PARK2 Depletion Connects Energy and Oxidative Stress to PI3K/Akt Activation via PTEN S-Nitrosylation

1Cell homeostasis is achieved through an orchestrated balance of cell signaling interactions that dictate the likelihood of a cell to escape from normal growth restraints or be eliminated from the replicative pool, resulting in cell death. Cancer is a group of diseases that are due to escape from cell death control, while Parkinson’s disease (PD) portrays a disease that results from accelerated cell death. It would seem unlikely that these diseases are related, yet there is increasing evidence to suggest that a subset of PD susceptibility genes are also associated with cancer.

2PARK2 was originally identified as a gene associated with the pathogenesis of familial PD in early-onset autosomal recessive juvenile parkinsonism. It has been reported to be mutated in as high as 77% of PD patients with an age of onset of <20 years, but only in 3% of patients with an age of onset of >30 years. Subsequently, it has been linked to a wide range of disorders, including leprosy, autism, type 2 diabetes mellitus, Alzheimer’s disease, cerebellar ataxia, resistance to intracellular pathogen infections, and cancer, where it is somatically deleted in a wide spectrum of tumor types. PARK2 is a bona fide haploinsufficient tumor suppressor, as depletion of a single PARK2 allele significantly increases adenoma development and polyp multiplicity in ApcMin/+ mice. PARK2 loss also renders mice more susceptible to hepatocellular and γ-irradiation-induced carcinomas, while ectopic PARK2 expression mitigates cell proliferation in colorectal, glioma, breast, hepatocellular, and non-small-cell lung cancer cells.

3The PARK2 gene encodes the E3 ubiquitin ligase Parkin, which mediates the ubiquitination of a number of substrate proteins, leading to their proteasomal degradation. Its activities go beyond the degradative ubiquitination, and it is implicated in the regulation of multiple cellular processes, including stress response, mitochondrial biogenesis, and stability of G1/S cyclins. Although the underlying mechanisms by which pathogenic PARK2 mutations contribute to PD are not entirely understood, mitochondrial dysfunction is considered to play a central role in stress-induced neuronal cell death associated with the pathogenesis of this disorder. Increased oxidative and nitrosative stress is a common phenomenon in both PD and cancer; hence, it is imperative to identify the molecular pathways underlying the functional contribution of PARK2 depletion in these processes.

4Compelling evidence shows that cancer cells utilize multiple pathways, including the phosphatidylinositol 3-kinase/Akt (PI3K/Akt) signaling pathway, to enhance their survival and prevent apoptosis under metabolic stress conditions. Importantly, PARK2 has previously been associated with the activation of the Akt pathway; however, the mechanistic evidence behind its functional contribution is unclear. One study showed that Parkin interacts with and ubiquitinates Eps15 to delay the internalization and degradation of its adaptor protein epidermal growth factor receptor (EGFR), thereby promoting PI3K/Akt signaling, while a more recent study suggests that Parkin directly interacts with and promotes the ubiquitination of EGFR, leading to diminished activation of EGF-induced PI3K/Akt signaling. These data highlight the need for further investigation of the molecular events underlying the role of PARK2 depletion in PI3K/Akt-mediated cellular survival.

5In this study, we identified PTEN as an important mediator behind the functional contribution of PARK2 depletion in the activation of the PI3K/Akt pathway, and we further characterized its pivotal role in the tumor suppressor function of PARK2 in vitro and in vivo. Additionally, our results reveal an important missing piece in the dynamic signaling and metabolic network connecting AMPK with Akt activation in the absence of mTORC1-S6K-dependent negative feedback loop mechanisms, demonstrating a compensatory survival mechanism for cancer cells under conditions of energy deprivation.

6We examined the degree of PARK2 deletion in the largest up-to-date collection of The Cancer Genome Atlas (TCGA), assembling data from 9,863 primary tumors from 28 different tumor types (Table S1). Focal deletions (Figure 1A, dark blue) of the PARK2 gene were most commonly found in colorectal (21%) and ovarian (25%) carcinomas, while a strikingly high number of tumors, including lung adenocarcinomas, melanomas, bladder, ovarian, and pancreatic, had an overall >40% DNA copy number (DCN) loss of the PARK2 gene (both focal deletion [Figure 1A, dark blue] and as part of whole or part chromosome arm losses [Figure 1A, light blue]). Cholangiosarcomas (CHO) showed a staggering 69% (25/36) rate of overall PARK2 deletions. In parallel, we also examined the degree of PARK2 mRNA decrease between cancer and corresponding normal tissues from a total number of 13,481 specimens, studied in 127 published microarray datasets (Table S2), covering 25 tumor types. Notably, there were many cancer types, including gliomas, cervical, and kidney carcinomas, with not the highest frequency of overall PARK2 deletion (<40%) but with widespread (up to 69%) decrease or loss of its mRNA expression compared to their normal counterparts (Figures 1B and S1A; >1-log2-fold [red] or >0.5-log2-fold [red and yellow]). Vice versa, there were some tumor types, e.g., ovarian cancer, where PARK2 deletion was found in >60% of the cases, yet only a small fraction were reported to have low PARK2 mRNA expression. This reflects in part the stringent criteria we used to report tumor types with >0.5-fold changes in mRNA expression compared to their normal counterparts, but also that the majority (>90%) of PARK2 deletions across all tumor types are heterozygous and, hence, less likely to cause a marked reduction in PARK2 mRNA expression.

7As high as two-thirds of glioma tumors had significantly reduced levels of PARK2 mRNA expression compared to their corresponding normal tissues (Figure 1B). We plotted the distribution of PARK2 expression, across different subgroups of gliomas reminiscent of distinct neural cell types, to show that PARK2 is significantly downregulated irrespective of the histological origin of this tumor type (Figure S1B). To identify relationships between PARK2 DCN and mRNA expression levels, we plotted the distribution of its expression across subgroups of glioblastomas (GBMs) with different PARK2 DCN. PARK2 mRNA expression was progressively lower across the different subgroups of PARK2 ploidy status (Figure 1C). To assess the prognostic significance of this gene across different types of genetic alteration, we also performed survival analysis between subgroups of patients with GBM, divided based on different molecular markers of the PARK2 gene. Interestingly, loss of PARK2 at the DNA, mRNA, and protein levels (Figure 1D) all correlated with significantly poorer survival in patients with GBM (Figure 1E). Of note, PARK2 expression was also associated with poorer survival in patients with breast and lung adenocarcinomas (Figures S1C and S1D). Collectively, our data demonstrate that PARK2 is altered in over a third of all human cancers. Its widespread loss or decrease across many molecular biomarker indicators (DNA, mRNA, and protein) significantly correlates with poorer survival, offering great prognostic and predictive value for clinical practice.

8The PI3K/Akt pathway is the single most frequently altered signaling cascade across all the tumor types, with recurrent PARK2 deletions and mRNA loss/reduction. Activation of the PI3K/AKT/mTOR pathway has also been associated with significantly poorer survival across many solid tumors. To investigate whether activation of this pathway is inherently associated with loss of the PARK2 gene, we stably expressed two independent hairpins targeted against PARK2 in HCT116 cells. Cells with PARK2 knockdown exhibited a pronounced increase in Akt phosphorylation compared to control GFP knockdown cells (Figure 2A). Staurosporine is a protein kinase inhibitor that can induce apoptosis across many different cell types, and Akt hyperactivation has been shown to attenuate sensitivity to staurosporine-induced cell death. We showed that PARK2 depletion augmented resistance to staurosporine-induced cell death (Figure 2B), and this was consistent with significantly lower apoptotic response, as indicated by caspase-3/7 activity (Figure 2C).

9Identification of candidate biomarkers that predict responsiveness to specific signaling cascade inhibitors is increasingly important in the era of personalized medicine. Accordingly, we assessed the role of PARK2 knockdown to the inhibition of various signaling nodes across the PI3K/Akt/mTOR and MAPK pathways. PARK2 depletion rendered cells more sensitive to inhibitors of PI3K (BKM120 and BEZ235), Akt (MK2206), and mTOR (BEZ235, Rapamycin, and Torin), but it decreased the effectiveness of MEK inhibitors PD0325901 and GSK1120212 (Figures 2D, 2E, S2A, and S2B). Taken together, these data suggest that PARK2 loss contributes in the activation of Akt signaling and the dependence of cells on this pathway, suggesting that PARK2 deficiency could be used as a biomarker of efficacy and favorable clinical response to inhibitors of the PI3K/Akt/mTOR pathway.

10To explore the significance of Akt activation in the molecular mechanisms that PARK2 may contribute to tumor suppression, we tested the effect of PARK2 overexpression in PTEN wild-type (WT) (HCT116 PTEN+/+ and H1299) and PTEN mutant (PC3 and HCT116 PTEN−/−) cell line models. Ectopic expression of PARK2 led to the inhibition of Akt phosphorylation in HCT116 cells (Figure 3A). Importantly, we showed that PARK2 overexpression led to the inhibition of Akt activation upon growth factor stimulation in PTEN WT H1299 cells (Figure 3B), but not in the PTEN mutant PC3 cells (Figure 3C). This reduction in Akt phosphorylation was also apparent when we ectopically expressed PARK2 at levels near the endogenous Parkin levels expressed in H460 cells, for both the EGFR WT (H1299) and EGFR null (SW620) cells (Figures S2C and S2D). Unlike in parental PTEN+/+ HCT116 cells, overexpression of PARK2 did not reduce Akt phosphorylation in isogenic PTEN−/− HCT116 cells (Figure S2E), further supporting a role for PTEN in PARK2-mediated regulation of PI3K/Akt signaling.

11Akt has been shown to play a central role in promoting growth factor-mediated cell proliferation and migration. Given that PARK2 inhibits Akt phosphorylation, we sought to investigate its functional contribution in suppressing some of the cellular functions mediated by Akt activation. HCT116 PTEN+/+ and H1299 cells exhibited a pronounced reduction in both cell proliferation (Figure 3D) and cell migration (Figure 3E) upon PARK2 overexpression. However, this was not true for the PTEN mutant PC3 and isogenic PTEN−/− HCT116 cells that showed no effect compared to control empty vector (EV)-expressing cells. No effect on cell proliferation and migration was also observed when overexpressing the E3-Ligase dead C431S Parkin mutant in both PTEN WT and mutant cells (Figures S2F–S2H).

12To investigate the tumor suppressor effect of PARK2 in vivo, we retro-orbitally injected luciferase-labeled H1299 cells overexpressing EV or human PARK2. Ectopic expression of PARK2 significantly mitigated the formation of lung metastases (Figures 3F–3I) in NOD/SCID mice. The luciferase activity in cells expressing the PARK2 gene was reduced on both weeks 3 and 4 post-injection (Figures 3F and 3G), and this was consistent with significantly lower number and smaller size of lung metastases, as histopathologically evaluated with H&E-stained sections (Figures 3H and 3I). These data show that suppression of Akt activation plays a central role in the tumor suppressor function of PARK2 and that the mutational status of PTEN determines the functional contribution of PARK2 loss in Akt-mediated cell proliferation and migration.

13To further investigate the role of PTEN in PARK2-mediated regulation of the PI3K/Akt pathway, we performed PARK2 knockdown analysis in cell lines with abundant Parkin but either WT PTEN present (HCT116 and H460) or absent (PC3 and U138). In line with our overexpression studies, PARK2 knockdown induced Akt phosphorylation in HCT116 and H460 cells, but not in PC3 or U138 cells (Figures 4A, S3A, and S3B) or in the isogenic PTEN null HCT116 (Figure 4B) and MCF10A cells (Figure S3C). Most importantly, PARK2 depletion led to a >30% reduction in PTEN protein levels across all PTEN WT cell lines (Figures 4A, 4B, and S3C), and this observation was confirmed in CRISPR/Cas9-mediated PARK2 knockout H460 cells (Figure S3D). Previous studies with hypomorphic Pten allelic series of mice have shown that even subtle reductions in Pten protein levels can have dramatic consequences in cancer progression. PTEN is regulated in cancer at the transcriptional, post-transcriptional, and post-translational levels.

14To gain further insight into this regulation by PARK2, we performed qPCR analysis of both PARK2 and PTEN following lentiviral knockdown of PARK2, and we showed that PARK2 depletion does not affect PTEN mRNA levels (Figures 4C and S3E). Importantly, PARK2 depletion resulted in a significantly lower PTEN enzymatic activity (Figures 4D and S3F) and higher levels of PI(3,4,5)P3 and PI(3,4)P2 in cells (Figure 4E), further supporting the functional contribution of PTEN in PARK2-mediated regulation of the PI3K/Akt pathway. Moreover, we were struck to observe a reciprocal correlation between Parkin and PTEN protein expression across ten breast cancer cell lines that have previously been reported to be WT for PTEN (Figure S3G).

15In view of our findings, we speculated that suppression of PTEN protein levels by PARK2 depletion might be regulated by proteasome-mediated protein degradation. Indeed, treatment with proteasome inhibitor MG132 led to an overall increase in PTEN protein levels and no significant difference between GFP and PARK2 knockdown cells (Figures 4F and S3H). Taken together, these results show that PARK2 depletion suppresses PTEN protein levels and activity, and they raise an interesting hypothesis that, mechanistically, this could be regulated, at least in part, by ubiquitin proteasome-mediated protein degradation.

16Recent evidence shows that Parkin loss leads to a marked decrease in mitochondrial biogenesis. To investigate the role of PARK2 depletion in cellular metabolism, we initially measured the effect of knocking down PARK2 expression on the oxygen consumption rate (OCR) of shGFP- and shPARK2-expressing cells, in the presence or absence of different mitochondrial stress inhibitors, using the Seahorse XFe96 Flux Analyzer. Notably, PARK2 depletion resulted in a marked decrease in basal, maximal, mitochondrial, and non-mitochondrial respiration (Figures 5A, 5B, S4A, and S4B).

17We next explored the effects of PARK2 depletion on cellular metabolism by performing targeted metabolomic analysis. Consistent with lower OCR levels, PARK2 knockdown cells showed a significant decrease in ATP levels (Figure 5C) and a concomitant activation of AMPK (Figures 5D and S4C), accompanied by the inhibition of mTORC1 signaling (Figures S4D and S4E). PARK2 depletion also caused a reduction across many of the cofactors that contribute in nitric oxide (NO) biosynthesis, involving the conversion of L-arginine to L-citrulline (Figures 5E and 5F). Importantly, although all the anaplerotic substrates for NO synthesis were detected to be lower in the PARK2 knockdown cells, L-citrulline levels were not significantly lower. Although other interpretations are possible, these results are consistent with a model in which NO synthase (NOS) is activated due to the loss of PARK2, such that the substrates for NOS are being consumed at a high rate while the product of NOS L-citrulline is maintained at a relatively constant level.

18Indeed, consistent with this model, we measured the NOS activity and oxidized NO (NO2 + NO3) levels, and we found that they are significantly higher in PARK2 knockdown cells (Figures 5G, 5H, and S4F). The latter also showed an overall increase in NADP/NADPH ratio (Figure 5I), higher reactive oxygen species (ROS) levels upon menadione treatment (Figures 5J and S4G), and lower reduced/oxidized glutathione (GSH/GSSG) levels (Figures 5K and S4H) compared to GFP control knockdown cells. Importantly, besides its function as a regulator of cellular redox status, GSH has a crucial role in modulating NO reactivity. NO as a free radical is a poor oxidant, hence, NO-dependent cysteine oxidation primarily occurs in the presence of oxygen or high ROS. Of note, we showed that GSH-MEE (reduced ethyl ester) treatment restored the suppression in PTEN protein levels observed in PARK2 knockdown cells (Figures 5L and S4I), while L-Buthionine-sulfoximine (BSO) treatment further enhanced the reduction in PTEN (Figure S4J). In addition, the activity of PTEN in PARK2 knockdown cells in the presence of DTT was comparable to the shGFP cells without the DTT pre-incubation (Figure S4K), reinforcing the conclusion that the inhibition of PTEN is generated by a redox-dependent modification of cysteine residues.

19Mitochondrial dysfunction and oxidative stress are well-recognized mechanisms leading to the activation of AMPK that can directly phosphorylate endothelial NOS (eNOS) to increase its catalytic activity. In line with this, we showed that PARK2 knockdown (Figure S4L) or allosteric activation of AMPK following treatment with the small molecule activator 991 (Figure 5M) led to an increase of eNOS phosphorylation and a reduction in PTEN protein levels. Of note, this reduction in PTEN protein levels and concomitant increase in Akt activation that were observed in 991-treated cells were decreased upon co-treatment with the NO scavenger carboxy-PTIO (cPTIO) (Figure S4M). Collectively, these results demonstrate that PARK2 depletion leads to mitochondrial dysfunction, high eNOS activity that is mediated, at least in part, by AMPK activation, and high ROS levels that together coordinate efficient NO oxidation.

20High NO production exerts a pleiotropic range of biological functions that are regulated, in part, by a post-translational redox-mediated modification known as S-nitrosylation of protein cysteine residues. PTEN has previously been identified as a target of such modification that leads to the inhibition of its enzymatic activity and downstream activation of Akt signaling. In view of our previous findings, we speculated that PARK2 depletion-mediated activation of the Akt pathway might be regulated by PTEN S-nitrosylation. Indeed, PARK2 depletion resulted in a marked induction of PTEN S-nitrosylation (SNO-PTEN), as indicated both by immunoblotting immunoprecipitated PTEN with the anti-S-nitrosocysteine antibody (Figures 6A and S5B) and by measuring the release of NO from S-nitrosothiol of recombinant PTEN using the quantitative fluorescent 2,3-diaminonaphthalene (DAN) assay (Figures 6B and S5C). Interestingly, PTEN S-nitrosylation has previously been shown to promote its ubiquitination, which could also explain the reduction in total PTEN protein levels and rescue following treatment with the proteasome inhibitor MG132 (Figures 4F and S3H). We performed an in-cell PTEN ubiquitination assay and showed that PARK2 depletion resulted in a marked increase in PTEN ubiquitination, with or without pre-treatment with MG132 for 6 hr (Figures 6C, 6D, and S5D). Consistent with enhanced PTEN ubiquitination, we also found an increased abundance of conjugated ubiquitin (ub) on PTEN in PARK2 knockdown HCT116 cells expressing HA-Ubiquitin (Figure S5A). An increase in phospho-Akt levels was observed in MG132-treated cells even in the absence of PARK2 knockdown (Figure 6C), perhaps reflecting the increase in ub conjugates and PTEN ubiquitination (Figure 6D) resulting from proteasome inhibition.

21Cysteine (Cys-83) has previously been identified as the critical cysteine thiol group within PTEN’s phosphatase domain that is predominantly targeted for S-nitrosylation, and, in line with this, C83S mutant PTEN is resistant to modification. Accordingly, we overexpressed WT and C83S mutant PTEN in the presence or absence of PARK2 lentiviral small hairpin RNA (shRNA) to show that the C83S mutant was completely devoid of S-nitrosylation and rescued Akt activation in PARK2 knockdown cells (Figure 6E). Importantly, since the activation of eNOS is also regulated by Akt-dependent phosphorylation, our observation raised a dilemma as to whether the contribution of PARK2 depletion to PTEN inhibition by S-nitrosylation is the consequence of Akt activation or is regulated via AMPK-mediated activation of eNOS independently of PI3K-Akt. To address this, we treated shGFP- and shPARK2-expressing HCT116 with the Akt inhibitor MK-2206 to show that inhibition of Akt activity does not rescue S-nitrosylation of PTEN (Figure S5E) nor the increased NO levels in PARK2 knockdown cells (Figure S5F). On the contrary, small interfering RNA (siRNA)-mediated downregulation of AMPK (AMPK α1 and α2) led to a reduction in PTEN S-nitrosylation in PARK2 knockdown cells (Figure 6F), further supporting the mechanistic implication of AMPK in PARK2-mediated activation of Akt.

22To determine the importance of AMPK activation to trigger PTEN S-nitrosylation, we tested if AMPK activation alone is sufficient to induce this modification in the absence of PARK2 depletion. Activation of AMPK following 991 treatment, glucose deprivation, or oligomycin treatment all led to a marked increase in PTEN S-nitrosylation (Figure 6G), identifying a functional cross-talk between AMPK and Akt activation.

23To better understand the physiological importance of PTEN S-nitrosylation in the proliferation and survival of cells under conditions of metabolic stress, we performed clonogenic assays in PTEN null PC3 cells overexpressing EV, WT, or SNO-resistant (C83S) PTEN in the presence or absence of the glycolysis inhibitor 2-Deoxy-D-glucose (2-DG). Although no significant difference in colony formation was detected in PTEN WT-overexpressing cells with or without 2-DG treatment, there was a marked reduction (∼30%) in the number of colonies following 2-DG treatment in PC3 cells with enforced expression of the C83S mutant PTEN (Figures S6A–S6C). Moreover, we compared the half-maximal inhibitory concentrations (IC50s) of 2-DG and dichloroacetate (DCA) (PDK inhibitor), both of which result in the activation of AMPK in response to ATP depletion, to show that PC3 cells expressing the C83S mutant exhibited significantly higher sensitivity to treatment with these drugs, not evident in PTEN WT-expressing PC3 cells (Figures S6D and S6E). Last but not least, and consistent with the ability of PARK2 knockdown cells to inactivate WT PTEN, we showed that shPARK2 cells proliferate significantly faster than shGFP-expressing cells in the presence of ectopic co-expression of WT (Figure S6F), but not C83S mutant, PTEN (Figure S6G). Taken together, these data demonstrate that PARK2 depletion contributes to the activation of Akt signaling through promoting S-nitrosylation and ubiquitination of the tumor suppressor PTEN. Notably, our data also highlight a previously unexplored mechanism contributing to AMPK-mediated activation of PI3K/Akt involving the inhibition of PTEN by S-nitrosylation, which appears to be critical for the proliferative capacity and survival of PTEN-proficient cancer cells under conditions of energy stress.

24Given the functional contribution of PARK2 depletion in PTEN inactivation by S-nitrosylation, we performed a bioinformatic analysis to report the copy number alterations (CNAs) of PARK2 and PTEN across 995 cancer cell lines of the Cancer Cell Line Encyclopedia (CCLE) project. We were struck to identify that almost one-half of cell lines with HET deletion in PTEN (n = 145/314, 46.2%) also have HET deletion in PARK2, with the coexistence for loss of heterozygosity (LOH) at both the PTEN and PARK2 loci being far more frequent than PTEN (n = 79/314, 25.2%) or PARK2 (n = 119/344, 34.6%) LOH alone (Figure S7A). We extended this bioinformatic analysis summarizing the percentage of PARK2 CNAs of primary tumors or cancer cell lines with PTEN LOH on 1,953 specimens across 13 different cancer types from the TCGA database to show that there is a strong selection for PARK2 LOH, when one copy of the PTEN gene is missing (Figure S7B).

25In an effort to study whether Park2 loss further exacerbates Pten-mediated tumorigenesis in vivo, mice with a targeted knockout of Park2 exon 3 were crossed with Pten+/− mice. The latter were born viable and developed prostatic intraepithelial neoplasia (PIN), as well as numerous neoplastic lesions in many organs, including skin, colon, endometrium, liver, thyroid, and thymus. Although Pten+/−/Park2+/− and Pten+/−/Park2−/− mice showed a similar spectrum of tumors like the ones detected in Pten+/−/Park2+/+ mice, they were significantly more tumor prone and exhibited features reminiscent of mice with a hypomorphic and a knockout Pten allele (Figure 7A). Park2 loss was associated with an increased incidence of high-grade PIN and low-grade prostate adenocarcinoma (Figures 7B and 7D), as well as significantly higher numbers of solid thyroid adenocarcinomas (Figure 7C). Most importantly, mice bearing Park2 loss developed tumor types, including hystiocytic sarcoma, multiple myeloma, and osteosarcoma in spine (Figures S7C and S7D), that have not been previously described for the Pten+/− mice. On two occasions, mice bearing both Park2 and Pten loss developed malignant pheochromocytomas and thyroid adenocarcinomas that metastasized to the lung (Figures S7C and S7D). As a result, Pten+/−/Park2+/− and Pten+/−/Park2−/− mice showed significantly poorer survival compared to Pten+/−/Park2+/+ mice (Figures 7E and 7F), while no difference in relative survival was found between Pten+/−/Park2+/− and Pten+/−/Park2−/− mice (Figure S7E).

26To further assess the contribution of the proposed model on the PARK2-mediated inactivation of PTEN in vivo, we performed immunohistochemical (IHC) analysis of PTEN and phospho-Akt on tumor sections derived from Pten+/− mice bearing WT, heterozygous (HET), or homozygous (HOM) deletion in Park2. Consistent with reduced PTEN and elevated phospho-Akt expression observed in the immunoblot analyses, Pten+/− tumor epithelia bearing one or no copies of Park2 expressed no or low PTEN protein levels, and they showed significantly higher phospho-Akt staining compared to lesions in Pten+/−/Park2+/+ mice (Figures S7F–S7H). Overall, our analysis showed that losses of function of Park2 and Pten displayed striking cooperativity to promote tumorigenesis and significantly shorten tumor-free survival.

27The PI3K/Akt pathway represents a complex signaling network that integrates numerous upstream stimuli to regulate diverse cellular processes, including cell growth, proliferation, survival, and migration. Interestingly, the PARK2 gene is associated with the pathogenesis of disease states that are characterized by different responses in cell fate determination; therefore, understanding its functional contribution in the activation of survival pathways is clearly of immense clinical benefit.

28In this study, we identify a PTEN-mediated role for PARK2 depletion in the activation of the PI3K/Akt pathway. Of note, recent evidence suggested that PARK2 interacts with EGFR to promote its ubiquitination, thereby inhibiting Akt activation. However, the role of EGFR in PARK2-mediated activation of PI3K-Akt signaling might be more complex and/or cell type specific than previously appreciated, as ectopic PARK2 expression suppresses Akt activation even in EGFR null SW620 cells.

29PTEN is a well-characterized haploinsufficient tumor suppressor, yet emerging evidence suggests that its function goes beyond its tumor suppressor role as a critical regulator of multiple CNS functions. PTEN loss and subsequent activation of the PI3K/Akt signaling promotes the activation of mTOR signaling, which is hyperactive in many cancers. Notably, the inhibition of mTORC1 by rapamycin treatment prevents PD symptoms in mice bearing a human mutation in the PARK2 gene. Furthermore, PTEN loss leads to the downregulation of PINK1 expression, supporting a mechanism to suggest that PARK2 LOH could repress PINK1 through inactivating PTEN and PINK1 repression could obliterate any residual activity of the remaining PARK2 allele, by abrogating Parkin translocation to mitochondria.

30An important mode of NO function primarily involves its reaction with oxygen or ROS and subsequent oxidation to nitrogen dioxide (NO2), dinitrogen trioxide (N2O3), or peroxynitrite (ONOO−), which ultimately lead to nitrosative stress and S-nitrosylation of target proteins and other biomolecules. Of note, NO signaling and S-nitrosylation have previously been detected to contribute to PTEN inactivation in neurodegeneration, but the role in cancer has been largely unknown. In this study, we showed that one of the most common genetic alterations across human cancers, PARK2 loss, contributes to S-nitrosylation of PTEN, thereby promoting its ubiquitin-dependent degradation by the proteasome. We also report that there is a strong selection for co-occurrence of PARK2 and PTEN LOH in cancer. This suggests that the high incidence of complete IHC loss of PTEN in PTEN HET tumors could be explained, at least in part, due to PARK2 LOH, leading to S-nitrosylation and ubiquitination of the protein encoded from the remaining PTEN allele. In line with our previous observations, PARK2 has many of the properties of a haploinsufficient tumor suppressor in that loss of a single Park2 allele exacerbates tumorigenesis without requiring complete inactivation of the remaining allele.

31Notably, activation of AMPK alone can trigger a marked increase in PTEN S-nitrosylation, pointing to a functional cross-talk between AMPK and Akt activation in the absence of any mTORC1-dependent negative feedback loop mechanisms. This redox-dependent modification in PTEN is important for supporting the survival and proliferative capacity of energy-deprived cancer cells, signifying an important compensatory role for AMPK to support cell homeostasis via PTEN-mediated activation of Akt signaling. Further studies are needed to determine the physiological functions of enhanced NO signaling across different disease states, but the reaction between NO and PTEN forms a nexus that opens up unique therapeutic opportunities for targeting dysregulated protein S-nitrosylation for a substantial fraction of tumors growing under conditions of energy deprivation.

32Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, George Poulogiannis (george.poulogiannis@icr.ac.uk).

33Pten+/− heterozygous mice were bred with mice harboring a targeted knockout of Park2 exon 3. All mice were previously backcrossed over ten generations and maintained in a C57BL/6 background and they were genotyped by PCR for Pten and Park2 alleles as described previously. Tumor-bearing mice were euthanized and subjected to whole-body histological analysis. Normal and tumor tissues were fixed in 4% PFA, embedded in paraffin, sectioned, and subjected to hematoxylin and eosin (H&E) staining for pathological evaluation. Survival analyses were performed using Kaplan–Meier curves and the Log-rank test. Six- to eight-week old NOD.CB17-Prkdcscid/J mice were injected via the retrobulbar sinus with 200 μl cell suspension (1x106 cells) of GFP flow-sorted luciferase labeled (pLV430G-oFL-T2A-eGFP) H1299 cells expressing pLenti6∕UbC∕V5-DEST- empty vector or PARK2. Development of metastases was monitored by luciferin injection and bioluminescence imaging at 3 and 4 weeks post-injection. Bioluminescence (BLI) signal intensities were determined by using the region of interest (ROI) tool in Living Image Software (PerkinElmer). After the last imaging session the mice were euthanized and the lungs were surgically removed and inflated with 4% formalin in PBS. Lung sections, 5 μm, were subjected to H&E staining and the number of lung tumor nodules was counted using a dissection microscope. All mice were maintained according to NIH-approved institutional animal care guidelines and the study was approved by the Institutional Committee at the Beth-Israel Deaconess Medical Center.

34293T and PC3 cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 μg/ml streptomycin. HCT116 parental and isogenic PTEN null cell lines were cultured in McCoy’s 5A, H460 and H1299 cells in RPMI, and U138 cells in DMEM-F12, all supplemented as above. The breast cancer cell lines MDA-MB-134, BT-474, UACC-812, MCF-7, Hs-578T, SK-BR-3, AU-565, BT-20 and JIMT-1 were cultured in DMEM, and the ZR-75-30 cells in RPMI, all supplemented as above. MCF10A parental and isogenic PTEN null cells were cultured in DMEM-F12 supplemented with 5% horse serum (HS) (ThermoFisher, 16050-122), 20 ng/ml EGF (PeproTech, AF-100-15), 0.5 mg/ml hydrocortisone (Sigma-Aldrich, H0888), 100 ng/ml cholera toxin (Sigma-Aldrich, C8052), 10 μg/ml insulin (Sigma-Aldrich, I1882), supplemented with penicillin and streptomycin as above. All cell lines were maintained at 37°C in a humidified incubator with 5% CO2 and were tested and confirmed to be negative for mycoplasma infection. For glucose starvation experiments, cells were washed twice with PBS and then incubated in DMEM without glucose and sodium pyruvate supplemented with 10% dialyzed FBS for 6 hr. For treatment with the allosteric AMPK activator 991, cells were serum-starved for 2 hr prior to 991 treatment for 5 hr. All cell lines used in the study were submitted to Eurofins Genomics for autosomal short tandem repeat (STR) authentication.

35For lentiviral gene knockdown, pLKO.1 shRNA sequences against human PARK2 or GFP control were transfected in 293T cells using the FuGENE 6 transfection reagent (Promega, E2691) according to the manufacturer’s protocols. Infected cells were selected in the presence of 2 μg/ml puromycin (InvivoGen, ant-pr-5b) for at least 4 days. Stably expressing cancer cell lines were established by infecting with virus encoding human wild-type PARK2, the E3 ligase dead C431S mutant or empty vector in the pLenti6∕UbC∕V5-DEST backbone (ThermoFisher, V49910). Infected cells were selected in the presence of 2-10 μg/ml blasticidin (InvivoGen, ant-bl-1) for 2 weeks. For stable expression of luciferase and GFP, H1299 cells were infected with pLV430G-oFL-T2A-eGFP, a lentivirus expressing the luciferase gene cloned in a eGFP expressing vector (gift from Dr Tina Yuan, Dana-Farber Cancer Institute, Harvard Medical School, Boston), and GFP-positive cells were sorted by flow cytometry 1 week after infection. H460 cells were transfected with PARK2 sgRNA CRISPR/Cas9 All-in-One Lentivector Target 3 (Applied Biological Materials, K1594608) or control LentiCRISPR v2 according to the manufacturer’s instructions. Infected cells were selected in the presence of 2 μg/ml puromycin for 4 days and isolation of PARK2 clonal deletion mutants was validated by immunoblotting analysis and DNA sequencing.

36Cells were washed with ice-cold PBS and lysed on ice for 30 min with cell lysis buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, supplemented freshly with a protease and phosphatase inhibitor cocktail (5872, Cell Signaling Technology), 10nM Calyculin A (Cell Signaling Technology, 9902) and 1 mM DTT (ThermoFisher, R0861). Lysates were subjected to centrifugation at 12,000 g for 10 min at 4°C and protein concentrations were determined using the Bradford assay (Bio-Rad, 5000006). Protein lysates were boiled for 10 min and subjected to SDS-PAGE electrophoresis. Densitometry was calculated using the Image Lab Software 5.2.1 (Bio-Rad).

37The GBM tissue microarray (TMA) was prepared and characterized as described previously from the Tumor Tissue Bank at University Health Network, University of Toronto. Briefly, the TMA was de-waxed in xylene followed by rehydration in a standard alcohol series. Antigen retrieval was by pressure cooking for 20 min in citrate buffer (pH 6.0), followed by blocking of endogenous peroxidase in 0.3% H2O2. Antibody incubation was performed using the ABC reagent kit (Vector Labs, PK-6100) as per manufacturers instructions. Briefly, Parkin antibody (ThermoFisher, PA1-38412) was diluted 1:100 in antibody diluent and added to the slides and incubated overnight at 4°C. Detection was performed using biotinylated secondary antibodies for 30 min, the ABC reagent kit and 3,3′-Diaminobenzidine chromogen. Sections were counter-stained with hematoxylin (ThermoFisher, 008001) for 30 s, dehydrated in 70, 80 and 100% ethanol, briefly washed in xylene and mounted in Permount (ThermoFisher, SP-15). For PTEN and phospho-AKT immunohistochemistry of mouse tumor sections, antigen retrieval was carried out by microwave heating for 20 min in citrate buffer (pH 6.0). Anti-PTEN (Dako, M3627) and anti-phospho-Akt (Ser473) (Cell Signaling Technology, 4060) antibodies were applied at dilutions of 1:200 and 1:50 respectively. PTEN and phospho-Akt IHC staining intensity was graded as 0 (negative), 1 (weak), 2 (moderate), and 3 (strong). Images were captured on a Nikon E-600 microscope and analyzed using Nikon ACT-1 software.

38For PTEN phosphatase activity assay, PTEN-transfected cells [pRK5-Myc-PTEN-WT,] were lysed in lysis buffer containing protease and phosphatase inhibitors in the absence of DTT. PTEN was immunoprecipitated from 1 mg of cell lysates and the PTEN phosphatase activity was measured using a PTEN activity ELISA kit (Echelon Biosciences Inc, K-4700), following the manufacturer’s instructions. Where indicated, PTEN immunoprecipitates were treated with 50 mM DTT at 4°C for 30 min prior to measuring PTEN phosphatase activity.

39S-nitrosylation of PTEN was detected using the 2,3-diaminonaphthalene (DAN) assay. Briefly, PTEN transfected cells (pRK5-Myc-PTEN-WT) were lysed in lysis buffer containing protease and phosphatase inhibitors and 40 mM N-ethylmaleimide in the absence of DTT. PTEN was immunoprecipitated from equal volume of the diluted cell lysates containing 1 mg of soluble protein. The immunoprecipitates were washed twice with lysis buffer and twice with PBS. The pellet was resuspended in 500 μL of PBS and 100 μM HgCl2 and 100 μM DAN was added. The samples were incubated in the dark at room temperature (RT) for 30 min and 1 M NaOH was added. The generated fluorescent triazole from the reaction of DAN with the NO released from PTEN was measured using an excitation wavelength of 375 nm and an emission wavelength of 450 nm. As a negative control, the PTEN antibody alone in lysis buffer was immunoprecipitated and the resulting background fluorescence intensity was subtracted from each sample. S-nitrosylated PTEN was confirmed by immunoblotting of immunoprecipitated PTEN with anti-S-nitrosocysteine antibody (Abcam, ab94930), under non-reducing conditions. For non-reducing SDS-PAGE, β-mercaptoethanol was omitted from the loading buffer and samples were not boiled.

40Cellular assays to measure PTEN polyubiquitination were performed as described previously. Briefly, cells were transfected with vectors encoding wild-type PTEN in the pRK5-Myc vector and HA-Ub and lysed in the presence of 40 NEM and supplemented freshly with a protease and phosphatase inhibitor cocktail. Equal volume of the diluted cell lysates containing 1mg of soluble protein was incubated with 1 μg mouse monoclonal anti-PTEN (Santa Cruz Biotechnology, sc-7974) or 2 μg mouse anti-HA (BioLegend, 901501) antibodies pre-coupled with protein G Sepharose beads (Sigma-Aldrich, P3296) and incubated at RT for 2 hr rotating. Beads were washed four times in immunoprecipitation buffer (50 mmol/L Tris, pH 7.6, 100 mmol/L NaCl, 2 mmol/L EDTA, and 0.2% Nonidet P-40) and the bound proteins were released by boiling in SDS-PAGE sample buffer for 10 min, prior to SDS-PAGE analysis and immunoblotting with the indicated antibodies.

41Subconfluent cells in 10-cm dishes were labeled in 8 mL of inositol-free DMEM and 10% dialyzed FBS supplemented with 160 μCi 3H-myo-inositol (PerkinElmer, NET1168001MC, specific activity = 20.1 Ci/mmol) for 24 hr. Deacylated phosphoinositides were resolved by HPLC using an Agilent 1200 Quaternary system, and radioactivity detected in-line using a Packard Flo-one Radiomatic detector. HPLC Buffer A is 1 mM EDTA, and Buffer B is 1 mM EDTA and 1 M NaH2PO4. An Agilent Zorbax SAX column (5 μm, 4.6 X 250 mm) was eluted by gradient program (from 100% A to 2% B at 1 min, 14% B at 30 min, 30% B at 31 min, 66% at 60 min, 100% B at 85 min, 100% A at 86 min, and hold at 100% A until 110 min) at a flow rate of 1 ml/min. PI(3,4)P2, PI(3,4,5)P3 standards were prepared by reacting PI3K with PI4P and PI5P, respectively and [γ-32P]ATP.

42Cell viability was assessed using the CellTiter96 AQueous non-radioactive cell proliferation assay [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt; MTS] (Promega, G5421) according to the manufacturer’s protocol. Caspase-3/7 activities were measured using the Caspase-Glo 3/7 assay (Promega, G8090) following treatment with 50 or 100 nM staurosporine (Sigma-Aldrich, S4400) for 1 hr. For both cell viability and caspase 3/7 assays, HCT116 cells were seeded at 3x104 cells/well in a 96-well plate and incubated overnight before treatment.

43Proliferation kinetics of H1299, PC3, HCT116 PTEN+/+ and HCT116 PTEN−/− cells transfected with pLenti6/UbC/V5-DEST-PARK2-WT, PARK2-C431S mutant or empty vector were determined using the Sulforhodamine B (SRB) (Sigma-Aldrich, S1402) assay over a period of 5 days. The same method was applied for shGFP and shPARK2-expressing PC3 cells transfected with pHR-SIN-PTEN-WT, PTEN-C83S and control empty vector. Briefly, 2x103 cells were seeded in a 96-well plate in triplicate and allowed to adhere for 24 hr (Day 0). Cells were fixed at days 0, 1, 3 and 5 as follows: they were fixed with 10% trichloracetic acid at 4°C for 60 min and washed 4x with 100 μL dH2O. After being left to dry at RT, the cells were stained for 10 min with 0.4% (w/v) SRB dissolved in 1% acetic acid, and protein-bound dye was extracted with 10 mM unbuffered Tris base pH 10.5. Absorbance was determined at 510 nm using a 96-well microplate reader and data presented indicate the average number of triplicate experiments ± standard error of the mean (SEM).

44For cell migration assays, cells were starved overnight in 0.5% FBS and 5x104 cells/well were seeded in 24-well transwell chambers (Corning, Lowell, MA, USA; 8 mm pore size) in a total volume of 300 μl 0.5% FBS-media. 750 μl of media (supplemented with 10%FBS) were added to the bottom of the transwells and the plates were returned to 37°C for 24 hr incubation. The transwell inserts were washed twice in PBS and cells were fixed in 10% trichloroacetic acid (TCA) for 1 hr. The TCA containing media was removed and the inserts were washed 3x in H2O. Cells were stained in 0.4% (w/v) SRB dissolved in 1% acetic acid for 10 min. The inserts were washed 3x in 1% acetic acid and the cells on the top of the filter were removed using a cotton swab. Migrated SRB-stained cells in the bottom of the transwell inserts were counted under a light microscope at 20x magnification in 5 fields/well.

45For determination of the IC50 values of 2-DG and DCA, shGFP- or shPARK2-expressing PC3 cells were co-transfected with pHR-SIN-PTEN-WT, PTEN-C83S and control empty vector. 48 hr post transfection, the cells were seeded at 5x103 cells/well in a 96-well plate and treated with various concentrations of 2-DG (0.5-50 mM) or DCA (5-80 mM) for 72 hr followed by SRB staining. For clonogenic assays, cells were plated at 1 × 102 cells/well in 24-well plates and treated with 10 mM DCA or 0.5 mM 2-DG. After 10 days, colonies were stained with a Crystal Violet solution (0.5% Crystal Violet, Sigma, 30% ethanol and 3% formaldehyde) and were counted in a GelCount™ (Oxford Optronix). Every experiment was performed in triplicate and data are presented as mean values ± SEM.

46Total RNA was extracted using the ReliaPrep RNA cell miniprep system (Promega, Z6010) according to the manufacturer’s instructions. Reverse transcription and real-time PCR reactions were carried out using the QuantiTect reverse transcription kit (QIAGEN, 205310) and SYBR select master mix (ThermoFisher, 4472908) respectively, using the TProfessional ThermoCycler from Biometra. The data presented are the mean values obtained ± SEM from triplicate reactions. Primer sequences used are as follows; Human PARK2, sense: TCAATCTACAACAGCTTTTATG, anti-sense: TGCACTAGTCCCAGGGCA, Human PTEN, sense: CAAGATGATGTTTGAAACTATTCCAATG, anti-sense: CCTTTAGCTGGCAGACCACAA, Human 18S, sense: GTGGAGCGATTTGTCTGGTT, anti-sense: CGCTGAGCCAGTCAGTGTAG.

47H2O2 levels were measured using the ROS-Glo H2O2 assay (Promega, G8820) according to the manufacturer’s instructions. Briefly, 1x104 shGFP control and shPARK2-expressing cells were seeded in a 96-well plate and incubated at 37°C and 5% CO2 for 18 hr. Cells were washed with PBS and treated in Hank’s BSS medium containing 20 μM menadione (Sigma-Aldrich, M5625) in the presence of 20 μl 125 μM H2O2 substrate (provided in the kit). Following incubation at 37°C for 2 hr, 100 μl of ROS-Glo detection solution was added to the wells and the plate was incubated for 20 min at RT. Luminescence was determined with a microplate reader and the average relative light units (RLU) ± SEM of triplicate data are reported.

48Nitric oxide (NO) levels were detected using the Nitric Oxide Assay Kit (abcam, ab65327) according to the manufacturer’s protocol. Briefly, shGFP and shPARK2-expressing cells were grown in regular culture medium for 24 hr. After washing with cold PBS, 2x106 cells were resuspended in 500 μl assay buffer and homogenized quickly by pipetting up and down a few times. The samples were centrifuged for 5 min at 4°C at top speed to remove any insoluble material. Triplicate 75 μl aliquots of the supernatant were transferred to a 96-well plate and 5 μl of enzyme cofactor and 5 μl of nitrate reductase were added to each of the reaction wells. The plate was incubated at RT for 3 hr and 5 μl of enhancer was added to each well, before further incubation for 30 min at RT. 5 μl of the DAN probe was added to each well, the plate was incubated at RT for 10 min and 5 μl of NaOH was added before final incubation at RT for 10 min. Fluorescence was determined with a microplate reader at Ex/Em = 360/450nm. All reagents were provided in the kit. GSH:GSSG levels were measured using the GSH/GSSG-Glo Assay kit (Promega, V6611) and nitric oxide synthase (NOS) activity was determined using the NOS activity assay kit (BioVision, K205-100) according to the manufacturer’s protocol.

49PARK2 and GFP stable knockdown cells were plated at 2x104/well in a 96-well Seahorse cell culture microplate and incubated overnight. The next morning, culture media was replaced with pH-adjusted (pH = 7.4 ± 0.1) bicarbonate-free DMEM with 10 mM Glucose, 1 mM sodium pyruvate and 2 mM L-Glutamine and the plate was incubated at 37°C for 1 hr in a non–CO2 incubator. Oxygen consumption rates were measured using the Seahorse XF Cell Mito Stress Kit (Agilent, 103015-100) on a XFe96 Analyzer. 2 μM Oligomycin, 0.5 μM FCCP, and 0.5 μM rotenone/antimycin (R/A) were used for all conditions. Cell numbers were normalized using Cyquant (ThermoFisher, C35012).

50Metabolite levels were determined by targeted liquid-chromatography tandem mass spectrometry (LC-MS/MS) analysis as described previously. Briefly, 48 hr prior to each experiment, 2.5x105 HCT116 cells expressing shGFP or shPARK2 were seeded in 6-cm dishes. The media for each plate was replaced 2 hr prior to metabolite extraction and after aspiration, 4 mL pre-chilled (at −80°C) methanol was added to the cells on dry ice for 15 min. Cell extracts were collected into 15 mL conical tubes and centrifuged for 5 min at 4200 rpm. Solvent in the resulting supernatant was evaporated using a centrifugal vacuum evaporator (“SpeedVac”) and samples were re-suspended in 20 μL HPLC-grade water for mass spectrometry. 8 μL were injected and analyzed using a Prominence UFLC HPLC system (Shimadzu) for hydrophilic interaction liquid chromatography (HILIC), coupled to a QTRAP 5500 hybrid triple quadrupole/linear ion trap mass spectrometer (SCIEX) operated in the selected reaction monitoring (SRM) mode. Peak areas of LC-SRM-MS traces for each metabolite were integrated using the MultiQuant v1.1 software (SCIEX).

51GISTIC (Genomic Identification of Significant Targets in Cancer) analyses were performed on DNA copy number data from the SNP pipeline version 3.0. 9,863 tumor samples across 28 cancer types were selected for this analysis using the 2015-06-01 stddata TCGA/GDAC tumor sample sets from FireHose, and the frequencies of PARK2 DNA copy number changes were plotted using a polar histogram. The Oncomine cancer profiling database was used to analyze PARK2 mRNA expression across 13,499 cancerous and 2,708 normal tissue specimens from 125 independent microarray datasets representing 27 different cancer types. The degree of PARK2 mRNA underexpression (≥0.5 and ≥ 1-log2-fold change in tumor versus normal) was plotted across all tumor types. Kaplan-meier analysis in Figure S1C was performed using the KM plotter. Student’s t test and one- or two-way ANOVA were used to evaluate the statistical significance among different variables as indicated in the respective figure legends; n.s. (not significant), ∗ (p < 0.05), ∗∗ (p < 0.01).

52G.P. and L.C.C. designed the study and wrote the manuscript. G.P., A.G., S.A.-V., N.K., M.D., A.V., Y.Z., Y.-H.C., and D.A. performed molecular biology experiments. S.A. and G.Z. performed Parkin IHC in GBM tumors. M.J.A. analyzed histopathology data. G.P. and J.M.A. performed metabolomics and analyzed the data. G.P. and R.C.-W. performed retro-orbital injection experiments and contributed to mouse colony management. G.P. performed bioinformatic analysis. All authors commented on the manuscript.