Protein Aggregation Patterns Inform about Breast Cancer Response to Antiestrogens and Reveal the RNA Ligase RTCB as Mediator of Acquired Tamoxifen Resistance

Simple Summary Acquired resistance to antiestrogenic therapy remains the major obstacle to curing luminal subtype breast cancer. While current treatment in acquired endocrine-resistant settings includes combined therapy with receptor tyrosine kinase or cyclin-dependent kinase inhibitors, progression to incurable disease remains possible. In recent years, the antioxidant system and the protein quality control network have been associated with the enhanced resistance of breast cancer cells to hormonal therapy. In this work, we raise the hypothesis that antiestrogen treatment induces the accumulation of protein aggregates in sensitive cells, which in turn could hinder the activation of survival pathways. We present evidence concerning a novel way to identify antiestrogen response and disclose a novel protein, RTBC, that controls acquired antiestrogen resistance. This work opens a new avenue for research towards finding breast cancer prognostic markers and therapeutic targets, where the identification of proteins prone to aggregate could help to identify antiestrogen response and understand mechanisms of disease. Abstract The protein quality control network, including autophagy, the proteasome and the unfolded protein response (UPR), is triggered by stress and is overactive in acquired antiestrogen therapy resistance. We show for the first time that the aggresome load correlates with apoptosis and is increased in antiestrogen-sensitive cells compared to endocrine-resistant variants. LC-MS/MS analysis of the aggregated proteins obtained after 4OH-tamoxifen and Fulvestrant treatment identified proteins with essential function in protein quality control in antiestrogen-sensitive cells, but not in resistant variants. These include the UPR modulators RTCB and PDIA6, as well as many proteasome proteins such as PSMC2 and PSMD11. RTCB is a tRNA and XBP1 ligase and its aggregation induced by antiestrogens correlated with impaired XBP1s expression in sensitive cells. Knock down of RTCB was sufficient to restore sensitivity to tamoxifen in endocrine-resistant cells and increased the formation of aggresomes, leading to apoptotic cell death. Analysis of primary human breast cancer samples and their metastases appearing after endocrine treatment showed that RTCB is only localized to aggresomes in the primary tumors, while total aggresomes, including aggregated RTCB, were significantly reduced in the metastases. Therefore, different protein aggregation patterns may indicate loss of function of essential proteins resulting in enhanced protein aggregation that can be used to identify antiestrogen-resistant breast cancer cells and improve the response to antiestrogenic therapy.

The UPR is initiated in response to the accumulation of misfolded proteins in the lumen of the endoplasmic reticulum (EnR). Upon BIP releasing itself from the EnR-resident proteins (IRE1α, PERK and ATF6) to bind the misfolded proteins, the EnR-resident proteins initiate the three UPR cascades to restore proteostasis by inhibiting translation and increasing chaperoning and EnR capacity [15]. In breast cancer cells, activation of IRE1α branch by estrogens leads to an anticipatory response prior to cell division, and is also enhanced in antiestrogen resistance [16]. Initiation of the IRE1α branch consists of an unconventional cytoplasmic splicing [17], where the IRE1α endoribonuclease domain removes an intron from the XBP1 mRNA followed by exon ligation by RNA-splicing ligase RtcB homolog (RTCB). The resulting XBP1 mRNA is translated into a functional transcription factor [18] that stimulates the expression of EnR chaperones such as BiP, lipid synthesis and EnR-associated protein degradation (ERAD) [15]. XBP1s is also an estrogen-responsive gene [19]; it can function as an ERα transcriptional co-activator, and upregulation of XBP1s is strongly associated with antiestrogen resistance [4,20]. Inhibition of IRE1α activity or autophagy can reduce endocrine resistance in breast cancer [6,21,22], supporting the notion that unresolved protein aggregation could enhance drug toxicity.
Toxic protein aggregates have been extensively linked to neurodegenerative diseases [23,24], but their significance has not been followed-up in cancer, with few exceptions, such as mutant p53, which undergoes prion-like aggregation to induce loss of wild-type tumor suppression function [25]. This supports the idea that protein aggregation induced by anti-cancer treatment can engage essential survival proteins, leading to their loss of function. Recently, HDAC6 inhibitors were shown to enhance sensitivity to radiation of breast cancer cell lines by increasing protein aggregation [26]. In this work, we aimed to characterize the aggregated proteome and identify essential proteins needed for endocrine-resistance. We show for the first time, that protein aggregation patterns differ between antiestrogensensitive and -resistant breast cancer cells, with aggregates containing essential proteasome and UPR proteins found exclusively in endocrine-sensitive cells. Specifically, by inhibiting RTCB function, we were able to re-sensitize endocrine-resistant cells to tamoxifen treatment. Therefore, this work opens new avenues for research with the aim of studying the aggregated proteome as a novel approach to improve the response to antiestrogen therapy.

Cell Culture
MDA-MB-231 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS and 5% PEST (Thermo Fisher Scientific, Bremen, Germany). MCF-7 and T-47D were grown in RPMI medium supplemented with 10% FBS and antibiotics. To obtain tamoxifen resistant cells (MCF-7R and T-47DR), endocrine responsive cells were grown in medium containing 500 nM 4-hydroxytamoxifen (TAM; Sigma-Aldrich, St. Louis, MO, USA) for 8 weeks, after which ER expression and cell number in response to TAM and ICI 182 780 (ICI; Sigma-Aldrich, St. Louis, MO, USA) were confirmed by cell counting and Western blot and showed to be either unaltered or increased (Supplementary Figure S1). This phenotype was tested every time the cells were thawed. To rule out clonal effects, MCF-7 and MCF-7R cell lines obtained from Dr. Julia Gee at Cardiff University were also used. MCF-7R and T-47DR cells were routinely grown in RPMI medium supplemented with 10% FBS, 100 nM TAM and antibiotics. All cell lines were grown at 37 • C in a humidified 5% CO2/95% air atmosphere. Twenty-four hours before treatment, the growth medium was replaced by phenol red-free RPMI 1640 or DMEM with 5% charcoal-treated FBS (DCC), 1% PEST and supplemented with 4 mM glutamine. ER ligands were dissolved in 100% ethanol (ETOH) and used at a final test concentration of 10 nM E2, 250 nM ICI, or 500 nM TAM. Autophagy inhibitors Bafilomycin (lysosomal inhibitor) or Spautin-1 (USP10 and USP13 inhibitor) were used at a final concentration of 1µM alone or in combination with 500 nM tamoxifen (TAM) for 8 h. MDA-MB-231, T-47D and MCF-7 cell lines were purchased from ATCC and used straight away, all cell lines were routinely checked for mycoplasma infection using PCR.

Cell Proliferation and Transfection Assays
The cells were plated in 24-well culture plates (5 × 10 4 cells/mL) in phenol red-free medium with 5% DCC. On day 1 after plating, cells were treated with the different study conditions in triplicate. On day 2 and day 3, medium was renewed, and on day 5 each well was treated with trypsin and cells were counted using a Neubauer chamber. For small interfering RNA (siRNA)-mediated knockdown of RTCB, cells were transfected with 1 nM of either the targeting siRNA (RTCBi) or a control (RTCBc) sequence (Sigma-Aldrich) using HiPerFect (Qiagen, Hilden, Germany) or Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA) for 24 h. For cell proliferation assays, cells were treated with TAM [500 nM] or equivalent volume of solvent (ETOH) for 72 h. For Western-blot hybridization or CASP-3 activity experiments cells were treated with 4µ8C (25 µM), Thapsigargin (Acros Organic, Geel, Belgium) (0.5 µM), TAM (500 nM) or equivalent volume of solvent (ETOH) for 3 h and 24 h, respectively.

Aggresome Detection and Co-Localization Studies
ProteoStat ® Aggresome Detection Kit Assay (Enzo Life Sciences, Farmingdale, NY, USA) was used for detection of protein aggregates in cells and tissues. ProteoStat ® is a 488 nm excitable fluorescent molecular probe which is like Thioflavin-T, a rotor moleculedye that is essentially nonfluorescent until it binds to structural features associated with aggregated proteins [27]. All procedures using the ProteoStat ® Aggresome Detection Kit were conducted according to the manufacturer's instructions using a positive control for protein aggregation that consisted of the same cells treated with the proteasome inhibitor MG132 run in parallel. Briefly, after cell fixation and permeabilization, unspecific dye binding was blocked with 10% FBS, 0.5% Triton X-100 in PBS for 1 h. Quantification of # of cells with protein aggregates (PA)/total # of cells was carried out in 3 independent experiments in which 6 to 8 random field images at 200× magnification were taken of each treatment (at least 10,000 cells per group were counted). Fluorescence microscopy was carried out in a Zeiss Axio Imager Z1 microscope (Zeiss, Oberkochen, Germany) equipped with a CCD monochromatic digital camera (Axiocam HRm) or in an Olympus IX-81 microscope (Olympus, Tokyo, Japan) equipped with a CCD monochromatic digital camera (F-view II). ImageJ digital open source software was used to calculate the numbers of positive stained cells and to measure the parameter correlated fluorescence [CF = fluorescence intensity-(area x mean fluorescence intensity of background]. For co-localization studies, ProteoStat ® [1:1000] was incubated for 30 min at room temperature (RT) with CellEventTMCaspase-3/7 Green (Invitrogen, Carlsbad, CA, USA), according to manufacturer's instructions, or with the following antibodies: SQSTM1 (sc-28359, Santa Cruz Biotechnology) 1:500, LC3 α/β (sc-398822, Santa Cruz Biotechnology) 1:500, β-amyloid 1:400 or RTCB 1:200, incubated overnight at 4 • C. Fluorescence confocal microscopy was carried out in a Zeiss LSM-510 META confocal microscope (Zeiss, Oberkochen, Germany) or in a Zeiss LSM 880 confocal microscope with Airyscan.

Tryptic Digestion, Mass Spectrometry Analysis and Protein Identification
Tryptic digestion was performed according to [28], with a few modifications. Protein bands were manually excised from the gel and transferred to 1.5 mL plastic tubes. The gel pieces were washed one time with 25 mM ammonium bicarbonate, three times with 25 mM ammonium bicarbonate/50% acetonitrile (ACN, VWR Chemicals, Radnor, PA, USA) and one time with ACN. The protein's cysteine residues were reduced with 10 mM DTT (45 min at 56 • C) and alkylated with 55 mM iodo-acetamide (30 min at RT). The gel pieces were washed with 25 mM ammonium bicarbonate, followed by 25 mM ammonium bicarbonate/50% ACN and finally with ACN. Gel pieces were dried in a SpeedVac (Savant, Thermo Fisher Scientific, Bremen, Germany) and rehydrated in digestion buffer containing 12.5 µg/mL sequencing grade modified trypsin (AbSCIEX) in 50 mM ammonium bicarbonate. After 30 min at 37 • C, the supernatant was removed and discarded, 50 µL of 50 mM ammonium bicarbonate were added and the samples were incubated overnight (for 16 h) at 37 • C. Extraction of tryptic peptides was performed by the addition of 5% formic acid (FA, Fluka, Mexico City, MX, USA) one time and 5% FA/50% CAN, twice. Tryptic peptides were lyophilized in a SpeedVac (Savant, Thermo Fisher Scientific, Bremen, Germany) and resuspended in 1% FA solution. The samples were analyzed with a Q Exactive Hybrid Quadrupole-Orbitrap (Thermo Fisher Scientific, Bremen, Germany) through the EASYspray nano ESI source (Thermo Fisher Scientific, Bremen, Germany) that was coupled to an Ultimate 3000 (Dionex, Sunnyvale, CA, USA) nano HPLC (high-pressure liquid chromatography) system. The trap (5 mm × 300 µm I.d.) and the EASY-spray analytical (150 mm × 75 µm) columns used were C18 Pepmap100 (Dionex, LC Packings) with a particle size of 3 µm. Peptides were trapped at 30 µL/min in 96% solvent A (0.1% FA). Elution was achieved with the solvent B (0.1% formic acid/80% acetonitrile v/v) at 300 nL/min. The 92 min gradient used was as follows: 0-3 min, 4% solvent B; 3-70 min, 4-25% solvent B; 70-90 min, 25-40% solvent B; 90-92 min, 40-90% solvent B; 92-100 min, 90% solvent B; 100-101 min, 90-4% B, 101-120 min, 4% solvent B. The mass spectrometer was operated at 2.2 kV in the data dependent acquisition mode. A MS2 method was used with a FT survey scan from 400 to 1600 m/z (resolution 70,000; AGC target 1E6). The 10 most intense peaks were subjected to HCD fragmentation (resolution 17,500; AGC target 5E4, NCE 28%, max. injection time 100 ms, dynamic exclusion 35 s). Spectra were processed and analyzed using Proteome Discoverer (version 2.2, Thermo Fisher Scientific, Bremen, Germany), with the MS Amanda (version 2.0, University of Applied Sciences Upper Austria, Research Institute of Molecular Pathology, Vienna, Austria) and Sequest HT search engines. Uniprot (TrEMBL and Swiss-Prot) protein sequence database (version of May 2017) was used for all searches under Homo Sapiens. Database search parameters were as follows: carbamidomethylation of cysteine, oxidation of methionine, and the allowance for up to two missed tryptic cleavages. The peptide mass tolerance was 10 ppm, and fragment ion mass tolerance was 0.02 Da. To achieve a 1% false discovery rate, the Percolator (version 2.0, Thermo Fisher Scientific, Bremen, Germany) node was implemented for a decoy database search strategy, peptides were filtered for high confidence and a minimum length of 6 amino acids and proteins were filtered for a minimum number of 1 peptide sequence. Only proteins with mean values above 25% vs. control (ETOH) condition were considered as upregulated in the insoluble fraction.

Statistical Analysis and Bioinformatics
Statistical analysis was performed using GraphPad Prism software version 6.0. Data from at least two independent experiments were used to calculate the mean ± standard deviation (SD). Mann-Whitney tests were used for the comparison of two independent groups for immunofluorescence, Western blotting, cell proliferation and siRNA assays. Wilcoxon tests were used for analysis of human breast cancer paired tissue samples. Differences were considered significant if p < 0.05. Venny 2.1 [29] was used to identify the upregulated proteins uniquely found in MCF-7 cells treated with TAM or ICI. For this, first up-regulated proteins in each treatment were obtained by comparison to the control (ETOH), second the upregulated proteins by E2 were deleted from the list, and finally up-regulated proteins by the antiestrogens in MCF-7 were compared to those found in MCF-7R cells.

Patient Sample Collection and Characterization
The clinicopathological data of the patient cohort are presented in Supplementary  Table S1. The samples used in this study are within the project "Exploring epigenetic profiling as prognostic/predictive markers of endocrine resistance in estrogen receptor positive breast cancer" approved by IPO Porto's Ethical Committee (CES IPO: 369/2017).

Aggresome Accumulation Correlates with Reduced UPR and Autophagy Activation in Antiestrogen-Sensitive Breast Cancer Cells
In antiestrogen-resistant (MCF-7R) cells, the XBP-1s/XBP-1u ratio increased already after 3 h incubation with TAM or ICI and increased throughout the 12 h evaluated, while in MCF-7 cells it did not significantly change ( Figure 1A). We ruled out the possibility that MCF-7 cells could be partially compensating EnR stress by activation of the PERK/EIF2A pathway, since neither the ratio p-PERK/PERK nor p-EIF2A/EIF2A increased following TAM or ICI treatment ( Figure 1A). To exclude the possibility that E2 could influence the effect of the antiestrogens alone, we carried out a control experiment co-incubating the cells with TAM + E2 and observed the same effects as with TAM alone (Supplementary Figure S3A). Therefore, the model of antiestogen resistance used herein shows efficient activation of the IRE1α/XBP1 branch in response to antiestrogens, in line with other studies [4,6].
The lack of UPR activation by MCF-7 cells in response to antiestrogens alone or combined with E2 resulted in protein aggregation and a significant increase in the aggresome levels as well as the % of cells containing aggresomes ( Figure 1B,C and Supplementary Figure S3B). Once again, co-incubation of E2 + TAM produced the same effect as TAM alone. This was also observed in T-47D cells, but not in MCF-7R or T-47DR cells. Treatment with E2, and antiestrogenic treatment in cells which do not express ERα (MDA-MB-231 cells) did not produce aggresome accumulation ( Figure 1B,C). This indicates that aggresomes result from antagonizing ERα function in cells that are sensitive to ICI and TAM. The aggresomes were found mostly dispersed in the cytoplasm of sensitive cells, indicating that the protein quality control (PQC) mechanisms were saturated [30], but in MCF-7R and T47-DR cells they were preferably found at the juxtanuclear quality control region that concentrates proteasomes and disaggregating chaperones [31,32], suggesting a functional PQC ( Figure 1B,C and Supplementary Figure S3B,C, arrowheads).
Aggresomes are preferably cleared by autophagy [33]; differences in autophagic patterns between MCF-7 and MCF-7R cells were evident already in basal conditions and were maintained after exposure to antiestrogens, with MCF-7R cells showing clear p62 and LC3-II puncta colocalization with aggresomes (Supplementary Figure S4A). On the other hand, in MCF-7 the number of puncta was lower, with a more diffuse p62 and LC3-II staining and higher number of aggresomes that did not colocalize (Supplementary Figure S4A). These results are compatible with lower autophagic efficiency in antiestrogen sensitive cells as compared to resistant ones. This is in agreement with the literature [7,34] and the lack of increase in XBP-1s [4], as shown in Figure 1A. It has previously been shown that cyto-protective autophagy can contribute to antiestrogen resistance [7,8,22,35,36], and several groups have shown that inhibition of autophagy can re-sensitize ERα+ cells to antiestrogen treatment [7][8][9]22,34,[37][38][39]. Moreover, co-incubation of MCF-7R cells with TAM and the autophagy inhibitors bafilomycin or spautin-1 for 24 h increased the aggresome load and their co-localization with LC-3 or p62 (Supplementary Figure S4B), supporting the notion that enhanced autophagy protects the cells from aggresome accumulation. Aggresome-associated cell death results from toxic gain of function by facilitating aberrant protein-protein interactions [40], or loss of function of essential proteins [41]. Sensitive and resistant cells showed nearly complete aggresomes colocalization with cleaved caspase 3 (cCASP-3) after antiestrogen treatment (not shown); consequently, the number of cells positive for aggresome/cCASP-3 colocalization was significantly higher in antiestrogen-treated MCF-7 and T-47D cells ( Figure 1D).
In summary, resistant cells are more capable of clearing protein aggregates that result from antiestrogen treatment because they maintain XBP1s expression and have more efficient autophagy; while in sensitive cells, the lack of aggresome clearance correlates with antiestrogen-induced cell death.

Protein Aggregation after Antiestrogen Treatment Targets Different Pathways in Sensitive and Resistant Cells
To isolate protein aggregates for subsequent mass spectrometry analysis, we enriched in proteins that were insoluble in ionic detergent. These included intrinsically disordered proteins, like Amyloid A4 proteins and β-amyloids, which, using MG132, accumulate in Proteostat-stained aggresomes (Figure 2A). The detergent-insoluble and total protein fraction were separated by SDS-PAGE and exhibited an increased insoluble/total protein ratio (IF/T) only in antiestrogen-sensitive cells after treatment ( Figure 2B,C). Notably, E2 protected proteins from detergent-induced destabilization, and the ERα-negative MDA-MB-231 cells did not exhibit increased levels of insoluble proteins ( Figure 2B,C). The insoluble fractions of MCF-7 and MCF-7R cells were analyzed by LC-MS/MS to identify the pathways that could be hindered by loss of protein function due to aggregation. A total of 640 and 605 proteins were consistently found in the detergent-insoluble fraction of MCF-7 and MCF-7R cells, respectively. Treatment of MCF-7 cells with TAM or ICI for 24 h resulted in 38 proteins aggregated with TAM or ICI, 54 with TAM treatment and 22 with ICI ( Figure 2D, top panel, and Table 1). For comparison, the unique proteins aggregated in MCF-7R cells after TAM or ICI treatment were also analyzed ( Figure 2D, second from top, and Supplementary Table S2).  Figure 2D, top panel, and Table 1). For comparison, the unique proteins aggregated in MCF-7R cells after TAM or ICI treatment were also analyzed ( Figure 2D, second from top, and Supplementary Table S2).  The aggregated proteins in MCF-7 cells significantly represented Proteasome pathway (pTAM = 0.003 and pICI = 7.03x10 −05 ; Table 2). Additionally, TAM aggregated mRNA surveillance proteins (p = 0.015), while ICI targeted Protein processing in endoplasmic reticulum (p = 0.027). These pathways were not overrepresented by the detergent-insoluble proteins from MCF-7R after either TAM or ICI treatment. Instead, MCF-7R cells possessed detergent-insoluble proteins enriched in Huntington's disease (pICI = 0.002). We also identified proteins with relevance in proteinopathies in MCF-7 cells, including Poly(U)-bindingsplicing factor PUF60 (PUF60), associated with amyotrophic lateral sclerosis (ALS) [42][43][44], Transaldolase-1 (TALDO1), associated with Parkinson's disease [45], and hydroxyacyl-CoA dehydrogenase trifunctional multienzyme complex subunit beta (HADHB), related to Creutzfeldt-Jakob disease [46]. To rule out higher levels of detergent-insoluble proteins being a result of higher expression levels, we compared the 640 proteins found to be insoluble in MCF-7 cells with their values in the total protein extract and found no significant correlation (Spearman rTAM = 0.019 and rICI = 0.13). Additionally, when analyzed individually, the total levels of the proteins found to be uniquely aggregated in MCF-7 cell after antiestrogen treatment were not up-regulated in the total protein fraction ( Figure 3A).

Protein Aggregation after Antiestrogen Treatment Impairs Proteasome and UPR Function
Proteosome pathway proteins that consistently aggregated after TAM and ICI treatment where PSMC5 and PSMC4 in either MCF-7 or MCF-7R cells. PSMA2 and UBA1 were also only detected as aggregated in MCF-7 cells with either of the antiestrogens, with

Protein Aggregation after Antiestrogen Treatment Impairs Proteasome and UPR Function
Proteosome pathway proteins that consistently aggregated after TAM and ICI treatment where PSMC5 and PSMC4 in either MCF-7 or MCF-7R cells. PSMA2 and UBA1 were also only detected as aggregated in MCF-7 cells with either of the antiestrogens, with PSMD11 and PSMC2 mostly aggregated after ICI treatment (Table 1 and Supplementary  Table S2). We confirmed using Western blot that even when aggregated in both sensitive and resistant cells lines, PSMD11 and PSMC2 were aggregated in higher levels in MCF-7 cells ( Figure 3B). Total levels of ubiquitinated proteins increased in MCF-7 cells after TAM or ICI treatment ( Figure 3C), which strongly supports the notion that in sensitive cells, antiestrogens impair the proteasome function by inducing aggregation of several 26S proteasome subunits.
RTCB and PDIA6 were also uniquely aggregated in MCF-7 cells ( Figure 3D). RTCB catalyzes the ligation of XBP-1 mRNA following cleavage by IRE-1α endonuclease [47], while PDIA6 maintains IRE-1α activation within a physiologically acceptable range [48]. Total RTCB and PDIA6 did not change, although the insoluble proteins significantly increased only in MCF-7 and T-47D cells ( Figure 3D). These results reinforce the idea that although RTCB and PDIA6 are expressed at similar levels in both sensitive and resistant cell lines, they are more prone to aggregation in sensitive cells in response to ERα antagonism, which may lead to cytotoxicity due to their loss of function. In support of this idea, RTCB aggregation occurred already after 3h (Supplementary Figure S5A,B) and correlated with lack of XBP1s protein increase only in sensitive cells ( Figure 1A). Moreover, RTCB subcellular localization is related to its dual function as t-RNA ligase (nuclear) [49] and XBP1 ligase (cytoplasm). We confirmed that in sensitive cells treated with antiestrogens, RTCB is initially localized in nuclear speckles but also in aggresomes, and after 24 h, RTCB was mostly found in the aggresomes, while in resistant cells, RTCB did not colocalize with aggresomes (Supplementary Figure S5B). E2 treatment showed a clear RTCB distribution throughout the cytoplasm and nucleus of MCF-7 cells. As a control, we combined E2+TAM and evaluated the pattern of RTCB staining in relation to the aggresomes after 24 h, where it was clear that addition of E2 did not prevent TAMmediated RTCB aggregation (Supplementary Figure S5C). Therefore, the RTCB instability triggered by antiestrogens in MCF-7 cells, impairs XBP1s-mediated proteostasis, promoting a vicious cycle of protein misfolding and aggresome accumulation.

Inhibition of RTCB Expression Downregulates XBP1s and Improves Sensitivity to Tamoxifen
To test whether RTCB function promotes survival to antiestrogens, we pre-treated MCF-7R cells with siRNAs to RTCB (siRTCB) for 24 h, after which they were treated with TAM + siRTCB for an additional 3 days. siRTCB significantly decreased proliferation of TAM-treated cells ( Figure 4A). The reversal of resistance to TAM was as efficient as the inhibition of proliferation observed in MCF-7 cells and was confirmed in T-47D/T-47DR cells. Moreover, similar to the IRE1α inhibitor 4µ8C, RTCB reduced the XBP1s/XBP1u ratio ( Figure 4B,C). In contrast to 4µ8C, siRTCB downregulated BIP protein ( Figure 4B,C), suggesting that blocking RTCB function may translate into longer lasting inhibition of UPR possibly due to the weak stability of the IRE1α-4µ8C complex [50]. Blocking XBP1s production by either siRTCB or 4µ8C in MCF-7R treated with TAM resulted in increased aggresome formation and co-localization with cCASP-3 ( Figure 4D). Interestingly, in siRTCB-treated cells, XBP1s and BIP levels were also reduced with ETOH, but protein aggregation was only observed after challenging the cells with TAM, confirming that RTCB function supports resistance to TAM.  To investigate the significance of RTCB aggregation in vivo, we tested a small cohort consisting of 22 paired samples of untreated primary tumor and the matched metastasis that appeared after completion of endocrine treatment (Supplementary Table S1). A statistically significant decrease in the percentage of cells with protein aggregates was found in metastatic tumors that represent disease recurrence after completion of endocrine treatment ( Figure 5A). Aggregation of RTCB showed higher colocalization with aggresomes in the primary tumor as compared to the metastasis ( Figure 5B). This supports the idea that protein aggregation in breast cancer tissue samples may provide information about tumor sensitivity to endocrine therapy and suggest that antiestrogen resistant cells are better capable of preventing RTCB aggregation. that appeared after completion of endocrine treatment (Supplementary Table S1). A statistically significant decrease in the percentage of cells with protein aggregates was found in metastatic tumors that represent disease recurrence after completion of endocrine treatment ( Figure 5A). Aggregation of RTCB showed higher colocalization with aggresomes in the primary tumor as compared to the metastasis ( Figure 5B). This supports the idea that protein aggregation in breast cancer tissue samples may provide information about tumor sensitivity to endocrine therapy and suggest that antiestrogen resistant cells are better capable of preventing RTCB aggregation.

Discussion
Enhancement of UPR and autophagy have been linked to endocrine therapy resistance [5,22]. Our results go a step further, showing that antiestrogen treatment causes less aggresome accumulation in resistant cells due to maintenance of soluble/functional RTCB, needed to maintain the protective effect of IRE1α/XBP1 pathway. Loss of RTCB function leads to aggresome accumulation, which can be used to predict antiestrogen response.
To date, the association of higher XBP1s with antiestrogen resistance had been related to co-activation of ERα by XBP-1s [51]. Here, we show that RTCB aggregation induced by antiestrogens and accumulation in aggresomes occurs only in sensitive cells and correlated with reduced XBP1s and BIP levels. Therefore, RTCB function could be targeted to revert endocrine resistance by reducing ERα co-activation and protection from BIP chaperoning activity. We tested this hypothesis using RTCB knock-down in antiestrogen-resistant cells to demonstrate that RTCB loss of function reproduced the aggresome pattern of sensitive cells and re-sensitized to tamoxifen. Notably, RTCB knock-down produced a much more stable and powerful inhibition of the stress response mediated by XBP1s than the IRE1α inhibitor 4µ8C, (measured as BiP protein levels). This may be because unfavorable pharmacokinetics limit the utility of 4µ8C and other aldehydes that function as IRE1α inhibitors [50]. Thus, RTCB targeting can potentially inhibit this pathway in a more stable manner.
Aggregated beta-amyloids are commonly found in proteinopathies [40] and have been associated with toxic gain of function through aberrant protein-protein interactions or loss of function [43,52]. Protein aggregates are known to remain insoluble in strong detergent buffers so, biochemical fractionation methods that involve the sequential use of buffers and detergents of increasing stringency and ultracentrifugation to separate the soluble and insoluble fractions have been applied to tissue and cellular homogenates [53][54][55]. The multistep protocol used in this study (Supplementary Figure S2) allows, at first, the separation of cytosolic soluble proteins and in a second phase, NP40 allows the solubilization and separation of membrane proteins. Additionally, as internal control, we were able to identify several proteins previously reported to be insoluble such as cytoskeletal proteins, collagens and matrix proteins [53] or with increased tendency to aggregate like TALDO1 [45], HADHB [46], TAR DNA-binding protein 43 (TDP-43) or PUF60 [42][43][44]. Interestingly, the number of aggregated proteins identified in the detergent-insoluble fraction of MCF-7 and MCF-7R cells was similar, although the identity and levels of specific proteins varied between the two phenotypes and did not correlate with total protein levels. This clearly shows that antiestrogens induce the aggregation of specific proteins. It has recently been shown that protein solubility increases when cells are undergoing cell division due to post-translational modifications regulating enzymatic activity [56]. In future studies it will be interesting to identify the stimuli that triggers this selective aggregation or protection of proteins.
Differential aggresome formation was found in therapy naïve primary luminal breast cancer vs. their metastasis, appearing after endocrine therapy. Only a small cohort was analyzed; however, the results are promising, as the metastases may have developed strategies to prevent RTCB aggregation. Future studies of luminal primary tumors from women that were treated with endocrine treatment, in which some developed early resistance/progression and others did not are needed. It is noteworthy that resistance to antiestrogens has been linked to emergence or selection of cancer stem-like cells [57,58]. There is an increasing body of evidence that points to important roles of UPR and autophagy in cancer stem-like subpopulaitons. Specifically, impairment of autophagy in breast cancer stem cells reduces expression of staminal markers, self-renewal capacity and tumorigenicity [59][60][61]. Autophagy is upregulated in mammospheres where Beclin1 and ATG4 are needed to maintain and expand the breas cancer stem-like populations [59,61,62]. UPR modulation in breast cancer stem-like cells has been less explored, yet it has been shown that BIP is enriched in these subpopulations isolated from multiple breast cancer cell lines [63] and activation of the IRE1/XBP1 branch was related to the maintenance breast cance stem cells, their tumorigenicity and resistance to chemotherapy [64,65]. Studies in other models have shown that damaged proteins are assymetrically partitioned during cell division [66][67][68]. Therefore, in the future, it will be interesting to explore whether protein aggregation load could function as a driving force for cancer stem-like features, through asymmetric partitioning and may explain the lack of protein aggregation in the metastasis arising after endocrine therapy.
In this work, we presented evidence concerning a new approach to identify antiestrogen response and disclosed RTCB as a novel player in controlling acquired antiestrogen resistance ( Figure 6). Cell stress induced by antiestrogen treatment (i.e., induced by ROS or metabolic imbalances) [69,70] results in protein misfolding. Since antiestrogen-sensitive cells have lower UPR and autophagic capacity [7,34], decreased proteasomal activity [71] and expression of chaperones [35,72], and the protein quality control mechanisms become saturated and aggresomes accumulate. Proteins trapped into aggresomes are essential to overcome proteotoxic stress (i.e., RTCB, proteasome proteins and chaperones) and their loss of function leads to a disruptive cycle of proteostasis loss leading to activation of apoptosis mediated by cCASP3 translocation to aggresomes. This is illustrated by the analysis of RTCB, which does not complete XBP1 splicing (dotted lines), and therefore, transcription of XBP1s target genes (i.e., BiP) is not activated. These alterations enhance EnR stress and may also decrease activation of ERα-mediated proliferation by XBP1s [73]. On the other hand, proliferation and survival signaling remains active in resistant cells treated with antiestrogens. Since these cells have increased activation of their proteostasis mechanisms, misfolded proteins are easily cleared and since XBP1s are induced upon antiestrogen treatment, aggresome levels are kept low. In addition, co-activation of ERα by XBP1s may be possible, sustaining proliferation and survival. However, if RTCB is knocked-down in resistant-cells treated with antiestrogens, reduced levels of XBP1s are sufficient to promote aggresome accumulation and apoptosis.  Additionally, antiestrogen therapy increases oxidative and metabolic stress, which hinders correct protein folding to increase their aggregation. The results presented herein show that the PQC network (autophagy, UPR and proteasome) may be rapidly saturated leading to aggresome accumulation. Aggresomes contain essential proteins needed to overcome proteotoxic stress (i.e., RTCB, chaperones and PQC Figure 6. Schematic representation of the findings presented in this work in the context of previously described mechanisms guiding endocrine resistance. (A) In antiestrogen-sensitive cells, treatment with antiestrogens induces ERα proteasomal degradation (ICI) or inhibits the transcription of ERα target genes (TAM); consequently, proliferation and survival signaling is halted. Additionally, antiestrogen therapy increases oxidative and metabolic stress, which hinders correct protein folding to increase their aggregation. The results presented herein show that the PQC network (autophagy, UPR and proteasome) may be rapidly saturated leading to aggresome accumulation. Aggresomes contain essential proteins needed to overcome proteotoxic stress (i.e., RTCB, chaperones and PQC proteins), reducing their function. Aggregated RTCB cannot complete XBP1 splicing, XBP1 levels drop and, consequently, the activation of proteostasis genes under XBP1 regulation, like BiP, becomes compromised. Altogether, these events may generate a positive feedback loop increasing the accumulation of aggresomes which, in turn, potentiates the proteome instability and contributes to impair stress responses. Activation of pro-apoptotic pathways with co-localization of cleaved caspase 3 (cCASP-3) in aggresomes ultimately results in cellular death. (B) Antiestrogen-resistant cells have a higher autophagic and proteasomal degradation capacity, which, coupled with induced expression of XBP1s and transactivation of its target genes (i.e., BiP, chaperones, ERAD response) allows resistant cells to cope with protein misfolding and aggregation, keeping aggresome levels low. In addition, co-activation of ERα by XBP1s is possibly maintained, resulting in increased proliferation and treatment with antiestrogens do not compromise survival signaling. (C) RTCB knock-down in resistant cells treated with antiestrogens, results in impaired XBP1 expression which is sufficient to promote aggresome accumulation. Impaired transcription of proteostasis genes by XBP1s possibly contributes to saturate the PQC network. Intolerable proteome instability and consequent proteotoxic stress may activate pro-apoptotic pathways resulting in cellular death as observed by cCASP3 accumulation. EnR-endoplasmic reticulum; Era-estrogen receptor alpha; PQC-protein quality control proteins.

Conclusions
Taken together, the findings presented herein suggest that protein aggregation patterns could be used to predict resistance to antiestrogen therapy, and they could be indicative of improved capacity to maintain a healthy proteome. This work opens a new avenue for research with respect to finding breast cancer prognostic markers and therapeutic targets, where the identification of proteins prone to aggregate could help identify antiestrogen response and understand mechanisms of disease.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/cancers13133195/s1, Figure S1. Estrogen receptor expression and response to antiestrogenic treatment of MCF-7, MCF-7R, T-47D, T-47DR and MDA-MB-231 cell lines, Figure S2: Schematic representation of the insoluble fraction recovery protocol used in this study, Figure S3: Characterization of antiestrogen-induced protein aggregation clearance and accumulation in breast cancer cell lines, Figure S4: Correlation between autophagy activation and aggresome accumulation following antiestrogen treatment, Figure S5: Changes in RTCB expression levels in the total (T) and Insoluble Fractions (IF) of MCF-7 and MCF-7R cells along a period of 24h following ETOH (control), 10 nM E2, 250 nM ICI or 500 nM TAM treatments, Table S1: Clinicopathological data of the patient cohort, Table S2  Informed Consent Statement: Patient consent was waived due to the fact that only archival tissue was used exclusively in a retrospective way, obviating the need for patient consent according to Portuguese Law Art 19º, n6, Law 12/2005 from January.

Data Availability Statement:
The data presented in this study are available in the article and supplementary material.