park2 promotes mitochondrial pathway of apoptosis and antimicrotubule drugs ... · 1 1 park2...
TRANSCRIPT
1
PARK2 promotes mitochondrial pathway of apoptosis and 1
antimicrotubule drugs chemosensitivity via degradation of 2
phospho-BCL-2 3
Hengxing Chen1, †, Yun Li1, †, Yu Li1, Zhen Chen1, Limin Xie1, Wenjia Li1, Yuanxin Zhu1, 4
Hong Xue2, H. Phillip Koeffler3,4, Wenjing Wu5，Kaishun Hu1, *, Dong Yin1, * 5
6
Running title: PARK2 is a biomarker for antimicrotubule drugs 7
8
1Research Center of Medicine Guangdong Provincial Key Laboratory of Malignant Tumor 9
Epigenetics and Gene Regulation, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, 10
Guangzhou 510120, China; 11
2Division of Life Science, Applied Genomics Centre and Centre for Statistical Science, Hong 12
Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China; 13
3Cancer Science Institute of Singapore, National University of Singapore, Singapore; 14
4Division of Hematology/Oncology, Cedars-Sinai Medical Center, University of California 15
Los Angeles School of Medicine, Los Angeles, CA, USA; 16
5Department of Breast Oncology, Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University, 17
Guangzhou, 510120, China; 18
19
†These authors contributed equally to this work. 20
* Corresponding authors 21
22
Corresponding authors: Prof. Dong Yin; Sun Yat-Sen Memorial Hospital, Sun Yat-Sen 23
2
University, Guangzhou 510120, China. Tel: +861892218251. E-mail: 24
[email protected] or Dr. Kaishun Hu; Sun Yat-Sen Memorial Hospital, Sun Yat-Sen 25
University, Guangzhou 510120, China. Tel: +8613560464095 E-mail: 26
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
3
Abstract 46
Rationale: Neoadjuvant chemotherapy has become the standard treatment of locally 47
advanced breast cancer. Antimicrotubule drugs and DNA-damaging drugs are the most 48
popular medicines used for neoadjuvant chemotherapy. However, we are unable to predict 49
which chemotherapeutic drug will benefit to an individual patient. PARK2 as a tumor 50
suppressor in breast cancer has been reported. While the role of PARK2 in chemotherapy 51
response remains unknown. In this study, we explore the impact of PARK2 on 52
chemosensitivity in breast cancer. 53
Methods: PARK2 expression in breast cancer patients with different neoadjuvant 54
chemotherapeutic regimens was studied using immunohistochemistry. Data was correlated to 55
disease-free survival (DFS), overall survival and pathologic complete response (pCR). The 56
functional roles of PARK2 were demonstrated by a series of in vitro and in vivo experiments. 57
Including mass spectrometry, Co-immunoprecipitation, isolation of subcellular fractionation, 58
fluorescence microscopy, in vivo ubiquitination assay and luciferase analyses. 59
Results: Highly expressed PARK2 predicted better response to antimicrotubule 60
drugs-containing regimen associated with higher rate of pathologic complete response (pCR). 61
In contrast, PARK2 expression did not predict response to the DNA-damaging drugs regimen. 62
Following antimicrotubule drugs treatment, levels of PARK2 was upregulated due to the 63
repression of STAT3-mediated transcriptional inhibition of PARK2. Moreover, 64
overexpression of PARK2 specifically rendered cells more sensitive to antimicrotubule drugs, 65
but not to DNA-damaging drugs. Depletion of PARK2 enhanced resistance to antimicrotubule 66
drugs. Mechanistically, PARK2 markedly activated the mitochondrial pathway of apoptosis 67
4
after exposure to antimicrotubule drugs. This occurred through downregulating the 68
antiapoptotic protein, phospho-BCL-2. BCL-2 phosphorylation can be specifically induced by 69
antimicrotubule drugs, whereas DNA-damaging drugs do not. Notably, PARK2 interacted 70
with phospho-BCL-2 (Ser70) and promoted ubiquitination of BCL-2 in an E3 71
ligase-dependent manner. Hence, PARK2 significantly enhanced the chemosensitivity of 72
antimicrotubule drugs both in vitro and in vivo, while loss-of-function PARK2 mutants did 73
not. 74
Conclusions: Our findings explained why PARK2 selectively confers chemosensitivity to 75
antimicrotubule drugs, but not to DNA-damaging drugs. In addition, we identified PARK2 as a 76
novel mediator of antimicrotubule drugs sensitivity, which can predict response of breast 77
cancer patients to antimicrotubule drugs-containing regimen. 78
79
Keywords 80
PARK2; antimicrotubule drug; docetaxel; phospho-BCL-2; breast cancer 81
82
83
84
85
86
87
88
Graphical abstract 89
5
90
91
Introduction 92
Breast cancer is one of the most leading cause worldwide of cancer morbidity and mortality in 93
women. Neoadjuvant chemotherapy (NAC) for advanced breast cancer is considered one of 94
the most crucial treatment in reducing mortality (1, 2). Thus, NAC become the standard 95
treatment of locally advanced breast cancer, which can downstage the disease and improve 96
surgical option (3). Antimicrotubule drugs and DNA-damaging drugs are commonly used 97
NAC drugs (1). Antimicrotubule drugs, such as taxane and vinca alkaloid suppress 98
microtubule dynamics, lead to G2/M cell cycle arrest, and trigger apoptosis through the 99
mitochondrial pathway (4-6). DNA-damaging drugs, such as alkylator and anthracyclines 100
cause cell death through directly destroying DNA. Sadly, only a small percentage of patients 101
receiving NAC achieve a complete response, while the others suffer from the side effects of 102
chemotherapy without benefit from it. Ideally, patients who would benefit from a particular 103
chemotherapy regimen will be identified, protecting the remaining from side effects. 104
Therefore, development of predictive biomarker are urgently need to guide the use of NAC 105
regimen. 106
6
Antimicrotubule drugs induce cellular apoptosis. However, antimicrotubule drugs also trigger 107
stress responses that inhibit apoptosis and promote emergence of an acquired 108
treatment-resistant phenotype. Antiapoptotic protein, BCL-2, has a key roles in these stress 109
responses. BCL2 protein inhibits apoptosis by binding to BAX protein and inhibiting the 110
activation of the latter (7-9). Antimicrotubule drugs induce phosphorylation of BCL-2 on its 111
serine (Ser) residues, Ser70. Phospho-BCL-2 (S70) greatly enhances BCL-2’s antiapoptotic 112
function. Briefly, mono- or multisite phosphorylation can enhance BCL-2’s survival activity 113
(10). JNK induced-BCL-2 phosphorylation at Ser70 prolongs cell survival following stress 114
stimuli (11). BCL-2 phosphorylation is required for full BCL-2 death-suppressing signaling 115
activity (12). Phospho-BCL-2 (S70) indeed enhanced BCL-2’s antiapoptotic activity in 116
MCF-7 cells (Figure S4E-G). 117
Loss-of-function mutants of PARK2 have been identified as one of the most common causes 118
of recessive familial early onset Parkinsonism (13, 14). PARK2 is mutated and/or deleted in a 119
variety of human malignancies (15, 16). It is a tumor suppressor in a wide variety of human 120
cancers including breast cancer, whose loss significantly increases tumorigenesis (17, 18). 121
PARK2 gene encodes a well-conserved RING-in-between-RING (RBR) type E3 ubiquitin 122
ligase Parkin, which targets specific substrates through the ubiquitin-proteasome system for 123
degradation (19). PARK2 is also involved in the regulation of various cellular processes, 124
including stress response and mitochondrial biogenesis (20). 125
In this study, we report a significant correlation between PARK2 expression and 126
chemosensitivity to antimicrotubule drugs-containing regimen in breast cancer patients. 127
Among patients who received antimicrotubule drugs, higher expression of PARK2 showed 128
7
better disease-free survival (DFS), overall survival and higher rate of pathologic complete 129
response (pCR). In contrast, no significant differences were noted between PARK2 expression 130
and survival in those who received DNA-damaging drugs (without antimicrotubule drugs). 131
Moreover, overexpression of PARK2 renders cells more sensitive to antimicrotubule drugs in 132
breast cancer cell lines, which is not observed in a range of DNA-damaging drugs. In addition, 133
our studies reveal that antimicrotubule drug stimulates PARK2 expression via downregulating 134
STAT3. PARK2 markedly activates the antimicrotubule drug-induced mitochondrial pathway 135
of apoptosis by degrading phospho-BCL-2 (Ser70). Taken together, our study shows that 136
PARK2 enhances chemosensitivity of breast cancer cells to antimicrotubule drugs and 137
provides a predictive biomarker for antimicrotubule drugs-containing regimen. 138
139
Materials and Methods 140
Reagents and antibodies 141
For details, please refer to Table 2. 142
143
Plasmids 144
The pEGFP-parkin WT, pRK5-HA-Ubiquitin-WT, and pRK5-HA-Ubiquitin-KO were 145
obtained from Addgene (Plasmid #45875, Plasmid #17608, and Plasmid #17603). Full-length 146
cDNA coding for PARK2 was obtained from the cDNA of human HEK293 cells using PCR 147
and was then cloned into pLVX-DsRed-Monomer-N1 vector also having a FLAG-tag 148
sequence. Full-length cDNA coding for BCL-2 was obtained from the cDNA of human 149
MCF-7 cells using PCR and was cloned into pLVX-DsRed-Monomer-N1 vector also having a 150
8
MYC-tag sequence. Point mutations were introduced by site-directed mutagenesis. 151
SFB-PARK2 was first subcloned into pDONR201 entry vectors and transferred into 152
destination vector with the indicated SFB tag using Gateway Technology (Invitrogen, 153
Camarillo, CA, USA). Doxycycline-inducible PARK2 constructs were cloned from human 154
HEK293 cDNA and inserted into pRL-TetON-EGFP vector using EcoR I and Sali sites. All 155
shRNA constructs for PARK2 were made using pLKO.1 backbone employing Age I and 156
EcoR I sites. For sequences of shRNAs and siRNAs, please refer to Table 2. 157
158
Western blotting and Co-immunoprecipitation 159
RIPA buffer (150 mM NaCl, 50 mM Tris-HCl [pH 8.0], 0.5% Nonidet P-40, 5 mM EDTA and 160
a protease and phosphatase inhibitor cocktail (Bimake, Houston, TX, USA) were used to lyse 161
cells. SDS-PAGE resolved the clarified lysates. Samples were transferred to PVDF 162
membranes for Western blots and detected by ECL detection reagents (Beyotime, Shanghai, 163
China). For immunoprecipitation, supernatants were first incubated with anti-FLAG agarose 164
overnight at 4°C. Subsequently, pellets were washed three times with NETN buffer. To 165
examine endogenous interactions, cell lysates were divided into two parts, incubated with 166
either anti-IgG, anti-BCL-2 or anti-PARK2 for 2 h, and then incubated with protein G agarose 167
(Life Technologies, Waltham, MA, USA) overnight. Beads were then washed three times with 168
NETN buffer; and the samples were resolved on SDS-PAGE. 169
170
In vivo ubiquitination assay 171
The process was carried out as previously described (21). Briefly, MCF-7 cells were 172
9
transfected with either pEGFP-parkin WT, pRK5-HA-Ubiquitin-WT or 173
pRK5-HA-Ubiquitin-KO plasmid for 18 h. Cells were either treated or not treated with 174
docetaxel (1 nM) for 24 h. Cells were then cultured with MG132 (10 mM) followed by 175
immunoprecipitation (IP) analysis with BCL-2 antibody. Endogenous BCL-2 was 176
immunoprecipitated using anti-BCL-2 antibody for 12 h at 4°C. Anti-HA antibody was used 177
to detect poly-ubiquitinated BCL-2. 178
MDA-MB-134-VI cells were stably either deleted endogenous PARK2 or NC (control). These 179
cells were transfected with pRK5-HA-Ubiquitin-WT for 18 h, and were either treated or not 180
treated with docetaxel (2 nM) for 24h. Cells were then treated with MG132 (10 mM) for 4h 181
and followed by immunoprecipitation (IP) analysis with BCL-2 antibody. 182
183
Cytotoxicity assay 184
Cells were seeded in 96-well plates at 1,000 cells/well in culture media with a variety of 185
chemotherapy drugs for 48 h. Cell viability was analyzed using the MTT as previously 186
described (22). Briefly, 10μL of 5 mg/mL MTT solution was added to each well. The wells 187
were incubated in a humidified chamber for 4h and terminated by addition of 100μL of 10% 188
sodium dodecyl sulfate (SDS). The samples were mixed thoroughly until the formazan was 189
dissolved and then detected by 570 nm absorbance. 190
191
Patients and tissue samples 192
205 paraffin-embedded breast cancer specimens, which were pathologically and clinically 193
diagnosed at the Sun Yat-sen Memorial Hospital, were utilized. Patients’ consent and approval 194
10
from the ethical board of Sun Yat-sen University were obtained for research in use of clinical 195
materials. Clinical information on these patients, including tumor size, 196
chemotherapy regimens, survival situation, and etc, was obtained from past medical records 197
and recent follow-up records (Table S1). All the clinical diagnostic information and 198
clinico-pathological variables were classified according to the American Joint Committee. 199
Pathologic complete remission (pCR) was defined as disappearance of invasive tumor lesion 200
in the surgically removed breast and axillary lymph nodes after chemotherapy. 201
202
Immunohistochemical staining and scoring 203
IHC staining was performed as described previously. The histopathological features of the 204
stained tumor were evaluated by two researchers who were unaware of the patient's clinical 205
characteristics. Intensity of PARK2 immunostaining of breast cancer tissues was scored as 206
negative (0), weak (1), medium (2) or strong (3). The extent of staining, defined as the percent 207
positive staining cells, was scored as 1 (≤ 10%), 2 (11–50%), 3 (51–75%) or 4 (>75%). 208
Staining index (SI) was calculated as follows: SI = staining intensity × proportion of 209
positively stained cells. Low and high expression were defined as scores of < 7 and ≥ 7, 210
respectively. 211
212
Mass spectrometry (label-free protein quantification) 213
4 plates of MCF-7 cells which stably expressed SFB-tagged PARK2 were divided into two 214
groups, either treated or not treated with docetaxel (1 nm) for 24 h, respectively. Followed by 215
cells lysed with NETN buffer for 20 min. Supernatant was removed via centrifugation at 216
11
15,000g to remove debris and then incubated with streptavidin-conjugated beads (Amersham 217
Biosciences, Franklin Lakes, NJ, USA) for 12 h at 4°C. Beads were washed three times with 218
NETN buffer, followed by incubated with S-protein beads (Sigma, St. Louis, MO, USA) for 6 219
h at 4°C. Beads were washed three times with NETN buffer followed by digestion of lysed 220
cells. Peptides were analyzed by LC-MS/MS. Label-free quantification was carried out at the 221
peptide level. These peptides were identified from individual second dimension fractions 222
which were assembled into a master list. Median of the computed peptide ratios represents 223
relative quantity of the protein in the two samples. Quantitative proteomics used a label-free 224
method to measure relative protein abundance. 225
226
Bioinformatics and data analysis 227
Copy-number data of breast cancer were collected from The Cancer Genome Atlas (TCGA) 228
via cBio Cancer Genomics Portal and analyzed with IGV software. Copy-number data of 229
breast cancer cell lines were obtained from the Cancer Cell Line Encyclopedia database. 230
Correlations between the overall survival of breast cancer patients and the expression of 231
PARK2 by their tumor were analyzed via KM plotter. Analysis of sensitivity of PARK2 gene 232
to chemotherapy drugs were collected from the ONCOMINE database. Correlations between 233
the BCL-2 protein level and PARK2 mRNA expression was analyzed by The Cancer Genome 234
Atlas and The Human Protein Atlas database. Somatic PARK2 mutations in cancer were 235
obtained from cBio Cancer Genomics Portal. 236
237
Statistical analysis 238
12
GraphPad Prism version 7 was used to analyze the data. Each group was analyzed three times. 239
All experiments subjected to statistical analysis were repeated at least three times. Results are 240
presented as mean ± standard error of mean (SEM) and comparisons were made using T-test 241
statistical analysis (non-parametric). 242
243
The rest of the Materials and Methods are shown in the Supplementary Information. 244
245
Results 246
PARK2 expression predicts clinical outcome of antimicrotubule drugs-containing 247
regimen 248
PARK2 gene is located on chromosome 6, it is often mutated and/or deleted in cancer. Copy 249
number loss of PARK2 gene is the most frequent somatic alteration (15, 16). Therefore, we 250
examined PARK2 deletion in The Cancer Genome Atlas (TCGA), assembling data from 251
1,283 primary breast tumors. A total of 34% of breast cancer patients exhibited a deletion of 252
PARK2 (Figure 1A). Copy number loss of PARK2 also occurred in 53% (32/60) of breast 253
cancer cell lines (Figure S1A). Next, the prognostic value of PARK2 expression in breast 254
cancer was analyzed. Higher expression of PARK2 was significantly correlated with better 255
survival of patients with breast cancer (Kaplan-Meier plotter database) (Figure S1B). We 256
classified breast cancer into different molecular subtypes, including basal, HER2+, luminal A, 257
and luminal B. Statistically, higher PARK2 expression was strongly correlated with better 258
survival in patients with basal, HER2+, and luminal A subtypes. Although the trend was the 259
same, this correlation within luminal B subtype was not statistically significant (Figure S1C). 260
13
Furthermore, we divided the data set into two groups: those who received chemotherapy, and 261
those who did not. Interestingly, for patients who did not receive chemotherapy, no difference 262
occurred in survival between patients with either high or low expression of PARK2. However, 263
among patients who received chemotherapy, higher PARK2 expression was positively 264
correlated with better survival (Figure S1D). Together, these finding suggest that prominent 265
PARK2 levels enhance chemotherapy in breast cancer. 266
Breast cancer patients received neoadjuvant chemotherapy including either antimicrotubule 267
drugs or DNA-damaging drugs, but the details of chemotherapy were not included in Kaplan 268
Meier plotter database. Indicators to determine the efficacy of chemotherapy include 269
disease-free survival (DFS), overall survival and pathologic complete response (pCR). Thus, 270
we collected 205 tumor tissues samples from breast cancer patients and analyzed the 271
expression of PARK2. Immunohistochemical analysis showed that 95 were PARK2-negative 272
(-), 111 were PARK2-positive (+) (Figure 1B). Highly expressed PARK2 showed significantly 273
better disease-free survival (DFS) and overall survival than those with tumors with low 274
PARK2 expression (Figure 1C). A total of 62 patients did not receive neoadjuvant 275
chemotherapy. No significant differences accured in their DFS and overall survival regardless 276
of the amounts of PARK2 in the tumors (Figure 1D). In contrast, for the remaining 143 277
patients, who receive neoadjuvant chemotherapy, higher expression of PARK2 in tumors was 278
associated with better DFS and overall survival (Figure 1E). This is consistent with the 279
previous database analysis. Furthermore, we divided the patients who received neoadjuvant 280
`chemotherapy into two groups: those who received antimicrotubule drugs-containing 281
regimen, and those who received DNA-damaging drugs regimen (without antimicrotubule 282
14
drugs). Among patients who received DNA-damaging drugs regimen (without 283
antimicrotubule drugs), PARK2 expression was not associated with DFS and overall survival 284
(Figure 1F). However, for patients who received antimicrotubule drugs-containing regimen 285
treatment, higher expression of PARK2 was significantly correlated with better DFS and 286
overall survival (Figure 1F). Among these patients, PARK2 level in the pathologic complete 287
response (pCR) group was substantially higher than that in the non-pathologic complete 288
response (non-pCR) group (Figure 1G&H). Moreover, we found that PARK2 mRNA level in 289
the Capecitabine/Docetaxel responding group was substantially higher than that in the 290
non-responding group (ONCOMINE database) (Figure 1I). 291
Collectively, these findings suggested that PARK2 expression could predict the clinical 292
outcome of antimicrotubule drugs-containing regimen therapy for breast cancer. 293
294
PARK2 regulates antimicrotubule drugs sensitivity in breast cancer cells 295
Chemotherapy drugs used in breast cancer include docetaxel, paclitaxel, adriamycin, 296
cyclophosphamide, 5-fluorouracil and vinorelbine. These can be divided into two categories: 297
antimicrotubule agents (Docetaxel, paclitaxel and vinorelbine) and DNA-damaging agents 298
(Adriamycin, cyclophosphamide and 5-fluorouracil) (4). 299
To confirm that overexpression of PARK2 only contributes to antimicrotubule drugs 300
sensitivity, we ectopically expressed PARK2 in MCF-7 and MDA-MB-231 cells. Cytotoxicity 301
assays were done, and because of close similarity of docetaxel and paclitaxel, only the former 302
was tested. Forced expression of PARK2 rendered breast cancer cells more sensitive to 303
docetaxel and vinorelbine in both MCF-7 and MDA-MB-231 cells (Figure 2A and S2A), but 304
15
overexpression of PARK2 did not change the sensitivity of DNA-damaging drugs (Figure 305
S2B). Flow cytometry showed that PARK2 overexpressing cells (MCF-7, MDA-MB-231) led 306
to marked increase in apoptosis when exposed to either docetaxel or vinorelbine (Figure 2B). 307
This was consistent with a significantly higher apoptotic response, as indicated by the 308
proteolysis of multiple caspase substrates (Figure 2C and 4E). Moreover, the higher PARK2 309
expression was paralleled by the greater sensitivity of docetaxel and vinorelbine in MCF-7 310
cells (Figure 2D). Furthermore, shRNA silencing of endogenous PARK2 resulted in 311
significantly decreased sensitivity to docetaxel and vinorelbine in MDA-MB-134-VI cells 312
(MDA-MB-134-VI: inherently high expression of PARK2) (Figure 2E and S2C). Flow 313
cytometry showed that silencing of endogenous PARK2 reduced cell apoptosis after either 314
docetaxel or vinorelbine exposure (Figure 2F). 315
Taken together, these results suggested that PARK2 conferred sensitivity to antimicrotubule 316
drugs in breast cancer cells. 317
318
Antimicrotubule drugs stimulates PARK2 expression 319
To evaluate changes in PARK2 expression after antimicrotubule drugs treatment, qPCR and 320
western blot were performed. Both mRNA and protein expression of PARK2 were 321
significantly increased in T47D, ZR-75-30 and MDA-MB-134-VI cells when exposed to 322
docetaxel (Figure 3A). Docetaxel also induced PARK2 expression in a dose-dependent 323
manner in T47D, ZR-75-30 and MDA-MB-134-VI cells (Figure 3B). 324
To understand how PARK2 levels are upregulated by antimicrotubule treatment in breast 325
cancer cells, we predicted putative transcription factors (TF) by overlaying the ChIP-seq data 326
16
from the ChIP-Atlas database, the UCSC Genome Browser and the virtual laboratory of 327
PROMO (Figure 3C). Intersection of data from the three databases showed seventeen TFs. 328
Among these TFs, STAT3 protein has been reported to be decreased in response to 329
antimicrotubule drugs treatment (23-25), whereas STAT3 is frequently enriched at the PARK2 330
promoter (Figure 3D). Indeed, Chip assay showed that STAT3 was enriched at the PARK2 331
promoter. After docetaxel treatment, the amount of STAT3 enriched on the PARK2 promoter 332
was significantly reduced (Figure 3E). Luciferase assay showing that STAT3 knockdown 333
increased PARK2 promoter activity in MDA-MB-134-VI cells (Figure 3F). We also found 334
that STAT3 knockdown increased the PARK2 levels in ZR-75-30 and MDA-MB-134-VI cells 335
(Figure 3G), whereas STAT3 overexpression downregulated the PARK2 levels in both cell 336
lines (Figure S2D). 337
338
PARK2 activates the mitochondrial pathway of apoptosis after exposure to 339
antimicrotubule drugs 340
Antimicrotubule drugs induce cells to enter the mitochondrial apoptotic pathway (5, 26). 341
Antimicrotubule drugs activate the pro-apoptotic protein, BAX, which forms oligomers. 342
These activated BAX dimers translocate to the mitochondria and induce release of 343
cytochrome C into the cytoplasm, resulting in proteolytic caspase activation followed by 344
DNA fragmentation, and ultimately lead to cell death (7, 27). 345
To examine whether PARK2 activated the mitochondrial apoptosis pathway upon 346
antimicrotubule drug treatment, immunofluorescence assays were performed. 347
Docetaxel-treated MCF-7-PARK2 cells were stained with immunofluorescent BAX (6A7) 348
17
antibody (this antibody specifically detects the activated form of BAX) (28). Overexpression 349
of PARK2 significantly promoted BAX activation and translocation into the mitochondria in 350
MCF-7 treated with docetaxel (Figure 4A). Meanwhile, fractionation assay of the 351
mitochondria showed that overexpression of PARK2 resulted in significant increase in BAX 352
oligomerization following docetaxel treatment of MCF-7 cells (Figure 4B). JC-1 assay 353
showed that overexpression of PARK2 significantly reduced the mitochondrial membrane 354
potential after docetaxel treatment of MCF-7 cells (Figure 4C). This is consistent with a 355
previous report that activated BAX dimer translocated into the mitochondria promoting 356
decrease of mitochondrial membrane potential (Δψm) (29). Moreover, we observed that 357
overexpression of PARK2 in MCF-7 cells significantly promoted the mitochondrial release of 358
cytochrome C (Figure 4D), followed by activated proteolytic caspases (Figure 4E). To 359
examine whether endogenous high expression of PARK2, activated and translocated BAX to 360
the mitochondria upon docetaxel treatment, immunofluorescence assay and western blot were 361
performed using MDA-MB-134-VI cells. We found that BAX protein was activated and 362
translocated into the mitochondria in MDA-MB-134-VI cells after docetaxel treatment. 363
ShRNA silencing of endogenous PARK2 resulted in significantly decrease BAX 364
oligomerization and mitochondrial translocation (Figure S3A&B). 365
In summary, PARK2 significantly activated the mitochondrial apoptotic pathway after 366
docetaxel exposure. 367
368
Antimicrotubule drugs promote the interaction between PARK2 and 369
phospho-BCL-2 370
18
To investigate the molecular mechanisms by which PARK2 activates the mitochondrial 371
apoptosis pathway after exposure to antimicrotubule drugs, in-depth proteomic profiling was 372
performed. Proteomic changes of protein binding to PARK2 before and after docetaxel 373
treatment of MCF-7-PARK2 cells were analyzed (Figure 5A). In total, 290 differential 374
PARK2 binding proteins were identified (Table 1) with excellent reproducibility (Figure S4A). 375
Docetaxel treatment of MCF-7-PARK2 cells exhibited a highly distinct enrichment of binding 376
proteins compared to non-treated cells (Figure 5B). Based on the Gene Ontology (GO) 377
Analysis, apoptosis-related proteins were significantly enriched after docetaxel treatment, 378
with 20 proteins associated with apoptosis (Figure 5C). The hub protein is the one that plays a 379
vital role in the biological processes. In related pathways, the regulation of other gene is often 380
affected by the hub protein (30, 31). To determine which gene among the 20 genes is the hub 381
protein, protein-protein interaction networks (PPIN) based on the STRING database were 382
constructed through the cytoHubaa plug-in unit. We screened 6 hub proteins by PPIN analysis 383
(Figure 5D). To narrow our list of candidate targets, we focused on 6 hub proteins that 384
exhibited the most robust differences between the groups of pre- and post- docetaxel 385
treatment. BCL-2 protein was the most highly enriched after docetaxel treatment (Figure 5E), 386
with a statistically significant p-value (Figure 5F). Exogenous co-immunoprecipitation (co-IP) 387
assays further confirmed that PARK2 interacted with BCL-2 before docetaxel treatment, 388
while this interaction markedly increased after docetaxel treatment as determined on western 389
blot (Figure 5G). Endogenous co-immunoprecipitation (co-IP) assays also supported this 390
conclusion (Figure 5H). The interaction between PARK2 and BCL-2 also increased following 391
vinorelbine treatment (Figures S4B and S3C). 392
19
Antimicrotubule drugs induce phosphorylation of BCL-2 on its serine (Ser) residues Ser70 (5). 393
We examined whether PARK2 interacts with phospho-BCL-2. First, we examined 394
antimicrotubule drugs versus DNA-damaging drugs. We found the former induced the 395
phosphorylation of BCL-2 (Figure S4D), this is consistent with previous reports (26, 32, 33). 396
We also examined whether phosphorylation of BCL-2 at ser70 enhances BCL-2 antiapoptotic 397
functions in breast cancer cell. We abrogated BCL-2 phosphorylation by introducing a 398
conserved, non-phosphorylated alanine (A) at ser70 sites. We mimic BCL-2 phosphorylation 399
by introducing a glutamate (E) at the ser70 sites. Mutant BCL-2 was stably transfected of 400
MCF-7 cells, and clones expressing quantitatively similar levels of BCL-2 were selected and 401
tested (Figure S4E). Flow cytometry assay showed that BCL-2 mutant S70E significantly 402
reduced apoptosis after docetaxel treatment in MCF-7 cells (Figure S4F). MTT assays 403
showed that BCL-2 mutant S70E rendered cells more resistant to docetaxel in MCF-7 cells 404
(Figure S4G). However, BCL-2 mutant S70A had much reduce antiapoptotic activity when 405
compared with BCL-2-WT. These results demonstrated that phosphorylation of BCL-2 at 406
ser70 enhanced BCL-2 antiapoptotic functions. Endogenous immunoprecipitation (IP) assays 407
showed that PARK2 interacts with phospho-BCL-2 (Ser70) before docetaxel treatment, while 408
this interaction markedly increased after docetaxel treatment (Figure 5I). Furthermore, BCL-2 409
mutants S70A did not interact with PARK2 (Figure 5J). Taken together, PARK2 interacted 410
with phospho-BCL-2 (Ser70). 411
Collectively, these data indicated that antimicrotubule drugs promoted the interaction between 412
PARK2 and phospho-BCL-2. 413
414
20
PARK2 promotes poly-ubiquitination and degradation of BCL-2 415
We have experimentally demonstrated that antimicrotubule drugs promote the interaction 416
between PARK2 and BCL-2; however, what is the relationship between PARK2 and BCL-2? 417
An analysis of the TCGA dataset showed that BCL-2 protein levels were negatively 418
correlated with PARK2 mRNA expression in breast cancers (Figure 6A). A total of 70% of 419
breast cancers had a negative correlation between levels of PARK2 protein and BCL-2 protein 420
according to The Human Protein Atlas database (Figure S5A&B). As PARK2 has E3 ligase 421
activity, we hypothesized that PARK2 might promote BCL-2 degradation. To confirm this 422
hypothesis, a series of western blotting assays were performed, showing that BCL-2 was 423
slightly downregulated in MCF-7-PARK2 cell; however, BCL-2 was significantly decreased 424
after docetaxel treatment of MCF-7-PARK2 cells (Figure 6B). Meanwhile, PARK2 425
overexpression markedly reduced levels of phospho-BCL-2 (Ser70) (Figure 6B). Silencing of 426
endogenous PARK2 slightly elevated BCL-2 protein, but BCL-2 was dramatically increased 427
after docetaxel treatment in sh-PARK2-MDA-MB-134-VI cells (Figure 6C). Silencing of 428
endogenous PARK2 also significantly increased levels of phospho-BCL-2 (Ser70) (Figure 429
6C). Docetaxel treatment resulted in a substantial accumulation of the pro-apoptotic protein 430
BAX; however, PARK2 expression had no effect on BAX levels (Figures 6B&C). 431
Cycloheximide chase assay showed that after docetaxel treatment, cells overexpressing 432
PARK2 caused a dramatic reduction in the half-life of BCL-2 (Figure 6D). Furthermore, 433
proteasome inhibitor MG132 mitigated downregulation of BCL-2 confirming that 434
PARK2-mediated degradation of antiapoptotic protein BCL-2 was proteasome-dependent 435
(Figure 6E). Overexpression of PARK2 increased the level of ubiquitination of BCL-2; and 436
21
this was more evident after docetaxel treatment (Figure 6F). Consistently, regardless of 437
docetaxel treatment, silencing of endogenous PARK2 reduced ubiquitination of BCL-2 438
(Figure 6G). In addition, overexpression of PARK2 significantly downregulated BCL-2 439
protein levels in MCF-7 cells in response to vinorelbine treatment (Figure S5C). Silencing of 440
PARK2 markedly increased BCL-2 protein levels in MDA-MB-134-VI cells after vinorelbine 441
treatment (Figure S5D). 442
Moreover, immunofluorescence assay showed that phospho-BCL-2 (Ser70) was recruited to 443
the mitochondria in MCF-7 cells after docetaxel treatment, whereas overexpression of 444
PARK2 significantly reduced the recruitment of phospho-BCL-2 (Ser70). Meanwhile, 445
overexpression of PARK2 markedly reduced fluorescence intensity of BCL-2 and 446
phospho-BCL-2 (Ser70) after docetaxel treatment (Figure S5E). 447
Together, these data identified that PARK2 promoted degradation of the antiapoptotic protein 448
BCL-2 through the ubiquitin-proteasome pathway, and this was markedly enhanced by 449
antimicrotubule drugs. 450
451
Following antimicrotubule drugs treatment, PARK2 is not able to induce cell 452
death when BCL-2 is not degraded 453
Degradation of BCL-2 was blocked by MG132 in the presence of docetaxel treatment 454
condition and then checked for activation of BAX. Immunofluorescence assay showed that 455
overexpression of PARK2 was not able to promote BAX activation and translocation into the 456
mitochondria in MCF-7 cells after MG132 treatment (Figure S6A). Meanwhile, fractionation 457
assay of the mitochondrial showed that overexpression of PARK2 was not able to promote 458
22
BAX oligomerization in MCF-7 cells after MG132 treatment (Figure S6B). Further, we 459
observed that with MG132 treatment, overexpression of PARK2 was not able to promote the 460
mitochondrial release of cytochrome C (Figure S6C). Overexpression of PARK2 was not able 461
to activate proteolytic caspases following MG132 treatment of MCF-7 cells (Figure S6D). 462
Overexpression of PARK2 is not able to induce cell apoptosis after mg132 treatment (Figure 463
S6E). 464
In summary, PARK2 was not able to activate BAX and thus was not able to induce cell death 465
when BCL-2 was not degraded. 466
467
Under the condition of antimicrotubule drugs treatment, the pro-apoptotic effect 468
of PARK2 is link to BCL-2, not to BCL-XL and MCL-1 469
Because PARK2 significantly promoted BAX activation and activated the mitochondrial 470
pathway of apoptosis after exposure to antimicrotubule drugs, we considered the possibility 471
that PARK2 might promote degradation of more BCL-2 family proteins to promote BAX 472
activation, leading to apoptosis. Therefore, we monitored the stability of the major 473
antiapoptotic BCL-2 family proteins expressed in MCF-7 cells (BCL-2, BCL-XL and MCL-1) 474
in response to antimicrotubule drugs. BCL-2, but not BCL-XL, MCL-1 was significantly 475
downregulated after docetaxel treatment of MCF-7-PARK2 cells (Figure 7A). PARK2 476
mediated sensitivity to docetaxel was blocked through knockdown of BCL-2, suggesting that 477
PARK2 mediated-antimicrotubule drugs sensitivity is mediated by downregulation of BCL-2 478
(Figure 7B and S7A). Because the preceding results suggested that PARK2 promoted 479
degradation of BCL-2 after exposure to antimicrotubule drugs, we wondered whether 480
23
knockdown of BCL-2 was sufficient to sensitize antimicrotubule-induced apoptosis. As 481
controls, we also silenced other members of the antiapoptotic subset of the BCL-2 family 482
(BCL-XL and MCL-1). Knockdown of BCL-2 in MCF-7-PARK2 cells greatly sensitized 483
apoptosis induced by antimicrotubule drugs. In contrast, knockdown of BCL-XL and MCL-1 484
had much more modest effects (Figure 7C and S7A). 485
486
PARK2 enhancing antimicrotubule drugs sensitivity depends on its E3 ubiquitin 487
ligase activity 488
PARK2 point mutations can abolish E3 ligase activity (34). Interestingly, in the cBioPortal 489
database, PARK2 loss-of-function mutants (T173A, T240M, R275Q, D280N, P294S, and 490
E444Q) were found in several human cancers (Figure 7D). As the degradation of BCL-2 by 491
PARK2 is dependent on its E3 ubiquitin ligase activity, we speculated that PARK2 492
loss-of-function mutants would not promote BCL-2 degradation. We chose three 493
tumor-associated PARK2 loss-of-function mutations (T173A, T240M, P294S), as well as a 494
known ligase-dead mutant C431S. Ubiquitination assays were performed on MCF-7 cells. 495
Compared with wild-type PARK2, all of the loss-of-function mutants showed decreased 496
ubiquitination of BCL-2 (Figure 7E and S7B). These loss-of-function mutations blocked the 497
degradation of BCL-2 (Figure 7F and S7C). Furthermore, these mutations failed to enhance 498
antimicrotubule drugs sensitivity (Figure 7G and S7D). 499
These data demonstrated that the PARK2 loss-of-function mutations blocked the degradation 500
of BCL-2 and failed to increase antimicrotubule drugs sensitivity. 501
502
24
PARK2 enhances antimicrotubule drug sensitivity in vivo 503
To confirm further whether PARK2 contributes to antimicrotubule drugs sensitivity in vivo, 504
we established a breast cancer xenograft model (Figure 8A). Doxycycline-inducible MCF-7 505
cells were also fluorescently-labeled and overexpressed either vector (control), wild-type 506
PARK2 or loss-of-function mutant of PARK2 (PARK2-T240M). These cells were 507
orthotopically injected. Induction of wild-type PARK2 expression significantly enhanced 508
docetaxel sensitivity (Figures 8B-E). Furthermore, wild-type PARK2 dramatically increased 509
apoptosis of the tumor cells (Figure 8F). In contrast, PARK2-T240M had no effect on 510
docetaxel sensitivity in vivo (Figures 8B-F). Docetaxel decrease STAT3 level. Wild-type 511
PARK2, but not PARK2-T240M, markedly reduced BCL-2 levels and increased BAX 512
activation following docetaxel treatment (Figure 8G). 513
Our data showed that PARK2 enhanced antimicrotubule drug sensitivity by promoting the 514
degradation of BCL-2 in vivo. 515
516
Discussion 517
Increasing efforts to identify what chemotherapy will be effective for individual patients 518
underscore the need for applicable biomarkers to predict response to specific drug, in order to 519
achieve effective personalized treatment while saving patients from side effects. Here, for the 520
first time, we demonstrated that PARK2 regulates sensitivity to antimicrotubule drugs (Figure 521
8H) and provide a potential biomarker for predicting response to antimicrotubule drugs in 522
breast cancer. 523
In our present study, we analyzed expression of PARK2 in 205 tumor samples from breast 524
25
cancer patients. We found that for patients who received antimicrotubule drugs, highly 525
expressed PARK2 in breast cancer tissues was strongly associated with higher rate of pCR 526
(Figure 1H). Meanwhile, higher expression of PARK2 also showed better DFS and overall 527
survival (Figure 1F). In contrast, for the patients who received DNA-damaging drugs (without 528
antimicrotubule drugs), no significant differences were observed in their DFS and overall 529
survival, regardless of the expression of PARK2 in the tumors (Figure 1F). In vivo and in 530
vitro, PARK2 sensitized apoptosis induced by antimicrotubule drugs (e.g. docetaxel and 531
vinorelbine) in breast cancer cells (Figure 2A). This effect was highly selective and was not 532
observed in DNA-damaging drugs (e.g. adriamycin, cyclophosphamide and 5-fluorouracil) 533
(Figure S2B). Thus, efficacy of antimicrotubule drugs, but not DNA-damaging drugs, is 534
associated with levels of PARK2. 535
We investigate the molecular mechanisms by which PARK2 mediated-antimicrotubule drugs 536
sensitivity. We found that PARK2 significantly activated the mitochondrial apoptotic pathway 537
after docetaxel treatment (Figure 4A-E). Meanwhile, our in-depth proteomic profiling showed 538
that BCL-2 protein was the most highly enriched protein binding to PARK2 after docetaxel 539
treatment. A series of co-immunoprecipitation (co-IP) assays also confirmed that PARK2 540
interacted with BCL-2. 541
Antimicrotubule drugs induce cytoprotective stress responses that suppress apoptosis and 542
promotes chemoresistance. Antiapoptotic protein, BCL-2, plays a key role in these 543
cytoprotective stress responses. BCL-2 is a crucial regulator of mitochondrial apoptotic 544
pathway. Antimicrotubule drugs induced BCL-2 phosphorylation at Ser70 (5). BCL-2 545
phosphorylation greatly enhances antiapoptotic function of the molecular. Phospho-BCL-2 546
26
confers chemoresistance of antimicrotubule drugs (10, 12, 35). Furthermore, we found that 547
PARK2 specifically interacts with phospho-BCL-2 (Ser70) which is augmented by 548
antimicrotubule drugs. PARK2 degraded phospho-BCL-2 in an E3 ligase dependent manner 549
(Figure 6A-G). Antimicrotubule drugs induced BCL-2 phosphorylation, whereas 550
DNA-damaging drugs do not (26, 32, 33). This may explain why PARK2 selectively confer 551
chemosensitivity to antimicrotubule drugs, but not to DNA-damaging drugs. Collectively, 552
PARK2 significantly improves sensitivity to antimicrotubule drugs via degradation 553
phospho-BCL-2. 554
Antiapoptotic family proteins includes BCL-2 and its close relatives BCL-XL and MCL-1. 555
PARK2 has also been reported to enhance apoptosis induced by mitochondrial depolarization 556
drugs by promoting poly-ubiquitination and degradation of MCL-1 (20). We considered the 557
possibility that PARK2 might promote degradation of additional BCL-2 family proteins after 558
antimicrotubule drugs treatment. We showed that BCL-2, but not BCL-XL, MCL-1 was 559
significantly downregulated after docetaxel treatment of MCF-7-PARK2 cells (Figure 7A). 560
Sensitivity to docetaxel was enhanced by PARK2 which was blocked by knockdown of 561
BCL-2 (Figure 7B). Thus, under antimicrotubule drugs treatment, the pro-apoptosis effect of 562
PARK2 is linked to BCL-2, not to BCL-XL and MCL-1. 563
In conclusion, we showed that elevated expression of PARK2 may play an important in 564
enhancing sensitivity to antimicrotubule drugs in breast cancer. In addition, PARK2 565
mediated-antimicrotubule drug sensitivity is, at least in part, mediated by downregulation of 566
phospho-BCL-2 which activated the mitochondrial pathway of apoptosis. Our studies provide 567
a rationale for using PARK2 as a potential biomarker to predict chemosensitivity of 568
27
antimicrotubule drugs-containing regimen. 569
570
Abbreviations 571
DFS: disease-free survival; pCR: pathologic complete response; ChIP: chromatin 572
immunoprecipitation; Co-IP: coimmunoprecipitation; TF: transcription factors; Δψm: 573
mitochondrial membrane potential; PPIN: protein-protein interaction networks; 574
575
Acknowledgements 576
This study was supported by Natural Science Foundation of China (81872140, 81572484, 577
81420108026 and 81621004 to DY, 81301732, 81772821 to KH, 81972466, 81402199 to 578
WW, 81903069 to YL); Guangdong Science and Technology Department (2019B020226003 579
to DY); Guangzhou Bureau of Science and Information Technology (201704030036 to DY); 580
Guangdong Natural Science Foundation (2017A030313471 to KH). 581
582
Author Contributions 583
Hengxing Chen and Yun Li performed all the experiments, performed the statistical analysis 584
and wrote the manuscript. Yu Li, Zhen Chen, Limin Xie, Wenjia Li, Yuanxin Zhu, Daning Lu 585
and Wenjing Wu analyzed the data and assisted in the experiment. H. Phillip Koeffler 586
reviewed and edited the manuscript. Kaishun Hu and Dong Yin conceived the study, 587
participated in the study design, funded the study and participated in manuscript revision. 588
589
Competing Interests 590
28
The authors declare that they have no conflict of interest. 591
592
References: 593
1. Iwata H, Sato N, Masuda N, Nakamura S, Yamamoto N, Kuroi K, et al. Docetaxel followed by 594
fluorouracil/epirubicin/cyclophosphamide as neoadjuvant chemotherapy for patients with primary 595
breast cancer. Jpn J Clin Oncol. 2011;41(7):867-75. 596
2. Dieras V, Fumoleau P, Romieu G, Tubiana-Hulin M, Namer M, Mauriac L, et al. Randomized 597
parallel study of doxorubicin plus paclitaxel and doxorubicin plus cyclophosphamide as neoadjuvant 598
treatment of patients with breast cancer. J Clin Oncol. 2004;22(24):4958-65. 599
3. Nowak AK, Wilcken NR, Stockler MR, Hamilton A, Ghersi D. Systematic review of 600
taxane-containing versus non-taxane-containing regimens for adjuvant and neoadjuvant treatment of 601
early breast cancer. Lancet Oncol. 2004;5(6):372-80. 602
4. Jordan MA, Wilson L. Microtubules as a target for anticancer drugs. Nature Reviews Cancer. 603
2004;4(4):253-65. 604
5. Bhalla KN. Microtubule-targeted anticancer agents and apoptosis. Oncogene. 605
2003;22(56):9075-86. 606
6. Zhai Y, Kronebusch PJ, Simon PM, Borisy GG. Microtubule dynamics at the G2/M transition: 607
abrupt breakdown of cytoplasmic microtubules at nuclear envelope breakdown and implications for 608
spindle morphogenesis. J Cell Biol. 1996;135(1):201-14. 609
7. Radha G, Raghavan SC. BCL2: A promising cancer therapeutic target. Biochim Biophys Acta Rev 610
Cancer. 2017;1868(1):309-14. 611
8. Yu Q, Qiu Y, Chen X, Wang X, Mei L, Wu H, et al. Chemotherapy priming of the Pancreatic 612
29
Tumor Microenvironment Promotes Delivery and Anti-Metastasis Efficacy of Intravenous 613
Low-Molecular-Weight Heparin-Coated Lipid-siRNA Complex. Theranostics. 2019;9(2):355-68. 614
9. Li Y, Zhang B, Xiang L, Xia S, Kucuk O, Deng X, et al. TGF-beta causes Docetaxel resistance in 615
Prostate Cancer via the induction of Bcl-2 by acetylated KLF5 and Protein Stabilization. Theranostics. 616
2020;10(17):7656-70. 617
10. Deng X, Gao F, Flagg T, May WS, Jr. Mono- and multisite phosphorylation enhances Bcl2's 618
antiapoptotic function and inhibition of cell cycle entry functions. Proc Natl Acad Sci U S A. 619
2004;101(1):153-8. 620
11. Deng X, Xiao L, Lang W, Gao F, Ruvolo P, May WS, Jr. Novel role for JNK as a stress-activated 621
Bcl2 kinase. J Biol Chem. 2001;276(26):23681-8. 622
12. Ito T, Deng X, Carr B, May WS. Bcl-2 phosphorylation required for anti-apoptosis function. J 623
Biol Chem. 1997;272(18):11671-3. 624
13. Kitada T, Asakawa S, Hattori N, Matsumine H, Yamamura Y, Minoshima S, et al. Mutations in the 625
parkin gene cause autosomal recessive juvenile parkinsonism. Nature. 1998;392(6676):605-8. 626
14. Park MH, Lee HJ, Lee HL, Son DJ, Ju JH, Hyun BK, et al. Parkin Knockout Inhibits Neuronal 627
Development via Regulation of Proteasomal Degradation of p21. Theranostics. 2017;7(7):2033-45. 628
15. Beroukhim R, Mermel CH, Porter D, Wei G, Raychaudhuri S, Donovan J, et al. The landscape of 629
somatic copy-number alteration across human cancers. Nature. 2010;463(7283):899-905. 630
16. Veeriah S, Taylor BS, Meng S, Fang F, Yilmaz E, Vivanco I, et al. Somatic mutations of the 631
Parkinson's disease-associated gene PARK2 in glioblastoma and other human malignancies. Nat Genet. 632
2010;42(1):77-82. 633
17. Gong Y, Zack TI, Morris LG, Lin K, Hukkelhoven E, Raheja R, et al. Pan-cancer genetic analysis 634
30
identifies PARK2 as a master regulator of G1/S cyclins. Nat Genet. 2014;46(6):588-94. 635
18. Lee SB, Kim JJ, Nam HJ, Gao B, Yin P, Qin B, et al. Parkin Regulates Mitosis and Genomic 636
Stability through Cdc20/Cdh1. Mol Cell. 2015;60(1):21-34. 637
19. Dawson TM, Dawson VL. The role of parkin in familial and sporadic Parkinson's disease. Mov 638
Disord. 2010;25 Suppl 1:S32-9. 639
20. Carroll RG, Hollville E, Martin SJ. Parkin sensitizes toward apoptosis induced by mitochondrial 640
depolarization through promoting degradation of Mcl-1. Cell Rep. 2014;9(4):1538-53. 641
21. Hong J, Hu K, Yuan Y, Sang Y, Bu Q, Chen G, et al. CHK1 targets spleen tyrosine kinase (L) for 642
proteolysis in hepatocellular carcinoma. J Clin Invest. 2012;122(6):2165-75. 643
22. Yuan J, Jiang YY, Mayakonda A, Huang M, Ding LW, Lin H, et al. Super-Enhancers Promote 644
Transcriptional Dysregulation in Nasopharyngeal Carcinoma. Cancer Res. 2017;77(23):6614-26. 645
23. Walker SR, Chaudhury M, Nelson EA, Frank DA. Microtubule-targeted chemotherapeutic agents 646
inhibit signal transducer and activator of transcription 3 (STAT3) signaling. Mol Pharmacol. 647
2010;78(5):903-8. 648
24. Zhang L, Xu X, Yang R, Chen J, Wang S, Yang J, et al. Paclitaxel attenuates renal interstitial 649
fibroblast activation and interstitial fibrosis by inhibiting STAT3 signaling. Drug Des Devel Ther. 650
2015;9:2139-48. 651
25. Zhang X, Wu X, Zhang F, Mo S, Lu Y, Wei W, et al. Paclitaxel induces apoptosis of esophageal 652
squamous cell carcinoma cells by downregulating STAT3 phosphorylation at Ser727. Oncol Rep. 653
2017;37(4):2237-44. 654
26. Wang LG, Liu XM, Kreis W, Budman DR. The effect of antimicrotubule agents on signal 655
transduction pathways of apoptosis: a review. Cancer Chemother Pharmacol. 1999;44(5):355-61. 656
31
27. Rosse T, Olivier R, Monney L, Rager M, Conus S, Fellay I, et al. Bcl-2 prolongs cell survival 657
after Bax-induced release of cytochrome c. Nature. 1998;391(6666):496-9. 658
28. Hsu YT, Youle RJ. Nonionic detergents induce dimerization among members of the Bcl-2 family. 659
J Biol Chem. 1997;272(21):13829-34. 660
29. Tait SW, Green DR. Mitochondria and cell death: outer membrane permeabilization and beyond. 661
Nat Rev Mol Cell Biol. 2010;11(9):621-32. 662
30. Zhou Z, Cheng Y, Jiang Y, Liu S, Zhang M, Liu J, et al. Ten hub genes associated with 663
progression and prognosis of pancreatic carcinoma identified by co-expression analysis. Int J Biol Sci. 664
2018;14(2):124-36. 665
31. Chin CH, Chen SH, Wu HH, Ho CW, Ko MT, Lin CY. cytoHubba: identifying hub objects and 666
sub-networks from complex interactome. BMC Syst Biol. 2014;8 Suppl 4:S11. 667
32. Blagosklonny MV, Giannakakou P, el-Deiry WS, Kingston DG, Higgs PI, Neckers L, et al. 668
Raf-1/bcl-2 phosphorylation: a step from microtubule damage to cell death. Cancer Res. 669
1997;57(1):130-5. 670
33. Haldar S, Basu A, Croce CM. Bcl2 is the guardian of microtubule integrity. Cancer Res. 671
1997;57(2):229-33. 672
34. Trempe JF, Sauve V, Grenier K, Seirafi M, Tang MY, Menade M, et al. Structure of parkin reveals 673
mechanisms for ubiquitin ligase activation. Science. 2013;340(6139):1451-5. 674
35. Wei Y, Pattingre S, Sinha S, Bassik M, Levine B. JNK1-mediated phosphorylation of Bcl-2 675
regulates starvation-induced autophagy. Mol Cell. 2008;30(6):678-88. 676
677
A Chr6
PARK2一「··皿皿皿U幽也L且Li」qter
4←－Samples一一一＃
c
Deleted Neutral Amplified
Total cases (Not receive chem。therapy+ receive chemotherapy)
100
e v 户UH 2 k mm p - ’
’’’’？。”
’
D00000 08642 1 百ZE3ω
＠E吁＠ω四”ω旦口
50 100 Time (months)
00 86 百ZE3的
e V AY C UH K M Dr
00 42 －E＠〉口 －」 PARK2 IHC -ve
150 。。 100 50
E
Receive chemotherapy
Time (months)
百 100>
·~ 80 il ’ 60 { I P=0.0027 ! 40
~ 20 －」 PARK2 IHC +ve
－』 PARK2 IHC -ve
50 100 Time (months)
100
.~ 80
~ 60
阳D「A崎
q’－
一－mw』ω〉。
－」 PARK2 IHC +ve
－』 PARK2 IHC -•e
0 150 0 50 100
Time (m。nths)
F
Antimicrotubule drugs-containing regimen
’00 百 100>
-~ 80 il 串 60 ,I; 目 4。
” ω -~ 20. 口 I －」 PARK2 IHC ve
。。 50 100
Time （阿10nths)
DNA-damaging drugs regimen (with。ut antimicrotubule drugs)
100 百 100
-~ 80 iil : 60 斗
!: 白
50 100 Time (months)
-~ 80 主iil 60
; 40 6 I 占 PARK21HC明
20 －』 PARK2 IHC -ve
0 150 0 50 100
Time (months)
吉 80 -~ iil 60
; 40 占 I 十 PARK2 1HC咄
20 －』 PARK2 IHC -ve
0 150 0 50 100
Time (months)
B PARK2 IHC +ve
(111/205)
响MKY刀XN
D
PARK2 IHC -ve
(95/205)
Not receive chemotherapy
100
回mm
+
i
C
C
HH 2
2 刚刚
AA PP -- J
00000 08
64
2
4, 百EE3的。E吁＠阶mw＠E口
150 50 100 Time (months)
G
吉 80 ~ iil 60
里 40
- 20
P=0.7477
－」 PARK2 IHC +ve
－」 PARK2 IHC -ve
150 50 100 Time (months)
150
Antimicrotubule drugs-containing regimen
pCR non-pCR
150 :
Antimicrotubule drugs-containing PARK2 Expression in Gluck Breast reaimen (Grouped by Capecitabine/Docetaxel
1:: T i； 自 ；； 十1so I j: t: ~
-3
IN ~
军 l:'>i坠，
H
~~"' ~~ 。~＂，v _,/1 f;、-.:,-- F ，。
可
150
-< -< "·- ~~· j寺命 ~<f:,.._ ~~－品’＿.，~~φ、
、 手电
1
Figure Legends
Figure 1. PARK2 predicts clinical outcome of antimicrotubule drugs-containing regimen. A.
IGV plots showed that PARK2 gene (located on chromosome 6) is often deleted in breast cancer
(436 of 1,283 breast cancer samples). B. Representative IHC staining intensities of PARK2
protein expression in breast cancer tissues. C. Higher PARK2 expression was associated with
better DFS and overall survival of breast cancer patients. High expression (n=111), Low
expression (n=94). D. Kaplan-Meier estimates of DFS and overall survival of the breast cancer
who did not receive chemotherapy. High expression (n=31), Low expression (n=31). E. Kaplan-
Meier estimates of DFS and overall survival of the breast cancers who received chemotherapy.
High expression (n=80), Low expression (n=63). F. Above graph: Kaplan-Meier estimates of
DFS and overall survival of the breast cancer who received antimicrotubule drugs-containing
regimen. High expression (n=42), Low expression (n=41). Below graph. Kaplan-Meier
estimates of DFS and overall survival of the breast cancer who received DNA-damaging drugs
regimen (without antimicrotubule drugs). High expression (n=38), Low expression (n=22). G.
Representative images from immunohistochemical staining of PARK2 in tumors from four
cases of breast cancer who receive antimicrotubule drugs with different pathologic complete
response (pCR). H. Expression levels of PARK2 in breast cancer patients, who received
antimicrotubule drugs-containing regimen. I. Upregulation of PARK2 mRNA expression in the
Capecitabine/Docetaxel responding group compared with the non-responding group.
80
. 直 60帽
>亏 40
<.> I ... Vector 20,
I -<’” PARK2 0
80
ε 喜 60>: 40
u l 司’ Vector20,
l 吃，』 PARK2
0
MCF-7
• Untreated
• Vinorelbine
• Docetaxel
D
80
43
c6o 95 _ @ ω
号 40
!2。”
43
32 F」
MCF-7 v p v p
++
c Vinorelbine
PARK2
PARP
MCF-7
100
MCF-7
100
A
<:, e舍、
，品，晶守矿
PARK2 cDNA(ng)
、
Caspase-9
CDC-27-P--1’ CDC-27-t
5() 。， 0.5 1 5 10 Vinorelbine (nM)
50 ，。。， 0.5 1 5 Docetaxel (nM)
95
MDA-MB-134-VI
飞、... shNC 4护 shPARK2 1 4 ’ shPARK2-2
o I 0 0.1 0.5 2.5 12.5 62.5 3’2.5
Vlnorelblne (nM)
’00
80
~ 'is 60
” >
吉 40
20
34
嗡－ shNC 毛护 shPARK2-1
4 • shPARK2-2 0←一←一←「0 0.1 0.5 2:5 12.5 62.5 312.5
Docetaxel (nM)
MDA-MB-134-VI
GAP DH
E
80
居豆 60 >
亏 40也J
20
MDA-MB-231
100
80
居'is 6。
” >: 40 u I ... Vector
20, I -<’- PARK2
0
MDA-MB-231
100
80
z. 罩 60 -~ >亏 40
'-' I ... Vect。r20"
I -<’- PARK2 0
：广 二14辙 J
，军苔
，空
102
Cl.
F
Untreated
Docetaxel
Vin。relbine
50 。， 0.5 1 5 10 Vinorelbine (nM)
50 ，。。， 0.5 1 5 Docetaxel (nM)
Docetaxel
Untreated
Vinorelbine
B
、旬,
吨，。。。
向,. .. -咯a.o••· ~, .. -
MDA-MB-231
/ 飞y,•
0 ~,<'
守护句。。。
50. I• Vector
!:40」 - PAR阳ω － ‘’ u 30 + >5 2。” ω 2 10 <
0 ,..,, ”’ _.,, ..
v"°'
MCF-7
。价。, .,,.,<'
、’P , 。。。。
60. I• Vector
,O '
己 I - PARK2 .!!? . 百 40ω
+ >巳
'iii 20 巳c ‘z
0 .I>
‘’ 。
4’ 心。
2
Figure 2. PARK2 regulates antimicrotubule drugs sensitivity in breast cancer cells. A.
Cytotoxicity assays showed the sensitivity to docetaxel and vinorelbine. Breast cancer lines
stably expressing either ectopic wild-type PARK2 or vector were treated with the indicated
compounds for 48 h. B. PARK2 overexpression increases apoptosis of breast cancer cells after
either docetaxel or vinorelbine treatment for 48 h. Representative FACS profiles are shown, on
which the population of cells in the quadrant (Annexin V± DAPI) represents apoptotic cells. C.
MCF-7 cells stably expressing either vector control (V) or PARK2 (P) were treated with
vinorelbine (1 nM) for 48 h. Cell lysates were immunoblotted for caspases and caspase
substrates. Phosphorylation of CDC27 is an indicator of mitotic arrest and we used it to indicate
effectiveness of antimicrotubule drugs. CDC27-P, phosphorylated CDC27. D. MCF-7 cells
were transfected with empty vector along with the indicated amount of PARK2 cDNA for 18 h.
After transfection, cells were either treated or not with either docetaxel (1 nM) or vinorelbine
(1 nM) for 36 h; cell death was analyzed by flow cytometry. E. Cytotoxicity assay measuring
the sensitivity to docetaxel and vinorelbine. ShNC and shPARK2 MDA-MB-134-VI cells
treated with the indicated compounds for 48 h. F. Silencing PARK2 (shPARK2) decreased cell
apoptosis of breast cancer cells after either docetaxel or vinorelbine treatment for 48 h. Data
show mean ± s.d. N = 3. *P < 0.05, **P < 0.01, ***P < 0.001
3
Figure 3. Antimicrotubule drugs stimulate PARK2 expression. A&B. Q-PCR and western blot
confirmed that the mRNA and protein expression of PARK2 after docetaxel treatment. C.
Prediction of TFs of PARK2 by overlaying ChIP-seq data from UCSC, Chip-Atlas and
PROMO. D. Chip-seq profiles for STAT3 at PARK2 locus in breast cancer cell lines. E.
Docetaxel treatment reduces enrichment of STAT3 on PARK2 promoter. F. Luciferase assay
showing that STAT3 knockdown increased PARK2 promoter activity in MDA-MB-134-VI
cells. G. Docetaxel induced downregulation of STAT3 and upregulation PARK2 expression,
whereas silencing STAT3 increased PARK2 levels in both breast cancer cell lines. Data show
mean ± s.d. N = 3. *P < 0.05, **P < 0.01
4
Figure 4. PARK2 causes apoptosis by activating mitochondrial pathway induced by
antimicrotubule drugs. A. MCF-7 cells stably expressing either ectopic wild-type PARK2 or
vector (control) were cultures either with or without docetaxel (1 nM) for 24 h. MCF-7 cells
were co-immunostained with TOMM20 antibody (red; anti-TOMM20), activated form of BAX
antibody (green; anti-BAX 6A7) and nuclei (blue; DAPI). Cells were analyzed by confocal
microscopy. Scale bars represent 5 µm. B. MCF-7 cells stably expressing either ectopic wild-
type PARK2 or vector control were treated with docetaxel (1 nM) for 0, 12, 24 and 36 h. After
docetaxel stimulation, mitochondria were fractionated. In order to crosslink the BAX
oligomerized proteins, mitochondria fractions were treated with a sulph-hydryl-reactive
crosslinker [bis-maleimidohexane (BMH)]. Samples were separated on a 4-12% gradient
NuPAGE gel. Blot were probed with anti-BAX, anti-CoxII and anti-β-tubulin. C.
Overexpression of PARK2 reduced mitochondrial membrane potential after docetaxel
treatment. Representative FACS spectrum is shown in which the population of cells in the Q3
quadrant（right, bottom）represented the decrease of mitochondrial membrane potential. D.
PARK2 promoted the release of cytochrome C from mitochondria to the cytosol. Isolation of
cytosol and mitochondrial fractions and detection with anti-cytochrome C. For internal control,
anti-Cox IV or anti-α-tubulin was used to probe the mitochondrial and cytoplasmic fractions,
respectively. E. MCF-7 cells stably expressing either ectopic wild-type PARK2 or vector
(control) were treated either with or without docetaxel (1 nM) for 48 h. Cell lysates were
immunoblotted for caspases and caspase substrates.
5
Figure 5. Antimicrotubule drug promotes interaction between PARK2 and BCL-2. A.
Proteomics workflow diagram. (MCF-7-SFB-PARK2: MCF-2 stably expressing SFB tag
PARK2). B. Heatmap cluster of proteomic changes with binding to PARK2 before and after
docetaxel treatment in MCF-7-PARK2 cell line. C. Gene Ontology (GO) analysis of increased
proteins after docetaxel treatment. D. PPIN analysis of apoptosis associated proteins to
search hub proteins. (Red nodes represent hub proteins, blue nodes represent non-important
proteins, and line represent the relationship between two proteins). E. Fold change in protein
level of 6 selected hub proteins. Illustrated is the fold change in protein level of post-docetaxel
treatment compared to pre-docetaxel treatment. F. Volcanic map of proteomic changes with
binding to PARK2 either before or after docetaxel treatment. Vertical dashed lines indicate cut-
off of log2Fold change (1 or –1). Horizontal dashed lines indicate the cut-off of p-value (0.05).
G. MCF-7 cells stably expressing either ectopic wild-type PARK2 or vector (control) were
treated with docetaxel (1 nM) for 24 h. Cells were lysed with RIPA buffer followed by
immunoprecipitation (IP) using anti-FLAG agarose and western blot with indicated antibody.
H&I. MDA-MB-134-VI cells were treated with either docetaxel or vinorelbine for 24 h. After
IP assays were performed with endogenous PARK2, BCL-2 or phospho-BCL-2 (Ser70) in
MDA-MB-134-VI cells. J. Association of Flag–PARK2 with Myc–BCL-2 mutants S70A.
6
Figure 6. PARK2 promotes poly-ubiquitination and degradation of BCL-2. A. Boxplots of
RPPA data showed a negative correlation between the mRNA levels of PARK2 and protein
levels of BCL-2. B. MCF-7 cells stably expressing either ectopic wild-type PARK2 (P) or
vector control (V) were treated with docetaxel (1 nM) for 24 h. C. MDA-MB-134-VI cells
stably deleting either endogenous PARK2 or shNC control were treated with docetaxel (2 nM)
for 24 h. B-C. Cell lysates were immunoblotted for BCL-2, phospho-BCL-2 (Ser70), BAX,
PARK2, and GAPDH. D. MCF-7 cells stably expressing either ectopic wild-type PARK2 or
vector (control) were treated with docetaxel (1 nM) for 24 h and then incubated with 10 µg/ml
cycloheximide (CHX) for the indicated periods of time and then immunoblotted (Left panel).
Quantification of BCL-2 protein levels, n = 3. Error bars indicate SEM. Relative level of BCL-
2 protein is normalized to 0 h (Right panel). E. MCF-7 cells stably expressing either ectopic
wild-type PARK2 or vector control were treated either with or without docetaxel (1 nM) for 24
h. These cells were then cultured in the presence of either DMSO control or MG132 (10 mM)
(6 h) followed by western blot analysis of BCL-2 levels. F. MCF-7 cells were transfected with
indicated vectors for 18 h. Cells were then treated with or without docetaxel (1 nM) for 24 h.
After docetaxel treatment, Cells were treated with MG132 (10 mM) and subjected to
immunoprecipitation (IP) analysis with BCL-2 antibody. HA-Ub KO mutant (incapable of
forming polyubiquitin chain) was used as a negative control. G. MDA-MB-134-VI cells stably
expressing indicated shRNAs were transfected with indicated vectors for 24hs. Cells were then
treated with MG132 (10 mM) either in the presence or absence of docetaxel (1 nM) for 36 h
and subjected to immunoprecipitation (IP) analysis with BCL-2 antibody.
7
Figure 7. PARK2 loss-of-function mutants fail to enhance antimicrotubule drugs sensitivity A.
MCF-7 cells stably expressing either ectopic wild-type PARK2 (P) or vector control (V) were
treated with docetaxel (1 nM) for 24 h. B. MCF-7 cells stably expressing either ectopic wild-
type PARK2 (P) or vector control (V) were transfected with control siRNA or siRNA targeted
against BCL-2 for 56 h. After transfection for 24 h, cells were treated with docetaxel for 48h.
C. MCF-7-PARK2 cells were transfected with control siRNA or siRNA targeted against the
indicated BCL-2 family member for 50 h. Knockdown efficiency was analyzed by western
blotting (Supplementary Figure. S5A). Cells were treated with docetaxel for 48h. D. Pan-cancer
analysis of PARK2 loss-of-function mutations. E. MCF-7 cells (Vector (control), PARK2 wild
type (WT), T173, T240M, P294S, and C431S clones) were transfected with indicated plasmid
for 18 h. Cells were then treated either without (left panel) or with (right panel) docetaxel (1
nM) for 24 h. MG132 was added 6 h prior to harvest cell lysate. Cells were lysed with RIPA
buffer followed by immunoprecipitation (IP) using anti-FLAG agarose and western blot was
done with indicated antibody. F. Overexpression of wild-type but not loss-of-function mutants
of PARK2 led to lower levels of BCL-2 protein. Cyclin D1 and Cyclin B1 protein was used as
a positive control. G. Overexpression of wild-type but not loss-of-function PARK2 mutants
increased apoptosis in breast cancer cells after docetaxel (1 nM) treatment. Data show mean ±
s.d. N = 3. **P < 0.01.
8
Figure 8. PARK2 enhances antimicrotubule drug sensitivity in vivo. A. Experimental schematic.
A-G. BALB/c nude mice were injected with MCF-7 cells which were stably transfected with
doxycycline-inducible, fluorescently-labeled constructs (Vector, PARK2-WT and PARK2-
T240M) in the breast orthotopic model. When tumors reached 100 mm3, doxycycline was
added in the mice drinking water, and docetaxel (10 mg/kg) was injected through tail vein at
the same time. B. Fluorescence images of tumor in BALB/c nude mice injected with
fluorescently-labeled MCF-7 cells overexpressing vector (control), wild-type PARK2 or
PARK2-T240M. Fluorescence intensity of wild-type PARK2 tumors were too low to be
detected. C. Images of resected tumors at the end point (day 42). D-E. When tumors reached
100 mm3, tumor volumes at the indicated times were measured. Weights of resected tumors at
the end point (day 42) (n=6). F. Representative immunofluorescent images for apoptotic cells
were examined by TUNEL staining in tumor tissues. Scale bar represents 20μm. And
quantification of apoptotic cells for each group (n = 6). G. Expression of BCL-2, BAX (6A7),
STAT3 and PARK2 were detected by Western blot in protein lysates of tumor tissue. H. Model
illustrates that following antimicrotubule drugs treatment, the expression of PARK2 was
upregulated due to the repression of STAT3-mediated transcriptional inhibition of PARK2.
PARK2 promotes degradation of phospho-BCL-2 following antimicrotubule drugs treatment,
which activates the mitochondrial apoptotic pathway. Data show mean ± s.d. N = 6, ***P <
0.001.