Parkin and its molecular associations in gliomas – a systematic review

Abstract

Parkin, a protein encoded by PRKN, discovered in the context of Parkinson’s disease, controls proteasomal degradation by protein ubiquitination and acts on cell cycle control and mitochondrial homeostasis, among other cellular processes. Parkin has been also implicated in several carcinomas, melanoma and leukemia. In the neoplastic setting, reduced parkin level usually indicates poorer prognosis. Some authors have described the associations between parkin and gliomas. Gliomas are a heterogeneous group of tumors that arise in the central nervous system, astrocytomas being the most common. The aim of this systematic review is to evaluate how parkin behaves in gliomas and the molecular pathways associated in this interaction. A search was conducted in PubMed, EBSCO and Scopus and 8 published articles were identified as eligible studies. The studies were categorized in three groups, according to their main emphasis: PRKN mutation patterns detected in gliomas, parkin effects on tumor growth and survival rates, and molecular interactions between parkin and other proteins. The studies showed higher PRKN mutation rates and lower parkin expression in high grade gliomas. Patients with higher parkin expression had better overall survival. Besides, different molecular pathways associated with parkin were described, some of them regarded as potential therapeutic targets.

Introduction

Parkin is a protein encoded by the parkin RBR E3 ubiquitin protein ligase (PRKN), a gene discovered two decades ago (Hattori and Mizuno 2017). Back then, it was linked to juvenile Parkinson’s disease, which explains the given name “parkin” (Shimura et al. 2000), and over time, the knowledge about parkin has increased substantially. Full-length parkin exhibits 2960 BP, equivalent to 465 amino acids (Kitada et al. 1998). In human brain tissue, parkin is expressed in the cytoplasm and nucleus of neurons and glia cells (Shimura et al. 1999; Yin et al. 2009). D’Agata et al. (2002) analyzed rat brain tissue and concluded that parkin is expressed throughout the brain. However, both in human and other mammals, the primary transcript suffers subsequent alternative splicing, which generates diverse mature RNA and therefore different protein isoforms, as demonstrated by Scuderi et al. (2014). These isoforms differ not only in structure but also in function, and they are expressed differently according to the tissue (Scuderi et al. 2014). D’Amico et al. (2016) showed that some parkin isoforms are expressed globally in rat brain samples while other isoforms differ according to the anatomic region. Parkin classical function is to control proteasomal degradation by inducing ubiquitination of damaged proteins (Martinez et al. 2017; Panicker et al. 2017; Zheng and Hunter 2015) and thus, parkin has a multifunctional role, involved in several cellular processes, such as cell cycle control; cell proliferation and migration; protection against oxidative stress; mitochondrial homeostasis; xenophagy; and tumor suppression (Drake et al. 2017; Shires et al. 2017; Durcan and Fon 2015; Gong et al. 2017). Moreover, PRKN mutations have been associated with several cancers, such as ovarian, colorectal, liver, melanoma and leukemia, among others (Cesari et al. 2003; Klimczak et al. 2016; Tay et al. 2010; Xu et al. 2014). Reduced parkin levels often indicate an increased risk of lymph node metastasis and increased mortality rates (Da Silva-Camargo et al. 2018). To date, it is still unclear whether these functions are carried out by one protein or different isoforms (Scuderi et al. 2014). Some authors have studied the role of parkin in human gliomas (Yeo et al. 2012; Lin et al. 2015). Maugeri et al. described (Maugeri et al. 2015) 21 parkin isoforms and demonstrated that three variants were found in all human glioblastoma (GBM) samples and GBM-derived cell culture, which could turn these variants into efficient prognostic tools .

The glioma category consists of a group of tumors in which astrocytomas are the most common. Oligodendrogliomas e ependymomas are less frequent gliomas (Ostrom et al. 2017). They are graded by the World Health Organization (WHO) according to their microscopic findings and molecular profile, and high-grade gliomas, including GBM, have poor prognosis and high mortality rates (Louis et al. 2016). Results showed that parkin expression decreases in gliomas, compared to normal brain tissue (Yeo et al. 2012; Mulholland et al. 2006; Yin et al. 2009). Also, parkin expression and tumor grade are inversely correlated, suggesting a loss of PRKN function along with tumor progression (Lin et al. 2015; Viotti et al. 2014). The same happens between parkin and p53, in different gliomas, such as GBM and oligodendrogliomas (Viotti et al. 2014). Regarding survival rates, patients with higher parkin levels have better outcome (Yeo et al. 2012; Lin et al. 2015). The present systematic review aims to group these studies involving parkin and gliomas, as well as evaluate the molecular associations described in this interaction.

Methods

Focused question

According to Preferred Reporting Items for Systematic Reviews and Meta-Analyses. (PRISMA) guidelines, a Population, Intervention, Comparison and Outcomes (PICO) question was developed. This is a strategy used in evidence-based practice, that helps to create the research question, as well as the main elements for the literature search (Methley et al. 2014). The final question was: “How does parkin behave in gliomas and what are its associated molecular pathways?”

Literature search

A search was conducted in three different databases: PubMed, EBSCO and Scopus, on November 5th, 2019. The following keywords were used: “PARK2” or “PRKN” or “Parkin”, in the fields Title and/or Abstract. PARK2 and PRKN are different names used for the gene that encodes parkin, the latter is the current name. We decided not to include terms related to gliomas, because when those keywords were added, the search showed at times zero results. Therefore, we searched each and every study involving the protein parkin or its gene (PRKN; PARK2) and selected the ones related to gliomas after reading title and abstract. No publication date or language restrictions were applied.

Registration in PROSPERO

According to recommendations, this review protocol was prospectively registered in PROSPERO – International Prospective Register of Systematic Reviews, and approved on April 27th, 2020, under the number CRD42020172043.

Inclusion and exclusion criteria

Every study had to evaluate parkin and at least one associated molecule or gene, in the context of gliomas. We decided to include both in vitro and in vivo studies, since the number studies was limited.

We first excluded studies unrelated to parkin. Then, studies evaluating parkin’s role in the context of other diseases, such as Parkinson’s Disease, acute myocardial infarction, infections and tumors other than gliomas were excluded. A more detailed flowchart is shown in Fig. 1.

Fig. 1

PRISMA flow diagram, detailing the inclusion and exclusion criteria and the number of studies included and excluded at each step of the literature search

Data collection and items

Two investigators analyzed the eight studies and extracted data separately. The information collected included: names of the authors, year of publication, whether it included patients or not, whether it included in vitro studies or not (if yes: glioma cell lines used, PRKN mutations, colony fold change), whether it included in vivo studies (if yes: animal model, size and weight of the tumor, mortality rates and prognosis), other proteins or genes associated with parkin. Eventually, discrepancies were discussed in order to reach consensus. Those cases in which agreement was not achieved, a third investigator were consulted.

Statistical analysis

This systematic review resulted in a descriptive review, due to the methodological heterogeneity of studies: some approached mutational aspects, some in vitro or in vivo molecular interactions. Thus, a meta-analysis was not performed.

Results

Study selection

Our search was based on title and/or abstract, resulting in 10,031 papers. After excluding 2023 duplicates, we screened title and abstract, and eliminated other 7999 studies, according to inclusion and exclusion criteria described above. Nine studies were thoroughly read, resulting in 8 eligible papers (Fig. 1).

Characteristics of included studies

Among the eight included studies, none of them were randomized clinical trials. Four of them evaluated the mutational patterns in PRKN locus and its impact on parkin expression in gliomas. Three of them evaluated correlation between PRKN mutations and tumor proliferation status in in vitro models and replicated analysis in in vivo models and/or used public glioma databases. Five studies assessed parkin and other molecular pathways associated.

Main outcome of the studies

We divided the results in three categories. First, we describe PRKN mutation patterns in gliomas and the impact on parkin expression. Then, we report how parkin impacts on tumor proliferation and prognosis. Finally, we outline how parkin interacts with other molecular pathways. These findings are summarized in Fig. 2.

Fig. 2

Parkin molecular interactions and its effects on gliomas

PRKN mutation patterns in gliomas

Four studies (Mulholland et al. 2006; Yin et al. 2009; Veeriah et al. 2010 and Yeo et al. 2012) showed the presence of different mutations in PRKN locus in gliomas, reducing parkin expression, as well as mutations in other genes in adjacent loci (Table 1). Evaluations were based in different techniques such as Fluorescence in situ hybridization (FISH) probes, Deoxyribonucleic acid (DNA) sequencing and comparative genomic hybridization.

Table 1 Description of included studies

Parkin impacts on tumor proliferation and prognosis

Three studies evaluated how parkin expression influences tumor growth using in vitro and in vivo studies (Table 1). In vitro experiments showed that mutated PRKN tumor cell colonies had increased growth compared to wild-type (WT)-PRKN colonies (Yeo et al. 2012; Lin et al. 2015; Veeriah et al. 2010). These authors showed also higher cyclin level in mutated colonies. In vivo assessments were also performed and showed compatible results to in vitro experiments, since tumor size and weight were inferior in WT-PRKN animal models (Yeo et al. 2012; Lin et al. 2015; Veeriah et al. 2010). Also, Yeo et al. (2012) and Lin et al. (2015) found that parkin expression correlates to better prognosis.

Parkin interacts with other molecular pathways

As for parkin associations with other molecular pathways, we have different approaches in five studies. Yeo et al. (2012) evaluated the Phosphatase and tensin homolog (PTEN) pathway, a well-established mutation in high grade gliomas. PTEN loss of function promotes abnormal cell proliferation via Phosphoinositide 3-kinases/Protein Kinase B (PI3K/Akt) activation. They showed that lower levels of parkin expression resulted in increased PI3K/Akt pathway phosphorylation, increasing cyclin expression and inducing more proliferation. Viotti et al. (2014) established an inverse correlation between p53 and parkin expression. Lin et al. (2015) analyzed two molecular pathways and how parkin interacts with them. Through the evaluation of Wnt-beta-catenin pathway, they postulated that parkin induces beta-catenin degradation via proteasome. The second molecular pathway evaluated was Epidermal growth factor receptor-Protein Kinase B (EGFR-Akt). Two studies focused on the fact that GBM is an aggressive neoplasia presenting highly invasive behavior. Maugeri et al. (2016), based upon the fact that GBM presents areas of hypoxia, which induces the expression of hypoxia-inducible factors (HIFs), stimulating pro-angiogenic factors and greater invasiveness, found that parkin suppresses HIFs expression (Maugeri et al. 2016). Wang et al. (2017) also evaluated how parkin influences glioma invasiveness and metastasis (Wang et al. 2017). They demonstrated that parkin downregulates Zinc finger E-box-binding homeobox 1 (ZEB1). This molecule controls epithelial-mesenchymal transition (EMT) markers, such as vimentin and e-caderin and influences tumor invasion. Vimentin expression was higher and e-caderin expression was lower in silenced PRKN GBM cells.

Discussion

Parkin is a multifunctional protein discovered in the context of juvenile Parkinson’s disease. Lately, this protein has been associated with several types of cancer, including gliomas. Gliomas encompass a heterogenous tumor group classified and graded by the WHO. GBM is the most aggressive high-grade glioma, with poor prognosis and high mortality rates. This systematic review gathered all published studies regarding how parkin behaves in gliomas and evaluated genetic and molecular associations presented in this interaction.

Some studies focused on analyzing mutation patterns in PRKN, located in 6q26, and deletions being the most common. Four out of 8 studies described complex genomic alterations in PRKN locus. Mulholland et al. (2006) found deletions in 6q25–27, in GBM cell lines and GBM samples affecting not only PRKN, but also Insulin-like growth factor 2 receptor (IGF2R), Parkin coregulated gene protein (PACRG) and Quaking homolog, KH domain RNA binding (QKI). Yin et al. (2009) had similar findings and showed that deletions in long arm of chromosome 6 were observed in about one third of the GBM samples evaluated, impacting on PRKN, PACRG, QKI and Phosphodiesterase 10A (PDE10A). Veeriah et al. (2010) analyzed GBM samples and detected deletions in 6q chromosome, affecting PRKN in 85% of the cases. Yeo et al. (2012) evaluated a public glioma database REMBRANDT (Madhavan et al. 2009) and Freje (Freije et al. 2004) and, like the previous studies, detected several mutations, including loss and gain, in PRKN locus in chromosome 6, and an inverse correlation between parkin expression and tumor grade.

Viotti et al. (2014) evaluated parkin expression in three different gliomas: oligodendrogliomas, oligoastrocytomas and GBM, and also observed an inverse correlation between parkin and tumor grade. Lin et al. (2015) also evaluated REMBRANDT database (Madhavan et al. 2009) and found that the frequency of deletions in PRKN locus was higher in glioblastomas compared to low-grade gliomas, which correlates to a lower parkin expression in high-grade gliomas.

One of the main associations regarding low parkin expression is increased tumor proliferation. Some authors evaluated this impact by measuring the number of colonies in in vitro studies or measuring the tumor (size and weight) using in vivo models (Yeo et al. 2012; Lin et al. 2015; Veeriah et al. 2010). Veeriah et al. (2010) showed that transfecting WT-PRKN into PRKN negative-GBM cells resulted in the reduction of colony formation. Yeo et al. (2012) also found that GBM cells with PRKN restoration increased parkin expression and reduced colony proliferation in agar plate, compared to control group. Lin et al. (2015) had similar findings, as WT-PRKN GBM cells had lower number of colonies, when compared to mutated PRKN.

Still regarding tumor growth, some authors investigated how parkin interacts with cyclins. Veeriah et al. (2010) showed that cyclin E expression was lower in WT-PRKN cells. This happened because there was as interaction between parkin and cyclin E, followed by cyclin degradation. When they knocked out PRKN in these cells, cyclin E expression increased. Yeo et al. (2012) and Lin et al. (2015) had the same response regarding cyclin D1, but not cyclin E.

In vivo assessments were also performed and showed compatible results to in vitro experiments. Veeriah et al. (2010) injected GBM cell line either having WT or mutated PRKN in murine. Most animals injected with PRKN mutations developed tumors, whereas the same happened in only 20% of the animals injected with WT-PRKN. As for tumor size, it was greater in mutated PRKN animals. Yeo et al. (2012) had the same results, showing greater tumor size and weight in the mutated PRKN group. Lin et al. (2015) also showed smaller volume in the mutated PRKN group, compared to control. These findings in in vivo studies correlate PRKN mutation with reduced tumor burden, which may indicate a tumor suppression function.

Survival rates were also assessed by these authors. Lin et al. (2015) had an extra group of mice followed up and detected higher mortality rates in control group. Therefore, functional PRKN tumors were smaller, and animals had better survival rates. Yeo et al. (2012) analyzed parkin expression in REMBRANDT and Freje public glioma databases (Madhavan et al. 2009; Freije et al. 2004) and results showed that higher parkin expression indicated better prognosis. Considering the role played by parkin in protein degradation, these findings suggest that absent or decreased parkin expression results in less cyclin degradation and, therefore, cell cycle progression. It is known that cyclin expression is increased in many tumors. Maybe the reason behind it in gliomas is PRKN mutation and less cyclin degradation.

As for parkin associations with other molecular pathways, we have different approaches by the authors. Yeo et al. (2012) evaluated PTEN pathway. Low parkin expression resulted in Akt phosphorylation, which increased cyclin D1. Molecules involved in Akt phosphorylation, such as mechanistic target of rapamycin (mTOR), Rapamycin-insensitive companion of mammalian target of rapamycin (Rictor), and Linsidomine (Sin1) as well as their associated substrates, such as Forkhead box O3 (FoxO3a) and Glycogen synthase kinase 3 beta (GSK3b) also had inverse correlation to parkin expression. The same happened in fibroblast cell lines, showing that it is not specific for glioma cells. Thus, low parkin expression increased indirectly cyclin D1 levels. This study suggests that molecules involved with PTEN pathway are associated with proliferation and tumor growth.

Lin et al. (2015) analyzed two different molecular pathways and how parkin interacts with them. The first was Wnt-beta-catenin. Molecules involved in this pathway, such as cyclin D1, c-myc and Transcription factor 4 (TCF4) were downregulated in WT-PRKN GBM cells, both in messenger Ribonucleic acid (mRNA) levels and protein expression, compared to control cells. When evaluated in Wnt3A medium, which stimulates the Wnt pathway, WT-PRKN cells showed lower levels of Axin-like protein (AXIN2), which is a Wnt target, compared to control cells. Also, parkin expression increased in this medium, indicating a possible feedback between PRKN and Wnt-beta-catenin. Beta-catenin levels were also evaluated. In silenced PRKN, beta-catenin expression increased, and after PRKN restoration, beta-catenin levels reduced. Hence, they postulated that parkin induces beta-catenin degradation via proteasome. The second molecular pathway evaluated by Lin and colleagues was EGFR-Akt. There is also an inverse correlation between parkin and EGFR expression. After silencing PRKN, EGFR expression increases. When cell lines were stimulated by EGF, control cells increased EGFR expression. Perhaps, like beta-catenin, parkin induces EGFR degradation. Thus, parkin inhibits tumor growth by acting on Wnt and EGFR pathways. The last two studies (Gong et al. 2017; Cesari et al. 2003) showed how parkin interacts in different molecular pathways that gain importance with a potential therapeutic target. This search becomes particularly important, since GBM therapy is still a challenge.

Viotti et al. (2014) established a connection between p53 and parkin. It is well known that high grade gliomas harbor Tumor protein p53 (TP53) mutations, causing overexpression of p53 protein. They evaluated p53 and parkin expression in gliomas and observed an inverse correlation between parkin and p53, and between parkin and tumor grade. In order to evaluate if p53 overexpressed in high grade gliomas was inactive, they evaluated bax and bcl2 expression: the former is stimulated by p53 and the latter, inhibited. Samples showed reduced bax and increased bcl2 expression, indicating that, even though p53 was expressed, it was not functional. GBM cell lines were evaluated and showed lower mRNA and parkin expression levels. And when TP53 was restored, parkin expression increased. The same results were observed in vivo. WT-TP53 and silenced TP53 glioma cells were injected in mice. Parkin expression and mRNA levels were higher in WT-TP53. When knocked out TP53 was restored, parkin expression increased. These are interesting findings because it could mean that not all high-grade gliomas have low parkin expression due to PRKN mutations. Some could present this feature because of TP53 mutation, which impacts on parkin expression.

Two studies based on the fact that GBM is an aggressive neoplasia with highly invasive behavior. Maugeri et al. (2016) focused on GBM’s hypoxic areas, which induce the expression of HIFs, stimulating pro-angiogenic factors and greater invasiveness. Hypoxic situations induce HIF1a/HIF3a expression, which contributes to neoangiogenesis and cell viability. Using GBM cell lines the authors compared parkin expression in normoxic and hypoxic situations. The main result showed that parkin expression was reduced in hypoxic situation. Then they evaluated HIF1a and HIF3a expression, in the presence and absence of parkin, in the context of normoxia and hypoxia. In the presence of normoxia and parkin, the hypoxia-inducible factors present opposite characteristics (low expression for HIF1a and high for HIF3a). After this the PRKN gene was silenced, and the expression of the hypoxia-inducible factors was changed (high expression for HIF1a and low for HIF3a). In the context of hypoxia, the presence of parkin was associated with high expression for both factors (HIF1a and HIF3a). After silencing PRKN, expression was surprisingly reduced in both proteins. So, they postulated that parkin interferes on HIF3a, so that HIF1a is overexpressed in hypoxic situations. This study showed increased HIF1a in GBM cells expressing parkin in low oxygen situations. Could this mean that in GBMs that do not harbor PRKN mutations, HIF1a contributes to a poorer prognosis, since it contributes to neoangiogenesis and cell viability? Future studies may answer this question.

Wang et al. (2017) evaluated how parkin influenced invasiveness and metastasis. This is regarded as the first study associating PRKN mutations and tumor invasion and metastasis. Wang et al. induced parkin overexpression in a GBM cell line using lentiviral construction. When parkin normal expressing and overexpressing cells were compared, the latter had less migration and invasion capacity. They also showed that silenced PRKN GBM cells, compared to cells with functional gene, had higher migration and invasion rates. Next, they analyzed ZEB1 expression in these two cell lines. It is known that ZEB1 has increased expression in high-grade gliomas. ZEB1 expression was lower in GBM cells overexpressing parkin, compared to silenced PRKN cells. Thus, parkin downregulates ZEB1. They also evaluated EMT (epithelial-mesenchymal transition) markers: vimentin and e-caderin. Vimentin expression was lower and e-caderin expression was higher in cells overexpressing parkin, while silenced PRKN cells showed the opposite. ZEB1 stimulates these markers because, once these cells had ZEB1 knocked out, e-caderin increased and vimentin decreased, which means parkin controls ZEB1, which impacts on EMT markers.

This systematic review could have some limitations. The number of GBM cell lines available is wide and not all authors used the same cell line. This could have some interference in results since different cell lines have different parkin expression levels, as demonstrated by Maugeri et al. (2015). The number of studies is also limited, and as new studies become available, future systematic reviews may show other molecular associations with greater impact in GBM therapy.

Conclusion

This systematic review showed that the frequency of different mutations in the PRKN locus is higher in high-grade gliomas. Parkin expression is inversely correlated with glioma grade. Higher parkin expression and lower proliferation rates were observed in low-grade gliomas. Parkin expression was also associated with better prognosis and outcome. Besides, it interacts with different molecular pathways, which could turn out to be therapy targets in the future. In conclusion, based on these studies, we can infer that parkin acts on reducing cell proliferation, reducing cell migration and invasiveness and improving prognosis, through different pathways.