrate, Italy), a rabbit polyclonal anti‐Caspase‐12 (SPA‐827; StressGen, Milan, Italy), a mouse monoclonal anti‐COX IV (MS407, Mitosciences, Milan, Italy), a rabbit polyclonal anti‐Grp94 (sc‐11402, Santa Cruz Biotechnology, Segrate, Italy), a mouse monoclonal anti‐Myc (sc‐40, Santa Cruz Biotechnology, Segrate, Italy), a rabbit polyclonal anti‐phospho PERK (Thr 981, sc‐32577, Santa Cruz Biotechnology, Segrate, Italy), a rabbit monoclonal anti‐PERK (#3192, Cell Signaling Technology, Boston, MA, USA), a rabbit polyclonal anti‐phospho eIF2α (Ser 51, #9271, Cell Signaling Technology, Boston, MA, USA) and a mouse monoclonal anti‐eIF2α antibody (sc‐133132, Santa Cruz Biotechnology, Segrate, Italy).

In order to compare expression in different tumor specimens, TRAP1 protein levels were quantified by densitometric analysis using the Quantity One 4.5.0 software (BioRad Laboratories GmbH, Segrate, Italy) and expressed as a time increase in tumors compared to levels in the respective peritumoral non‐infiltrated gland.

2.3. RNA extraction and Real Time RT‐PCR analysis

Total RNA from cell pellets and tumor specimens was extracted using the TRIzol Reagent (Invitrogen, San Giuliano Milanese, Milano, Italy). For first strand synthesis of cDNA, 1 μg of RNA were used in a 20 μl reaction mixture utilizing a Transcriptor First Strand cDNA Synthesis Kit (Roche, Mannheim, Germany). For Real Time PCR analysis, 0.5 ng of cDNA sample was amplified using the LightCycler 480 SYBR Green I Master (Roche, Mannheim, Germany) in an Light Cycler 480 (Roche, Mannheim, Germany). The following primers were used: BiP/Grp78, forward 5′‐GTGGAATGACCCGTCTGTC‐3′, reverse 5′‐CGTCTTTGGTTGCTTGGC‐3′, (PCR product 254 bp); GAPDH forward 5′‐AGGCTGAGAACGGGAAGC‐3′, reverse 5′‐CCATGGTGGTGAAGACGC‐3′, (PCR product 135 bp); TRAP1 forward 5′‐CGCAGCATCTTCTACGTGC‐3′, reverse 5′‐CTGATGAGTGCGCTCTCC‐3′ (PCR product 200 bp). Primers were designed to be intron spanning. Reaction condition were as follows: pre‐incubation at 95 °C for 5 min, followed by 45 cycles of 10 s at 95 °C, 10 s at 60 °C, 10 s at 72 °C. GAPDH was chosen as an internal control.

2.4. Apoptosis assay

Apoptosis was evaluated by cytofluorimetric analysis of Annexin‐V and 7‐amino‐actinomycin‐D (7‐AAD)‐positive cells using the fluorescein isothiocyanate (FITC)–Annexin‐V/7‐AAD kit (Beckman Coulter, Milan, Italy). Stained cells were analyzed using the FACSCaliburTM (Becton Dickinson). Positive staining for Annexin‐V as well as double staining for Annexin‐V and 7‐AAD were interpreted as signs of early and late phases of apoptosis respectively (Landriscina et al., 2010b).

2.5. Statistical analysis

The paired Student's T test was used to establish the statistical significance between different levels of apoptosis or gene expression in controls and treated cells or in transfected cells and related scramble controls. The chi‐square test was used to establish statistical significance of TRAP1 and BiP/Grp78 co‐expression in human BCs. Statistically significant values (p < 0.05) are reported in the Results and Figures Legends.

3. Results

3.1. TRAP1 and BiP/Grp78 are co‐expressed in human BCs and induced by paclitaxel treatment in BC cells

Based on recent evidence that TRAP1 is involved in protection from ER stress in human malignancies (Siegelin et al., 2011; Amoroso et al., 2012), a cohort of human BCs and the respective normal, non‐infiltrated peritumoral glands was screened for TRAP1 expression by immunoblot analysis and the ER stress marker, BiP/Grp78 (Lee, 2005) by quantitative RT‐PCR. Figure 1A shows TRAP1 protein and BiP/Grp78 mRNA levels in 8 BCs, chosen as representative analytical results of all samples. Patient characteristics are reported in Supplementary Table 1, while protein and mRNA levels in the entire series are reported in Table 1. TRAP1 protein and BiP/Grp78 mRNA expression levels were seen as up‐regulated if they had increased at least twofold in comparison to the corresponding non‐infiltrated peritumoral gland. Indeed, TRAP1 and BiP/Grp78 were up‐regulated in respectively 57.1% (16/28 cases) and 56.5% (13/23 cases) of BCs. Interestingly, among 23 BCs characterized for both TRAP1 and BiP/Grp78 expression, we observed a significant co‐expression of the two genes (chi square test, p = 0.022). These data suggest that UPR is activated in a significant subgroup of human BCs, as a result of chronic exposure to ER stress conditions, and that, remarkably, TRAP1 may well be part of this cell response to ER stress.

Figure 1.

TRAP1 and BiP/Grp78 expression in human breast carcinomas and the role of ER stress induction in paclitaxel cytotoxicity. A. Total cell lysates from 8 human breast carcinomas (T) and the respective non‐infiltrated peritumoral gland (C) were separated by SDS–PAGE and immunoblotted with anti‐TRAP1 and anti‐GAPDH antibodies. BiP/Grp78 mRNA levels are reported as time increase in tumors compared to the respective control. B. Apoptotic levels in MCF7 cells treated with increasing concentrations of paclitaxel or 1 μM thapsigargin (Tg) for 24 h. p‐values versus vehicle‐treated cells: *p < 0.02; **p = 0.001; ***p < 0.0001. C. Bip/Grp78 mRNA levels in MCF7 cells treated with increasing concentrations of paclitaxel or 1 μM thapsigargin (Tg) for 24 and 48 h. p‐values versus vehicle‐treated cells: *p < 0.0001. D. Total cell lysates from MCF7 cells exposed to 1 and 10 μM paclitaxel for 24 h were separated by SDS–PAGE and immunoblotted with anti‐TRAP1, anti‐Grp94, anti‐Caspase12 and anti‐GAPDH antibodies. Error bars, ±SD.

Since adaptation to ER stress conditions represents a mechanism of resistance to taxanes (Liao et al., 2008; Wang et al., 2009; Mhaidat et al., 2009), we questioned whether the cytotoxic activity of paclitaxel may involve the induction of ER stress. Paclitaxel‐induced apoptosis was evaluated in parallel with markers of ER stress, i.e. BiP/Grp78, the ER‐associated HSP90 chaperone Grp94 (Argon and Simen, 1999), and the degradation product of pro‐caspase‐12 (Badiola et al., 2011). Indeed, paclitaxel induced apoptotic cell death in a dose‐dependent manner in MCF7 cells (Figure 1B) and, within the range of 1–10 μM, enhanced BiP/Grp78 (Figure 1C) and Grp94 expression, and caspase 12 activation (Figure 1D). It is noteworthy that, within the same concentration range, paclitaxel induced TRAP1 protein levels in BC MCF7 (Figure 1D) and MDA‐MB231 cells (Supplementary Figure 1A), without affecting its mRNA expression (Supplementary Figure 1B).

3.2. TRAP1 is relevant in protection from paclitaxel‐induced cytotoxicity

In order to analyze the role of TRAP1 in the protection from paclitaxel‐induced apoptosis, scramble and shTRAP1 MCF7 cells and MBA‐MB231 cells transfected with TRAP1 siRNA were exposed to increasing concentrations of paclitaxel (Figure 2A and Supplementary Figure 1C). TRAP1 knock down (KD) increased apoptosis upon paclitaxel treatment in both BC cell lines, whereas TRAP1 re‐expression in shTRAP1 resulted in significant protecting activity (Figure 2A). Analysis of the UPR response in shTRAP1 MCF7 cells showed reduced BiP/Grp78 levels compared to scramble cells (0.21 ± 0.06 versus 1.0 ± 0.14), in agreement with the co‐expression of TRAP1 and BiP/Grp78 in human BCs. By contrast, paclitaxel increased BiP/Grp78 expression in sh/siTRAP1 MCF7 and MDA‐MB231 cells (Figure 2B and Supplementary Figure 1D) and Grp94 levels and caspase 12 cleaved product in shTRAP1 compared to scramble cells (Figure 2C). Consistently with apoptotic results and the role of TRAP1 in protecting from ER stress, TRAP1 re‐expression in shMCF7 cells completely suppressed the induction of BiP/Grp78 (Figure 2B) as well as caspase 12 cleavage (Figure 2C). Furthermore, Grp94 levels were not further induced by paclitaxel upon TRAP1 up‐regulation, likely because its level was already enhanced (Figure 2C).

Figure 2.

TRAP1 protects from paclitaxel‐induced apoptosis and ER stress. A. Apoptotic levels in scramble and shTRAP1 MCF7 cells and in shTRAP1 cells transfected with pMock or TRAP1 cDNA (pTRAP1) treated with increasing concentrations of paclitaxel or 1 μM thapsigargin (Tg) for 24 h. p‐values versus scramble cells (*p = 0.014; **p < 0.0001; °°p = 0.004) or versus shTRAP1 cells transfected with pMock (#p < 0.0001; ***p < 0.006; °p = 0.003). B. BiP/Grp78 mRNA levels in scramble, shTRAP1 MCF7 cells transfected with pMock or TRAP1 cDNA (pTRAP1) treated with 10 μM paclitaxel or 1 μM thapsigargin (Tg) for 24 h p‐values versus scramble cells (*p < 0.0001) or versus shTRAP1 cells transfected with pMock (**p < 0.0001; ***p = 0.003). C. Total cell lysates from scramble and shTRAP1 cells transfected with pMock or TRAP1 cDNA (pTRAP1) exposed to 10 μM paclitaxel (paclit) for 24 h were separated by SDS–PAGE and immunoblotted with anti‐TRAP1, anti‐Grp94, anti‐Caspase12 and anti‐GAPDH antibodies.

In a parallel experiment, MCF7 cells were chronically adapted to paclitaxel and stabilized to final concentrations of 300 nM. In actual fact, paclitaxel‐resistant MCF7 cells exhibited increased levels of TRAP1, and Grp94 (Figure 3A), as well as up‐regulation of BiP/Grp78 expression (Figure 3B). Interestingly, in addition to being insensitive to paclitaxel, paclitaxel‐resistant cells exhibited resistance to agents that induce ER stress (i.e., thapsigargin and bortezomib), whereas down‐regulation of TRAP1 by siRNA (Figure 3C, insert) re‐established sensitivity to both paclitaxel and ER stress‐inducing agents (Figure 3C).

Figure 3.

TRAP1 is responsible for resistance to paclitaxel and ER stress agents in paclitaxel‐resistant MCF7 cells. A. Total cell lysates from scramble and paclitaxel‐resistant (Paclit‐R) MCF7 cells were separated by SDS–PAGE and immunoblotted with anti‐Grp94, anti‐TRAP1, and anti‐GAPDH antibodies. B. BiP/Grp78 mRNA levels in scramble and paclitaxel‐resistant (Paclit‐R) MCF7 cells. p‐values versus scramble cells: *p = 0.004. C. Apoptotic levels in scramble and paclitaxel‐resistant MCF7 cells upon transfection with control siRNA (Neg siRNA) and TRAP1 siRNA treated for 24 h with 1 μM paclitaxel, 1 uM thapsigargin (Tg) or 100 nM Bortezomib. p‐values versus scramble cells (*p < 0.0001; **p = 0.001) or versus paclitaxel‐resistant cells transfected with negative siRNA (***p < 0.0001). Insert: TRAP1 levels in paclitaxel‐resistant MCF7 cells upon transfection with negative (1) and TRAP1 siRNA (2).

Our finding of the role played by TRAP1 in resistance to paclitaxel‐induced apoptosis by protecting from ER stress prompted us to evaluate whether cross‐resistance to these agents could be observed in these cell populations. With this aim in mind, MCF7 cells adapted to chronic exposure to thapsigargin and bortezomib were selected and stabilized to growth in the presence of 600 nM thapsigargin and 20 nM bortezomib. Remarkably, thapsigargin‐ and bortezomib‐resistant cell lines exhibited increased Grp94 and TRAP1 levels (Figure 4A) and, interestingly, both cell models showed cross‐resistance to paclitaxel in addition to the respective selection agent (Figure 4B–C). It is worth noting that TRAP1 KD by siRNA in both ER stress‐adapted cell lines (Figure 4B–C, inserts) restored sensitivity to both the specific selection agent and paclitaxel (Figure 4B–C). These results suggest that adaptation to paclitaxel and ER stress share common molecular mechanisms and that TRAP1 is likely to play a role in resistance to paclitaxel by protecting from ER stress.

Figure 4.

Breast carcinoma cells adapted to thapsigargin and bortezomib are cross‐resistant to paclitaxel. A. Total cell lysates from scramble, thapsigargin (Tg)‐ and bortezomib (Bort)‐resistant MCF7 cells were separated by SDS–PAGE and immunoblotted with anti‐Grp94, anti‐TRAP1, and anti‐GAPDH antibodies. B. Apoptotic levels in scramble, bortezomib‐resistant (Bort‐R) and bortezomib‐resistant MCF7 cells transfected with TRAP1 siRNA treated with 100 nM bortezomib or 1 μM paclitaxel for 24 h. C. Apoptotic levels in scramble, thapsigargin‐resistant (Tg‐R) and thapsigargin‐resistant MCF7 cells transfected with TRAP1 siRNA treated with 1 μM thapsigargin or 1 μM paclitaxel for 24 h. B and C. p‐values versus scramble cells (*p < 0.0001; **p = 0.002) or versus drug‐resistant cells transfected with negative siRNA (***p < 0.0001) each treated with the same agent. Inserts (B–C): TRAP1 levels in bortezomib‐resistant (B) and thapsigargin‐resistant (C) MCF7 cells upon transfection with negative (1) and TRAP1 siRNA (2).

3.3. The role of TRAP1 in protein quality control in the endoplasmic reticulum is relevant for its antiapoptotic function

Since previous studies suggested that the antiapoptotic role played by TRAP1 relies mainly on its regulatory function on cyclophilin D and the opening of the MTP (Kang et al., 2007; Chae et al., 2012), we further evaluated the relevance of TRAP1 in protecting from ER stress and protein quality control for its antiapoptotic activity. To address this issue we used the Δ1‐59TRAP1 deletion mutant, which lacks the mitochondrial targeting sequence and is thus unable to enter into mitochondria (Figure 5A), while still able to bind TBP7 (Amoroso et al., 2012) and protect from ER stress (Figure 5B and Amoroso et al., 2012). We questioned whether this TRAP1 mutant was able to protect from apoptosis induced by agents that enhance ER stress and/or paclitaxel. With this aim in mind, wild type TRAP1 or the Δ1‐59TRAP1 deletion mutant were transfected in shTRAP1 MCF7 cells before treatment with ER stress inducers, i.e. tunicamicin, bortezomib, Mg132, thapsigargin and paclitaxel. Interestingly, we observed significant cytoprotective activity toward these agents due to re‐expression of both wild type TRAP1 and the mitochondrial import‐defective TRAP1 mutant (Figure 5C–D).

Figure 5.

The ER stress‐protecting role of TRAP1 is relevant for its antiapoptotic activity. A. Total cell lysates and mitochondrial fractions from shTRAP1 MCF7 cells transfected with pMock, wild type TRAP1 cDNA or the Δ1‐59TRAP1 mutant were separated by SDS–PAGE and immunoblotted with anti‐TRAP1 and anti‐Myc antibodies. B. BiP/Grp78 mRNA levels in shTRAP1 MCF7 cells transfected with pMock, wild type TRAP1 cDNA or the Δ1‐59TRAP1 mutant treated for 24 h with 10 μM paclitaxel (Paclit) and 1 μM thapsigargin (Tg). p‐values versus shTRAP1 cells transfected with pMock: *p < 0.0001. C and D. Apoptotic levels in shTRAP1 MCF7 cells transfected with pMock, TRAP1 cDNA or the Δ1‐59TRAP1 mutant upon treatment with 1 mM tunicamicin (Tun), 100 nM Bortezomib (Bort) or 100 nM Mg132 (C) and with 10 μM paclitaxel (Paclit) or 1 μM thapsigargin (Tg) (D). p‐values versus shTRAP1 cells transfected with pMock each treated with the same agent: *p < 0.0001; **p = 0.001; ***p = 0.002; °p = 0.012.

3.4. Quality control of 18 kDa Sorcin by ER‐associated TRAP1 is involved in apoptosis prevention

Since TRAP1 is responsible for the quality control of specific mitochondria‐destined proteins, including the 18 kDa mitochondrial isoform of Sorcin (Landriscina et al., 2010b; Amoroso et al., 2012), a Ca²⁺‐binding protein involved in chemoresistance (Zheng et al., 2012) and contributing to TRAP1 cytoprotective functions (Landriscina et al., 2010b), we theorized the possibility that ER‐associated TRAP1 could control the mitochondrial apoptotic pathway. Indeed, the simultaneous treatment of shTRAP1 MCF7 with paclitaxel and cyclosporine A, a inhibitor of MTP opening (Boengler et al., 2010), completely abolished paclitaxel‐induced apoptosis, (Figure 6A); remarkably, this process is analogously prevented by transfection in TRAP1 KD cells of the Δ1‐59TRAP1 deletion mutant (Figure 6A). Interestingly, several observations support the hypothesis that ER‐associated TRAP1 regulates mitochondrial apoptosis by exerting quality control regulation on mitochondrial Sorcin: i) shTRAP1 MCF7 cells are characterized by reduced levels of 18 kDa Sorcin (Figure 6B), but not cyclophillin D (data not shown), ii) the transfection of either TRAP1 wt or the Δ1‐59TRAP1 deletion mutant restored Sorcin expression in shTRAP1 cells (Figure 6B), and iii) 18 kDa Sorcin re‐expression protected shTRAP1 MCF7 cells against paclitaxel‐induced apoptosis (Figure 6A). Consistently with the lack of the mitochondrial targeting sequence in the TRAP1 deletion mutant, the TRAP1 band observed in mitochondria of MCF7 cells transfected with the Δ1‐59TRAP1 mutant (Figure 6B–C) showed no immunoreactivity with the anti‐Myc antibody (data not shown), as Figure 5A reports. Furthermore, as previous findings suggested stability control by TRAP1 on mitochondrial proteins (Chae et al., 2012), our group demonstrated reciprocal stability regulation between Sorcin and TRAP1 due to their direct interaction in mitochondria in colon carcinoma cells (Landriscina et al., 2010b). In support of this finding, we observed that Sorcin up‐regulation moderately increased TRAP1 protein levels (Figure 6B). Additionally, the fact that proteasome inhibition by Mg132 rescued Sorcin and TRAP1 levels in shTRAP1 cells (Figure 6C) reinforces the hypothesis that regulation of protein degradation/stability is one of the mechanisms responsible for Sorcin antiapoptotic function.

Figure 6.

The TRAP1‐dependent quality control on 18 kDa Sorcin is relevant for its regulation of mitochondrial apoptotic pathway. A. Apoptotic levels in shTRAP1 MCF7 cells transfected with pMock and treated with 10 μM paclitaxel (Paclit) in the presence and the absence of 1 μM cyclosporine A (CsA), and in shTRAP1 MCF7 cells transfected the Δ1‐59TRAP1 mutant or 18 kDa Sorcin cDNA and treated with 10 μM paclitaxel. p‐values versus shTRAP1 cells exposed to paclitaxel (*p < 0.0001) or versus shTRAP1 cells transfected with pMock each treated with paclitaxel (**p < 0.0001; ***p = 0.001). B. Mitochondrial fractions from scramble and shTRAP1 MCF7 cells transfected with pMock, TRAP1 cDNA, the Δ1‐59TRAP1 mutant or 18 kDa Sorcin cDNA were separated by SDS–PAGE and immunoblotted with anti‐TRAP1, anti‐Sorcin, and anti‐Cox 4 antibodies. C. Mitochondrial fractions from scramble and shTRAP1 MCF7 cells treated with 250 nM Mg132 upon transfection with pMock, TRAP1 cDNA, and the Δ1‐59TRAP1 mutant were separated by SDS–PAGE and immunoblotted with anti‐TRAP1, anti‐Sorcin, and anti‐Cox 4 antibodies. D. Total cell lysates from scramble and shTRAP1 MCF7 cells, treated with 1 μM or 10 μM paclitaxel, were separated by SDS–PAGE and immunoblotted with anti‐TRAP1, anti‐phosphoPERK, anti‐PERK, anti‐phospho eIF2α, anti‐eIF2α and anti‐βactin antibodies.

3.5. TRAP1 modulates the activity of the PERK/eIF2α in response to pacltaxel

An important stress sensor of UPR is PKR‐like ER‐associated kinase (PERK), which mediates phosphorylation of the α subunit of the eukaryotic translation initiation factor eIF2 (Harding et al., 1999), leading in turn to global inhibition of protein synthesis and parallel ATF4‐dependent preferential translation of genes involved in cell metabolism and cytoprotective functions (Harding et al., 1999). Accordingly, we questioned whether TRAP1 modulates the activity of PERK/eIF2α pathway in response to paclitaxel‐induced ER stress. As reported in Figure 6D, while PERK and eIF2α are efficiently phosphorylated in scramble MCF7 cells upon 10 μM paclitaxel exposure, down‐regulation of TRAP1 resulted in reduced activation of the PERK/eIF2α pathway, suggesting that ER‐associated TRAP1 modulates the activation of stress pathways responsible for regulation of protein synthesis rate and survival responses.

3.6. Pharmacological inhibition of TRAP1 partially reverts resistance to paclitaxel

Since our results suggest that TRAP1 protects from mitochondrial apoptosis through adaptation to ER stress, we questioned whether the pharmacological inhibition of TRAP1 may be clinically relevant in reverting paclitaxel resistance, as previously observed by siRNA strategies (Figure 3). TRAP1 inhibition was achieved by gamitrinib, a novel mitochondrial‐targeted small‐molecule HSP90 inhibitor (Siegelin et al., 2011; Kang et al., 2009). The combination of sub‐cytotoxic concentrations of gamitrinib with paclitaxel resulted in additive apoptotic effects in BC MCF7 cells (Figure 7A) as well as partially reverting resistance to this agent in paclitaxel‐resistant MCF7 cells (Figure 7B). In this context, while paclitaxel induced TRAP1 and Grp94 expression and caspase 12 activation, gamitrinib per se minimally induced Grp94 expression and the combination of gamitrinib with paclitaxel enhanced neither UPR response nor TRAP1 expression (Figure 7C). These results suggest that strategies combining TRAP1 inhibition with paclitaxel may achieve clinically‐relevant activities.

Figure 7.

The cytotoxic activity of paclitaxel is enhanced by TRAP1 inhibition. A and B. Apoptotic levels in MCF7 cells (A) and paclitaxel‐resistant MCF7 cells (B) exposed to 10 μM paclitaxel, 10 μM gamitrinibs or the combination of the 2 agents for 24 h. p‐values versus vehicle‐treated cells (*p = 0.006; **p < 0.0001) or versus paclitaxel‐treated cells (***p < 0.0001). C. Total cell lysates from MCF7 cells exposed to 10 μM paclitaxel, 10 μM gamitrinibs or the combination of the 2 agents for 12 h were separated by SDS–PAGE and immunoblotted with anti‐TRAP1, anti‐Grp94, anti‐caspase 12 and anti‐GAPDH antibodies.

4. Discussion

The ER plays an essential role in the regulation of protein folding, protein synthesis, cellular responses to stress, although recent studies suggest that it participates in apoptosis through various mechanisms (Gorman et al., 2012). Crosstalk between ER and mitochondria has indeed been widely demonstrated, as it is involved in the regulation of cell death and drug resistance in human tumors (Rodriguez et al., 2011). Our group recently demonstrated that TRAP1, originally described as a mitochondrial protein, is also localized at the interface between ER and mitochondria where it is involved in crosstalk between these organelles, exerting a role of protection from ER stress, a regulatory function on protein ubiquitination and a quality control on specific mitochondrial client proteins (Amoroso et al., 2012). No evidence exists at present as to whether this novel TRAP1 function in the ER is connected with its well‐established antiapoptotic role. In actual fact, most TRAP1 protein is localized in mitochondria, where it is responsible, along with HSP90, for folding regulation of cyclophillin D and MTP opening (Kang et al., 2007). Based on this rationale, we queried whether the ER stress protecting activity of TRAP1 is relevant for its antiapoptotic activity and role in favoring drug resistance. We addressed this issue in BC cell models exposed to paclitaxel, a microtubule stabilizing agent (Murray et al., 2012) widely used in BC therapy and known for inducing ER stress (Liao et al., 2008; Wang et al., 2009; Mhaidat et al., 2009). We initially demonstrated that a significant cohort of human BCs are characterized by the concomitant overexpression of TRAP1 and BiP/Grp78, likely being exposed to mild chronic conditions of ER stress. Incidentally, this is the first demonstration of TRAP1 induction in this tumor type. Furthermore, we confirmed that paclitaxel induces ER stress in BC cells within its cytotoxic concentration range. It is worth noting that our study suggests that i) TRAP1‐protecting activity toward ER stress is crucial for its antiapoptotic function and role in inducing resistance to paclitaxel, ii) the antiapoptotic activity of ER‐associated TRAP1 depends mostly on quality control of its client protein 18 kDa Sorcin, involved in Ca²⁺ homeostasis and protection from apoptosis in mitochondria (Landriscina et al., 2010b; Suarez et al., 2012), and iii) adaptation to ER stress and resistance to paclitaxel share common molecular mechanisms, with TRAP1 relevant in inducing resistance to both conditions.

These results are relevant for basic and translational researchers. Indeed, our findings that quality control exerted by ER‐associated TRAP1 on its mitochondrial client proteins and, among others, Sorcin is critical for its survival activity shed some light on the relevance of this novel TRAP1 function in the drug resistance phenotype of human malignancies. Previous studies from Altieri's group suggested that the antiapoptotic function of TRAP1 depends mostly on the folding regulation of cyclophillin D within mitochondria (Kang et al., 2007; Chae et al., 2012) and that TRAP1 inhibition directly in mitochondria by the anti‐HSP90 family agent, gamitrinib, results in enhanced apoptosis and secondarily ER stress response (Siegelin et al., 2011). Recently, this TRAP1 folding regulation on cyclophillin D has been shown to be responsible for tumor bioenergetics regulation (Chae et al., 2012). Indeed, HSP90s compartmentalized in mitochondria control glycolysis and oxidative phoshorylation and interference in chaperone regulation of mitochondrial protein folding results in decreased ATP production, along with activation of an integrated signaling pathway with phosphorylation of AMPK (Mihaylova and Shaw, 2011), inhibition of mTORC1 (Zoncu et al., 2011) and a secondarily cytoprotective ER UPR (Chae et al., 2012; Siegelin et al., 2011). Our results reveal a mirror/parallel, non‐redundant, complementary mechanism used by TRAP1 to regulate the mitochondrial apoptotic pathway, based on quality control of specific mitochondrial proteins in the ER and through the regulation of ER stress adaptation. Indeed, ER‐associated TRAP1 is responsible for ubiquitination/quality control on specific proteins along with parallel activation of a cytoprotective UPR response (this study and Amoroso et al., 2012). Specifically, a number of findings suggest the relevance of this mechanism for TRAP1 control of the mitochondrial apoptotic pathway and drug resistance: i) TRAP1 modulates PERK/eIF2α pathway activity, which is known for attenuating protein synthesis and favoring a parallel preferential translation of genes involved in ER stress protection (Harding et al., 1999), ii) specific TRAP1 client proteins, including mitochondrial Sorcin, are down‐regulated in a TRAP1‐low background (shTRAP1 cells) (Amoroso et al., 2012 and Figure 6), iii) Sorcin protein levels are restored upon re‐expression of ER‐associated TRAP1 deletion mutant in shTRAP1 cells, iv) up‐regulation of mitochondrial Sorcin rescues the pro‐apoptotic phenotype of shTRAP1 cells, and v) mitochondrial Sorcin is critical for TRAP1 cytoprotective function in cancer cells (Landriscina et al., 2010b). Viewed in this context, previous studies from our group demonstrated that human colorectal carcinomas are characterized by the co‐expression of TRAP1 and its client proteins, including 18 kDa Sorcin (Amoroso et al., 2012), while recent findings suggest that 18 kDa Sorcin in responsible for cytoprotective functions in cardiac myocytes by regulating Ca²⁺ homeostasis in mitochondria (Suarez et al., 2012). Furthermore, in addition to this mechanism, our results support the hypothesis that TRAP1 and Sorcin are responsible for mutual stability/folding regulation within mitochondria, consistent with previous observations by our group showing direct interaction between the two proteins in mitochondria and their reciprocal requirement for Sorcin mitochondrial localization and TRAP1 stability (Landriscina et al., 2010b). It is accordingly reasonable to hypothesize the responsibility of TRAP1 for dual control on mitochondrial apoptotic pathway: a folding/stability regulation on cyclophillin D and, likely, other client proteins directly within mitochondria (Kang et al., 2007; Chae et al., 2012; Landriscina et al., 2010b) as well as quality control regulation on specific mitochondrial client proteins in the ER, most of which are extremely important regulators of mitochondrial apoptosis (Landriscina et al., 2010b; Amoroso et al., 2012; Suarez et al., 2012). In this perspective, TRAP1 is emerging as a key regulator of a bidirectional crosstalk between ER and mitochondria, as it is present in both cellular compartments, playing a critical role in coupling organelle proteostasis, adaptation to stress and cell survival.

These results are of the highest relevance in the context of studying TRAP1 pathway as a molecular target and designing novel agents to inhibit TRAP1 chaperoning activity. Indeed, while up‐regulation of TRAP1 and its client proteins observed in several human malignancy is considered a general mechanism responsible for resistance to apoptosis and protection from the cytotoxic activity of antiblastic agents (Costantino et al., 2009; Landriscina et al., 2010b), this study suggests its relevance in inducing resistance to paclitaxel by favoring the adaptation of BC cells to ER stress conditions. Indeed, paclitaxel is currently used in the treatment of various malignancies, and, among others, BC in either the adjuvant or the advanced settings (Murray et al., 2012). While paclitaxel activity depends on its ability to stabilize microtubule dynamics, thereby inducing apoptosis (Murray et al., 2012), several studies suggest a role for adaptation to ER stress in the emerging of clinical resistance to taxanes (Liao et al., 2008; Wang et al., 2009; Mhaidat et al., 2009). Furthermore, human tumors are characterized by elevated rates of protein synthesis in order to fulfill the high metabolic demand of proliferating cells, which represents a chronic condition of ER stress (Liu and Ye, 2011), as demonstrated by the activation of the UPR in BCs (this study and Lee et al., 2006) and other human malignancies (Zhang and Zhang, 2010). While prolonged exposure to ER stress results in the activation of the apoptotic cascade (Ozcan and Tabas, 2012), low/mild chronic ER stress conditions are likely to be beneficial for tumors (Ma and Hendershot, 2004). Indeed, high BiP/Grp78 levels are protective against apoptotic stimuli in cancer cells (Chae et al., 2012), favor resistance to anti‐estrogen therapy and chemotherapeutics in BCs (Lee et al., 2006; Scriven et al., 2009; Cook et al., 2012) and correlate with shorter overall survival in human prostate (Tan et al., 2011) and lung (Chae et al., 2012) tumors. In this context, our results suggest a mechanistic link between adaptation to ER stress, protection from mitochondrial apoptosis by TRAP1 and resistance to paclitaxel, thus providing preclinical evidence for considering TRAP1 as a target to prevent adaptation to ER stress and resistance to taxanes. Furthermore, the cross‐resistance of BC cells adapted to ER stress agents or paclitaxel highlights the presence of common mechanisms of resistance between the two conditions. Finally the evidence that down‐regulation of TRAP1 by siRNA or sub‐cytotoxic concentrations of gamitrinib, a mitochondrial HSP90 family inhibitor (Kang et al., 2009), enhances the cytotoxic activity of placlitaxel and restores paclitaxel sensitivity in drug‐resistant BC cells is a strong rationale for considering TRAP1 pathway as a molecular target for reverting drug resistance in clinical setting. Seen in this light, the evidence that ER‐associated TRAP1 plays a crucial role in regulating mitochondrial apoptosis needs to be taken into account in the design of novel TRAP1 inhibitors. Indeed, it is likely that mitochondria‐directed agents may inhibit only the cyclophillin D direct folding regulation (Kang et al., 2007), as they are unable to block the quality control pathway in the ER, as demonstrated by the lower re‐sensitizing activity of gamitrinib, compared to siRNA strategies, in paclitaxel‐resistant cells. Thus, further studies are needed to address the issue of the best pharmacological strategy to inhibit this emerging survival pathway in human cancers.

5. Conclusions

The results herein presented support the hypothesis that ER‐associated TRAP1 is responsible for an extramitochondria control of apoptosis and, therefore, an interference of ER stress adaptation through TRAP1 inhibition outside of mitochondria may be considered a further compartment‐specific molecular approach to rescue drug‐resistance.

Conflict of interest

The authors declare that they have no conflict of interests to disclose.

Supporting information

The following are the supplementary data related to this article:

Supplementary data

Supplementary Figure 1 TRAP1 protection from paclitaxel and ER stress in breast carcinoma MDA‐MB231 cells. A. Total cell lysates from MDA‐MB231 cells exposed to 1 μM paclitaxel (paclit) for 48 h were separated by SDS–PAGE and immunoblotted with anti‐TRAP1 and anti‐GAPDH antibodies. B. TRAP1 mRNA levels in MDA‐MB231 and MCF7 cells exposed to 0.5–1 (MDA‐MB231) and 1–10 μM paclitaxel for 24 h. C. Apoptotic levels in MDA‐MB231 cells transfected with control (Neg siRNA) or TRAP1 siRNA and treated with increasing concentrations of paclitaxel for 24 h. p‐values versus scramble cells: *p = 0.006; **p < 0.0001. Insert: TRAP1 levels in MDA‐MB231 cells upon transfection with negative (1) and TRAP1 siRNA (2). D. BiP/Grp78 mRNA levels in MDA‐MB231 cells transfected with control (Neg siRNA) or TRAP1 siRNA and treated with increasing concentrations of paclitaxel or 1 μM thapsigargin (Tg) for 24 h. p‐values versus scramble cells: *p = 0.02; **p = 0.005; ***p = 0.002.

Supplementary data 1.

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.molonc.2013.04.009.

Maddalena Francesca, Sisinni Lorenza, Lettini Giacomo, Condelli Valentina, Matassa Danilo Swann, Piscazzi Annamaria, Amoroso Maria Rosaria, La Torre Giuseppe, Esposito Franca, Landriscina Matteo, (2013), Resistance to paclitxel in breast carcinoma cells requires a quality control of mitochondrial antiapoptotic proteins by TRAP1, Molecular Oncology, 7, doi: 10.1016/j.molonc.2013.04.009.