A mitochondrial unfolded protein response inhibitor suppresses prostate cancer growth in mice via HSP60

Mitochondrial proteostasis, regulated by the mitochondrial unfolded protein response (UPRmt), is crucial for maintenance of cellular functions and survival. Elevated oxidative and proteotoxic stress in mitochondria must be attenuated by the activation of a ubiquitous UPRmt to promote prostate cancer (PCa) growth. Here we show that the 2 key components of the UPRmt, heat shock protein 60 (HSP60, a mitochondrial chaperonin) and caseinolytic protease P (ClpP, a mitochondrial protease), were required for the development of advanced PCa. HSP60 regulated ClpP expression via c-Myc and physically interacted with ClpP to restore mitochondrial functions that promote cancer cell survival. HSP60 maintained the ATP-producing functions of mitochondria, which activated the β-catenin pathway and led to the upregulation of c-Myc. We identified a UPRmt inhibitor that blocked HSP60’s interaction with ClpP and abrogated survival signaling without altering HSP60’s chaperonin function. Disruption of HSP60-ClpP interaction with the UPRmt inhibitor triggered metabolic stress and impeded PCa-promoting signaling. Treatment with the UPRmt inhibitor or genetic ablation of Hsp60 inhibited PCa growth and progression. Together, our findings demonstrate that the HSP60-ClpP–mediated UPRmt is essential for prostate tumorigenesis and the HSP60-ClpP interaction represents a therapeutic vulnerability in PCa.

. HSP60 regulates ClpP expression in multiple types of cancer cells. Figure S2. Overexpression of ClpP in HSP60 depleted PCa cells. Figure S3. HSP60 interacts with ClpP but not with LONP1. Figure S4. HSP60 interacts with both HSP10 and ClpP, whereas HSP10 and ClpP do not compete for binding with HSP60. Figure S5. Schematic of various constructs of HSP60 (mutants) and their interactions with ClpP. Figure S6. Transcript levels of UPR mt components in prostate cancer. Figure S7. HSP60 or ClpP silencing inhibits clonal growth and reduces c-Myc and EZH2 in PCa cells. Figure S8. DCEM1 disrupts HSP60-ClpP interactions, and induces a very little or no apoptosis in immortalized untransformed prostate epithelial cells. Figure S9. DCEM1 induces mitochondrial ROS (mitoROS) production and caspase activation, and inhibits clonogenicity in PCa cells. Figure S10. DCEM1 inhibits c-Myc and EZH2, and induces mitoROS, polyubiquitination, and apoptosis in TKO cells. Figure S11. N-acetyl cysteine (NAC) antagonizes DCEM1-induced cell death. Figure S12. The components of UPR mt were not affected by DCEM1 treatment with exception in PC-3 cells. Figure S13. DCEM1 and c-Myc inhibitor reduced expression of EZH2 and AR in PCa cells. Figure S14. Mitochondrial Chaperonin Activity Assay (MiCAA). Figure S15. LONP1 but not PARL inhibits DCEM1-induced accumulation of poly-Ub proteins and DEVDase activity. Figure S16. Reduced ATP/ADP ratio upon Hsp60 or ClpP silencing and by DCEM1 treatment in PCa cells. Figure S17. DCEM1 treatment did not cause organ or systemic toxicities. Figure S18. DCEM1 treatment of C57BL/6 mice did not change hematological parameters. Figure S19. DCEM1 treatment of C57BL/6 mice did not change clinical chemistry parameters. Tables   Table S1. List of antibodies used in the study Table S2. List of shRNAs sequence and their sequences Table S3. List of siRNAs and their sources

Whole cell lysates preparation, subcellular fractionation, and Western blotting
Preparation of whole cell lysates (WCL), mitochondrial and cytosolic fractions, and Western blotting were performed as previously described (1)(2)(3)(4). The quantification of protein was carried out by micro BCA protein assay kit using BSA as standard (Thermo Fischer, Grand Island, NY).
In brief, proteins were separated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis using Criterion Precast Gels (Bio-Rad, Hercules, CA) and transferred on to nitrocellulose membrane (Bio-Rad, Hercules, CA). Nonfat milk (5%) was used to block the membrane for 30 minutes and washed with PBS-T (1XPBS and 0.1% Tween 20) and incubated overnight with indicated primary antibodies and corresponding HRP-conjugated secondary antibodies followed by immunodetection using ECL reagent (BioRad, Herculer, CA). Actin or GAPDH, lactate dehydrogenase B (LDHB) and TOM20 were used as loading and quality controls for WCL, cytosolic and mitochondrial extracts, respectively.

TCGA mRNA expression and gene correlations
The publicly available TCGA IlluminaHiSeq datasets (February 24, 2015) were downloaded in their normalized format from the UCSC Xena Browser and the average transcript reads were calculated for both tumor and matched non tumor specimens. The Read per kilo-base per million mapped reads (RPKM) method was used to quantify the mRNA gene expression from RNA sequencing data by normalizing for total read length and the number of sequencing reads.
Spearman's correlation was performed to determine the degree to which two genes are related.

Immunohistochemistry (IHC)
Tissues were fixed in 10% normal buffered formalin for 24 h prior to processing. Tissues were processed and embedded in paraffin and then sectioned at 5 microns (5 µm) thickness. Tissue sections were de-paraffinized using xylene and rehydrated with graded alcohol followed by double distilled water (H2O). Antigen retrieval was performed by boiling with 1X sodium citrate buffer, pH 6.0 (Zymed Laboratories, San Francisco, CA) for 20 minutes. The Dako Autostainer was used to stain slides. Endogenous peroxidase was blocked by immersing slides in 3% H2O2 in methanol for 15 minutes. To block non-specific binding, tissues were incubated with 10% normal serum for 30 minutes, followed by avidin/biotin block (Vector Labs Cat#SP-2001). Primary antibodies for mentioned proteins were diluted in 1% BSA solution and incubated for 30 minutes at room temperature (RT), followed by the biotinylated secondary antibodies for 15 minutes. For signal enhancement, ABC reagent (Vector Labs Cat #PK 6100) was applied for 30 minutes. To reveal endogenous peroxidase activity, slides were incubated with 3, 3'-diaminobenzidine tetrahydrochloride (DAB) substrate (Dako Cat #K3467) for 5 minutes and then counterstained with DAKO Hematoxylin for 20 seconds. Slides were then dehydrated through several baths of graded alcohols and xylenes and then mounted via DAKO Counterstaining solution. All images were taken at three fields (10X, 20X & 40 X magnifications) that showed representative areas of the human prostate tumor tissue microarray (TMA) (3,4).

Immunoprecipitation
LNCaP, PC-3, and DU145 cells (treated with DCEM1 or vehicle treated); and TKO prostatic tumor tissues were lysed, precleared with mouse or rabbit (depending on the primary antibodies used) IgG-conjugated agarose beads, and incubated with primary antibodies against HSP60, ClpP or IgG (as control antibody), followed by addition of anti-rabbit, or anti-mouse IgG beads. Finally, the beads were pelleted, washed thoroughly, boiled in SDS sample buffer, and analyzed by TrueBlot Western blotting system (eBioscience).
In another experiment, various HSP60-V5 WT and mutant expressing constructs and ClpP-FLAG construct were co-transfected in PC-3 cells. After 48 h of transfection, cells were harvested and lysed, precleared with mouse or rabbit (depending on the primary antibodies used) IgG-conjugated agarose beads, and incubated with primary antibodies against V5 tag, FLAG tag or IgG (as control), followed by addition of anti-rabbit, or anti-mouse IgG beads. Finally, the beads were pelleted, washed thoroughly, boiled in SDS sample buffer, and analyzed by TrueBlot Western blotting system (eBioscience) (3).

Cellular thermal shift assay (CETSA)
CETSA was performed as described previously (8,9) . To determine the binding of DCEM1 with   HSP60 within the cells, 2 million PC-3 cells were treated with either DMSO or DCEM1 (20 µM) for 1 hr. Protease inhibitor cocktails (PIC) were directly added to media and cells were harvested by scrapping and washed once with PBS, then resuspended in 500 µl PBS supplemented with 1X PIC and DCEM1 (20 µM). 100 µl cell suspension was distributed into three PCR tubes and incubated at 55 0 C, 57 0 C and 58 0 C for 3 minutes in a PCR machine. Immediately after heating, cell suspensions were further incubated at room temperature for 3 minutes and then snap frozen in liquid nitrogen. Cells were lysed by three freeze-thaw cycles in liquid nitrogen was performed.
Cell lysates were collected and centrifuged at 20,000 g for 20 min at 4 0 C to remove cell debris and aggregated proteins. Supernatant was collected and 10 µl was heated with SDS sample buffer at 95 0 C for 7 minutes and subjected to Western blot analysis for detection of HSP60 and actin.

Biotin-DCEM1 pull down assay
To confirm the direct binding of DCEM1 to endogenous HSP60 protein in PCa cell lysate, biotin conjugated DCEM1 was synthesized (Enamine Ltd). Biotin-DCEM1 pull down of HSP60 was performed according to recent published protocol (8). Whole cell lysates from 60-70 % confluent PC-3 cells were prepared in lysis buffer (25 mM HEPES, ph 7.4, 150 mM NaCl, 1% NP-40, 1 mM EDTA). Protein concentrations were measured using micro BCA protein estimation kit and 100 µg whole cell lysate (total volume 200 µl) was pre-cleared with streptavidin magnetic bead conjugate (Cell Signaling, Cat # 5947) for 1 h at 4 0 C. After pre-clear step, lysates were incubated with 10 and 20 µM Biotin-DCEM1 or 20 µM D-Biotin or DMSO on a tube rotator overnight at 4 0 C. Streptavidin magnetic beads (20 µl) were added to each sample and further incubated on a tube rotator for 2 h at 4 0 C. Streptavidin magnetic beads were extensively washed 3 times with lysis buffer containing 0.1% BSA followed by 3 washings with lysis buffer without BSA. Finally, beads were incubated at 95 0 C in 2x SDS sample buffer for 7 minutes. Samples were centrifuged and supernatant was subjected to Western blot analysis for HSP60 and ClpP proteins.

Dot Blot Far Western analysis
To detect direct interaction between HSP60 and ClpP, we performed Dot Blot Far Western analysis as previously described (10). Recombinant HSP60 (Abcam, ab113192) and ClpP (Origene, TP300301) proteins were purchased from commercial vendors. 10 and 20 ng of either HSP60 or ClpP protein were dot blotted on a nitrocellulose membrane using a dot blot apparatus (BioRad).
The membrane was blocked with 5% BSA and incubated with 5 µg/ml purified ClpP (for HSP60 dot blot) or HSP60 (for ClpP dot blot) protein in 5% BSA with or without 20 µM DCEM1 for 2 h at room temperature. Membranes were washed with TBST and then incubated with HSP60 and ClpP specific antibodies for 2 h at room temperature. After washing with TBST, membranes were incubated with appropriate secondary antibodies for 40 minutes at room temperature. After washing, blots were developed using standard chemiluminescence detection. To show specificity, in one set we denatured protein by heating reaction mixture at 80 0 C for 10 minutes.

In vitro co-immunoprecipitation
To detect direct interaction between HSP60 and ClpP, we performed in vitro coimmunoprecipitation using recombinant proteins as described previously with slight modifications (10). Briefly, 200 ng of purified HSP60 protein was incubated with 200 ng of purified ClpP protein for 2 h at 4 °C in immunoprecipitation (IP) buffer (50 mM Tris HCl, 150 mM NaCl, 1% NP-40, 5% glycerol and PIC) with or without 20 µM DCEM1. 1 µg of either HSP60 or ClpP antibody was added to the mix and further incubated at 4 0 C for overnight. Next day, 25 µl of precleared Protein A/G magnetic beads (ThermoFisher Scientific) was added, and the mixture was incubated for another 2 h at 4 °C. The beads were then washed 5 times in the above-mentioned IP Buffer at room temperature using a DynaL magnetic stand (ThermoFisher Scientific). Respective IgG was used as negative control. Bound proteins were eluted in 1.2x SDS loading buffer heated at 80 °C for 10 minutes.

Quantification of apoptosis by Annexin V and propidium iodide (PI) staining
PCa cells were harvested at indicated time points and percentage apoptotic cell death was determined using annexin-V-alexafluor 488 / PI kit (Thermo Fisher Scientific, Cat # V13245) according to the manufacturer's instructions. The stained cells were analyzed by flow cytometry (LSRIIA, BD Biosciences) collecting 10,000 events and data were analyzed using Win List 3D software (4).

DEVDase activity measurement
The caspase-3/7 (DEVDase) activity was measured using colorimetric approach method described earlier (4,13). In a 96 well plate, 5 µL of WCL from each group was incubated with 40 µL of was taken as a standard. The fluorescence obtained were normalized by their respective protein concentration. The results were presented as fold change compared to respective controls.

Double immunofluorescence (DIF) and confocal microscopy
Briefly, cells (5000 cells) on coverslips were stained live with DAPI (1 μg/ml) for 15 minutes at 37°C in CO2 incubator. Cells were fixed with 4% paraformaldehyde containing 5% sucrose for 30 minutes at RT followed by permeabilization with 1% Triton X-100 in PBS for 10 minutes.
Following washing and blocking with 10% goat serum containing 1% Triton X-100 in 1X PBS and washed with 1X PBS twice, respective primary antibodies (HSP60, HSP10 and ClpP) were applied for overnight at 4°C. Alexafluor-488-conjugated and alexafluor-594-conjugated secondary antibodies were added for 2 h. After washing with 1X PBS twice, coverslips were mounted on glass slides using ProLong Gold Antifade Mountant with DAPI as mounting medium. Fluorescent images were acquired using a laser-scanning confocal system (Zeiss LSM710) on an inverted microscope equipped with an oil immersion lens (3,4,6).

Cellular reactive oxygen species (ROS), mitochondrial ROS (mitoROS) and mitochondrial membrane potential (mitoMP) using flow cytometry
Flow cytometry analysis was carried out to illustrate cellular and mitoROS, and mitoMP in control and treated groups as described previously (4,14). The cellular ROS, mitoROS, and mitoMP were estimated in the control and treated groups using DHR-123, MitoSox Red, MitoTracker Green, and MitoTracker Orange probes (Life Technologies, USA), respectively. In brief, control and treated cells were harvested and centrifuged at 1000 rpm for 5 minutes. Cells were then resuspended in 1 ml of 1X PBS and centrifuged for 5 minutes at 2500 rpm. The pellets were stained with respective probes and incubated for 30 minutes in the dark at 37°C. Cells were centrifuged at 1500 rpm for 5 minutes and pellets were re-suspended in 1X PBS and 10,000 cells were analyzed using flow cytometry (LSR II, BD Biosciences). Data were analyzed by WinList 3D 7.1 software and presented as the fold changes of geometric mean in comparison to untreated control.

Quantification of ATP levels
Hsp60 and ClpP silenced LNCaP and PC-3 cells were seeded in a 96 well white culture plate (5000 cells per well). After 24 h, ATP levels were assayed using luminescent ATP detection assay kit (Abcam, Cat # ab113849) as per manufacturer's instructions. Luminescence signals were read using BioTeK luminometer plate reader. Similarly, ATP levels were estimated in DU145 parental and HSP60 +/cells. In another experiment, LNCaP and PC-3 cells were seeded in a 96 well white culture plate (5000 cells per well). After 24 h cells were treated with DCEM1 (5, 10, and 20 µM), followed by ATP measurement after 24 h as described above. TKO prostatic tumor tissues were homogenized in 1% NP40 buffer and 10 µg protein were diluted in PBS to the volume of 50 µl and ATP levels were estimated as described above.

Measurement of ATP/ADP ratio
In a white 96 well plate, 5 x 10 3 cells were transfected with siRNA against Hsp60 and ClpP. After 48 h the ATP/ADP ratio was measured using the ATP/ADP ratio assay kit (Millipore Sigma) according to the manufacturer's protocol. In another set of experiment, cells were treated with DCEM1 for 24 h and ATP/ADP ratio was measured. TKO prostatic tumor tissues were homogenized in 1% NP40 buffer and 10 µg protein were diluted in assay buffer to the volume of 10 µl followed by estimation of ATP/ADP levels.

Chemical cross-linking and oligomerization assays
Freshly harvested cells were suspended in 45 μl of HIM buffer (10 mm HEPES-KOH, 200 mm mannitol, 70 mm sucrose, 1 mm EGTA, pH 7.5) followed by addition of freshly prepared bismaleimidohexane (BMH) or ethylene glycol bis(succinimidylsuccinate) to a final concentration of 2 mm and incubated at room temperature for 30 minutes. Cells were then mixed with protein sample buffer and subjected to Western blotting (6).

Mitochondrial DNA damage analysis
Mitochondrial DNA damage was analyzed as described previously (17). Briefly, PCa cells were

Seahorse XF96 bioenergetics mitostress assay
Different parameters associated with mitochondrial respiration such as oxygen consumption rate, basal respiration, maximal respiration, ATP production, and spare respiratory capacity in PCa cell lines were assessed by MitoStress assay by using Seahorse XF96 (Seahorse Biosciences, North Billerica, MA, USA) analyzer as described previously (4). Coli strain and were isolated using a Zymo Plasmid MidiPrep kit (Zymo research, cat # D4200).
LNCaP cells were seeded in 96 well plates and transfected with either pLightSwitch empty vector or CLPP promoter reporter clones using a lipofectamine 3000 transfection kit (ThermoFisher Scientific, Waltham, MA). Luciferase activity assay was determined after 48 h using LightSwitch Assay Reagent (SwitchGear Genomics, Menlo Park, CA. Cat # LS010).

Viability assay and clonogenic survival assay
Cell viability was assayed via Trypan blue exclusion. Briefly, 10 μl of cell suspension was added to 10 μl of trypan blue, and 10 μl of the cell suspension-trypan blue solution was added to a hemocytometer. Live and dead cells were counted, and viability was calculated. For the clonogenic survival assays, stable Hsp60 and ClpP silenced LNCaP and PC-3 PCa cells, as well as DU145 parental and DU145 Hsp60 +/cells, were seeded at a density of 500 cells per well in a 6-well tissue culture dish and were cultured for 7 days. Cells were washed with 1X PBS, fixed with chilled 95% ethanol for 10 minutes, and stained with crystal violet solution (0.05% in distilled water) for 30 minutes at room temperature and observed under phase contrast microscope (Carl Zeiss, Thornwood, NY) as previously described (18). In another experiment, PCa cells were seeded at a density of 500 cells per well in a 6-well tissue culture dish and after 24 h treated with DCEM1 (5 µM). After 6 days of treatment, cells were washed with 1X PBS, fixed with chilled 95% ethanol for 10 minutes, and stained with crystal violet solution as described above.

ClpP protease activity in purified mitochondria
Purified mitochondria from PCa cells were isolated as described above. ClpP protease activity in mitochondria was assayed as described earlier (19). In brief, 0. were stored at 4°C in the dark. For human PCa TMA slides, the same methods were followed as mentioned in the IHC protocol up to antigen retrieval and further followed the exact protocol of the in situ PLA protocol mentioned above in this section. Subcellular localization of proteinprotein interactions was detected using confocal microscope (Zeiss LSM710) (20).

Magnetic resonance imaging (MRI) imaging
Magnetic resonance imaging was carried out on a 4.7 Tesla preclinical scanner using the  (21) and previously reported MR images of mouse prostate (22). Volumes for each ROI were calculated in Analyze and are reported herein.

Mitochondrial Chaperonin Activity Assay (MiCAA)
We analyzed mitochondrial chaperonin activity assay as reported earlier with modifications (23).
To analyze the endogenous mitochondrial chaperonin (protein folding) activity, we co-transfected

Structure-based virtual drug screening for HSP60 inhibitors
Virtual screening experiments to identify HSP60 inhibitors were performed with Schrödinger software (Small molecule drug discovery suite, v2018-2). The X-ray crystal structure of the HSP60-HSP10 human mitochondrial chaperonin complex (PDB 4PJ1) was used as the model for docking experiments. Proteins were prepared for docking studies using the Protein Preparation Wizard (25). Specifically, we elected to assign bond orders, add hydrogen atoms, create bonds between sulfur atoms within 3.2 Å, and remove crystallographic water molecules > 5 Å from heteroatom groups. The protein hydrogen-bonding network was optimized using the automated optimization at pH 7.0 and the overall structure was minimized to the lowest energy state using the OPLS3 force field. The virtual chemical library was developed from the Enamine 3D Diversity Set (based on 3D shape diversity) and Enamine pharmacological diversity set (based on compounds with known pharmacological effects). Ligands were prepared using LigPrep (Schrödinger) and settings were as follows: OPLS3 force field, ionization states at pH 7.0 +/-2.0, ligands were desalted if necessary, any possible tautomers were generated, specified chiralities were retained, and the lowest energy ring conformation was kept. Using the GLIDE program as described (26), we set up a virtual screening platform to identify HSP60 inhibitors. All docking settings were set to default. To progressively filter out hits from non-hits, we performed three stages of GLIDE docking, each stage with an increased stringency: HTVS (High-Throughput

Virtual Screening), SP (Standard Precision), and XP (Extra Precision). Top scoring molecules
were sorted based on glide score (-kcal/mol) and the top 25% from HTVS were processed to SP docking and the top 15% from SP were processed to XP docking. The top 50-100 scored molecules after XP docking were visually inspected for differences in molecular interactions, erroneous poses, or similar binding modes (pharmacophores). The top molecules were subsequently clustered based on structural similarity to identify around 10-20 structurally distinct clusters. A single representative molecule from each cluster was selected for purchase.

Xenograft studies
Roswell          Actin serves as a loading control. (B) PARL protein was over-expressed in LNCaP cells followed by DCEM1 (10 µM) treatment and level of poly-ubiquitinated proteins and DEVDase activity was analyzed. Actin serves as a loading control. Data are mean ± SD (n=3). * p<0.05 compared to respective controls by 1-way ANOVA followed by Tukey's multiple-comparison test (A and B). Body weight (gm)

Supplementary Figures and Tables
Week 0 Week 1 Week 2 Week 3 Week 4 Week 5 ns ns