Linking Parkinson’s Disease and Melanoma: Interplay between α-Synuclein and Pmel17 Amyloid Formation

Abstract

Parkinson’s disease (PD) is a neurodegenerative disorder associated with the death of dopaminergic neurons within the substantia nigra of the brain. Melanoma is a cancer of melanocytes, pigmented cells that give rise to skin tone, hair and eye color. While these two diseases fundamentally differ, with PD leading to cell degeneration and melanoma leading to cell proliferation, epidemiological evidence has revealed a reciprocal relationship where PD patients are more susceptible to melanoma and melanoma patients are more susceptible to PD. The hallmark pathology observed in PD brains are intracellular inclusions, of which the primary component are proteinaceous α-synuclein (α-syn) amyloid fibrils. α-Syn has also been detected in cultured melanoma cells as well as tissues derived from melanoma patients, where an inverse correlation exists between α-syn expression and pigmentation. While this has led to the prevailing hypothesis that α-syn inhibits enzymes involved in melanin biosynthesis, we recently reported an alternative hypothesis in which α-syn interacts with and modulates the aggregation of Pmel17, a functional amyloid that serves as a scaffold for melanin biosynthesis. In this perspective, we review the literature describing the epidemiological and molecular connections between PD and melanoma, presenting both the prevailing hypothesis and our amyloid-centric hypothesis. We offer our views of the essential questions that remain unanswered to motivate future investigations. Understanding the behavior of α-syn in melanoma could not only provide novel approaches for treating melanoma, but also could reveal insights into the role of α-syn in PD.

Introduction

The primary pathological hallmarks of Parkinson’s disease (PD),¹ a neurodegenerative disorder associated with the death of dopaminergic neurons in the brain, are intracellular inclusions containing α-synuclein (α-syn) protein aggregates, called Lewy bodies.² This aggregation process involves the conversion of natively unfolded α-syn monomers to β-sheet-rich fibrils known as amyloids.³ More broadly, protein aggregation and amyloid formation are involved in many neurodegenerative diseases, such as the aggregation of amyloid-β and tau proteins in Alzheimer’s disease, prion protein in Creutzfeldt-Jakob disease, and huntingtin protein in Huntington’s disease.⁴

Nearly 50 years ago, Skibba et al. reported the unusual development of multiple melanomas, a cancer of melanocytes, in a patient suffering from PD.⁵ While this was first attributed to levodopa administration, a common treatment for PD, additional scientific inquiry in the past 20 years has since solidified unique epidemiological and molecular connections between these two diseases (described in detail below). This leads to puzzling questions, given that PD and melanoma induce opposing cellular responses, namely cell death and cell growth, respectively.

In this perspective, we first highlight the literature which substantiate the epidemiological and molecular connections between PD and melanoma. Then, we describe the prevailing hypothesis that explains one of the observed links, and offer an additional, amyloid-centric hypothesis first described in our recent work.⁶ Finally, we conclude with outstanding questions that need to be addressed in future investigations to substantiate our amyloid-centric hypothesis.

Epidemiological Connections Between PD and Melanoma

Development of melanoma among individuals with PD

Since the initial report by Skibba et al. in 1972, several epidemiological studies have been conducted to evaluate cancer incidence among patients suffering from PD. Generally, the overall cancer incidence and mortality rates among PD patients are lower than expected, with standard incidence/mortality ratios (SIR/SMR), odds ratios (OR), or relative risk estimates (RR) ranging from ~0.3–0.9.^7–15 However, some cancers such as melanoma,^{7, 11, 13, 15} nonmelanocytic skin cancer,^{11, 13, 16} breast cancer,^{11, 13} brain cancer,¹³ and prostate cancer¹² are found to be higher than expected in the PD population. Numerous investigations have focused on melanoma, reporting an increase in incidence ranging from 1.4–20-fold among individuals with PD compared to healthy patients (Table 1).^{10–15, 17–25}

Table 1.

Incidence of melanoma among PD patients

Size of PD Cohort	Observed Incidence	Expected/Control Incidence	Fold-Change	Reference
107,598a	NA	NA	1.41+	Bajaj et al.10
20,343	65	46	1.41%	Rugbjerg et al.11
8,090	46	128b	1.44^	Olsen et al.17
196	3	2	1.5^	Elbaz et al.12
692	6	5	1.6^	Lo et al.18
70,730	23	70b	1.66&	Ryu et al.25
14,088	44	22.5	1.95%	Olsen et al.13
2,998	48	24.6	1.95+	Kareus et al.19
188	9	21b	2.72^	Becker et al.14
800	5	1.5	3.3%	Constantinescu et al.20
974	26	21	3.8#	Dalvin et al.21
14,088	39	-	4.6^	Olsen et al.22
487	9	0	6.15+	Driver et al.15
2,106	24	3.86	6.2+	Bertoni et al.23
806	9	0.43	20.9^	Schwid et al.24

^aThis study is a meta-analysis of 29 individual studies.

^bThe total number of control cases analyzed exceeds the total number of PD cases analyzed.

(NA) not applicable,

^(-)unknown/not reported,

⁽⁺⁾relative risk,

^{(^)}odds ratio,

^(#)logistic regression analysis,

^(%)standard incidence ratio,

^(&)hazard ratio

Initial reports suggested that the increased incidence of melanoma among the PD population could arise due to L-3,4-dihydroxyphenylalanine (L-DOPA or levodopa) administration, given that L-DOPA is intricately involved in both dopamine and melanin biosynthetic pathways in dopaminergic neurons and melanocytes, respectively.²⁶ To address this, Olsen et al. published a follow up study to determine melanoma incidence prior to PD diagnosis (and therefore prior to levodopa treatment), and found it was still increased compared to control populations (OR = 1.44).¹⁷ This was later supported by similar studies which have shown higher rates of melanoma among PD patients regardless of whether patients received levodopa treatment or not.^{20, 22, 24, 27} Collectively, these studies show that melanoma occurs more frequently among the PD population, which cannot be explained by the administration of levodopa for PD symptoms, pointing to a genetic and/or molecular link between the two diseases.

Development of PD among individuals with melanoma

The increased incidence of melanoma among PD patients immediately leads to question if the reverse is also true, that is whether or not melanoma patients are at a higher risk of developing PD than healthy individuals. Among individuals with a family history of melanoma (first-degree relative), Gao et al. found that the RR of developing PD was 1.85-fold higher than those without a family history of melanoma, whereas no association with PD was observed among other cancer types (Table 2).²⁸ This observation is supported by another study, where melanoma patients were found to be at 4.2-fold increased risk of developing PD (Table 2).²¹ Epidemiological studies investigating PD mortality among cancer patients found a 1.5–3-fold higher SMR among melanoma patients compared to those with other types of cancer (Table 2).^{19, 29–31} Thus, current data suggest that the epidemiological connection between PD and melanoma is reciprocal, where PD patients are more likely to develop melanoma and melanoma patients are also more likely to develop PD, again pointing to an underlying genetic and/or molecular link.

Table 2.

Incidence of PD among melanoma patients.

Size of Melanoma Cohort	Observed Incidence	Expected/Control Incidence	Fold-Change	Reference
1,985	87	59	1.5$	Freedman et al.29
7,841	48	29.1	1.65+	Kareus et al.19
3,389	24	592a	1.85+	Gao et al.28
24,098	12	-	1.99$	Baade et al.30
127,037	126	-	2.7$	Baade et al.30
1,544	43	14	4.2&	Dalvin et al.21

^aThe total number of control cases analyzed exceeds the total number of melanoma cases analyzed.

^(-)unknown/not reported,

^($)standard mortality ratio,

⁽⁺⁾relative risk,

^(&)Cox proportional hazards model

Genetic Links Between PD and Melanoma

While the epidemiological link between PD and melanoma is now well established, fewer studies have investigated potential underlying genetic links. Recent review articles have discussed these potential genetic links,^{26, 32–39} however we feel it is important to highlight a few articles here. Recently, a meta-analysis of four individual genome-wide association studies (one melanoma and three PD) was performed and found a positive genetic correlation between melanoma and PD, but not melanoma and other neurodegenerative disorders.⁴⁰ This work also identified 606 PD-associated and 168 melanoma-associated genes, of which seven were above the false discovery rate threshold for both PD and melanoma.⁴⁰ Inzelberg and co-workers performed whole exome sequencing of 246 cutaneous melanoma tissue samples, specifically investigating whether somatic mutations in familial PD genes (PARK genes) were present.⁴¹ In all, somatic mutations were found in 14 of the 15 PARK genes analyzed, with ~50 % of melanoma tissue samples containing at least one PARK gene mutation.⁴¹ The highest rates of mutations were found in PARK8 (LRRK2, present in 16.2 % of samples), PARK2 (PRKN, 12.6 %), PARK18 (EIF4G1, 12.6 %), and PARK20 (SYNJ1, 11.4 %).⁴¹ This is supported by other studies which have found increased rates of PARK2 (PRKN) mutations in melanoma tissue,^{42, 43} a gene which encodes parkin, a protein shown to act as a tumor suppressor.^{43, 44} Thus, loss-of-function parkin mutations associated with familial PD may also promote melanoma progression.⁴³ While these studies highlight a potential genetic factor between PD and melanoma that should be explored further, mutations in PARK genes only account for ~5% of PD cases,⁴⁵ suggesting the prevalence of other molecular links.

α-Synuclein: A Molecular Link Between PD and Melanoma

Cutaneous detection of α-synuclein within melanocytes

The aggregation and deposition of α-syn within Lewy bodies is a pathological hallmark of PD, where post-translational phosphorylation at S129 has been identified as a biomarker of aggregated α-syn.⁴⁶ In recent years, the deposition of S129 phosphorylated α-syn (pS129 α-syn) within peripheral tissues has been observed and proposed as a pre-symptomatic diagnostic tool for PD.⁴⁷ Along these lines, cutaneous detection of pS129 α-syn from skin biopsies has been explored extensively,^48–63 and may prove to be one of the more accurate approaches for PD diagnosis.⁴⁷ Tsukita et al. performed a systematic review and meta-analysis of 41 primary literature articles which examined immunogenic detection of α-syn in skin tissue samples, finding that detection of phosphorylated α-syn as opposed to non-phosphorylated α-syn had enhanced selectivity in identifying PD vs. healthy patients.⁴⁷ Additional studies have shown that α-syn inclusions found in the skin of PD patients are resistant to protease degradation, a hallmark of α-syn amyloids,⁶⁰ and that ex vivo material from skin tissue of PD patients, but not healthy controls, can template recombinant α-syn amyloid formation in vitro via real-time quaking-induced conversion (i.e., seeding) assays.^{64, 65} Based on these results, it is concluded that pathogenic α-syn amyloid deposits are found within the skin of PD patients.

α-Syn has also been detected within specific cell types found in the skin, including fibroblasts,^{66, 67} keratinocytes^{48, 59, 68} as well as melanocytes^{54, 68} from PD patients. The latter is interesting given the aforementioned epidemiological connection between PD and melanoma. Rodriguez-Leyva et al. recently compared the deposition of non-phosphorylated α-syn within the epidermal layer of healthy, PD, and melanoma patients and found that α-syn deposition was scarcely found within melanocytes of healthy patients (accounting for ~1% of the analyzed area), but increased within both melanocytes and keratinocytes of PD (~3% of the analyzed area) and melanoma (~14% of the analyzed area) patients.⁶⁸ This is supported by gene expression analysis, which found that SNCA expression was ~3-fold higher in metastatic growth phase melanomas compared to healthy controls⁶⁹ as well as immunoblotting of primary and metastatic melanoma tissue samples, where 86% and 85% were positive for α-syn, respectively.⁷⁰ Cultured melanoma cell lines have also been used to evaluate endogenous α-syn expression, where varying levels of α-syn are detected when comparing various cell lines under identical experimental conditions (Table 3).^69–74 Altogether, this reveals that α-syn is endogenously expressed within melanocytes of both PD and melanoma patients, and thus, could represent a molecular link between the two diseases.

Table 3.

Endogenous expression of α-syn in melanoma cell lines.

Cell Line	Relative Expressiona	Referenceb
A-375	+	Lee et al.74
WM266–4	++
SK-MEL 28	++
MeWo	+++
SK-MEL 5	++++
WM1158	+	Turriani et al.69
WM852	+
SK-MEL 5	++
WM983-A	+++
WM983-B	+++

^aRelative expression was qualitatively determined by comparing band intensities, where ++++ > +++ > ++ > + and comparison between cell lines from different references is not appropriate.

^bWhile several studies have reported the endogenous expression of α-syn in melanoma cell lines, these two studies are highlighted because they compared α-syn expression between various cell lines under the same conditions. Epitopes for antibodies used were mapped to α-syn residues 115–122 (LB509, Lee et al.⁷⁴) and 118–123 (MJFR1, Turriani et al.⁶⁹).

Inverse correlation between α-syn and pigmentation

While the presence of α-syn within melanoma tissue and cells is well characterized, its role within melanoma cells is ill-defined. One of the primary functions of melanocytes is to synthesize eumelanin and pheomelanin (collectively termed melanin for simplicity), pigmented compounds which give rise to skin tone, hair and eye color.⁷⁵ Matsuo and Kamitani analyzed 17 pigmented melanoma tissue sections from patients, finding that cells with increased amounts of α-syn had little to no pigmentation,⁷⁰ an observation later supported by a similar study.⁶⁸ In addition, Pan et al. showed that knocking down SNCA expression in hypopigmented SK-MEL 28 cells increased overall melanin production, while over-expressing α-syn in pigmented A375 cells decreased overall melanin production.⁷² This consistent observation suggests that α-syn acts to disrupt melanin production in melanoma cells leading to amelanotic (non-pigmented) melanomas which evade early detection. This, combined with the fact that melanin serves to absorb carcinogenic UV radiation,⁷⁶ the disruption of melanogenesis by α-syn may play a role in the development and/or progression of melanoma.

Current Hypotheses

α-Syn disrupts tyrosine hydroxylase and/or tyrosinase activity

The prevailing hypothesis^{26, 32–34, 37, 39} to describe the inverse correlation between α-syn expression and melanin production states that α-syn disrupts the activities of catalytic enzymes involved in melanin biosynthesis, namely tyrosine hydroxylase (TH) and tyrosinase (TYR), which is based on a few primary research studies.^{72, 77, 78} In dopaminergic brain cells, α-syn has been shown to co-immunoprecipitate with TH and negatively regulate its activity by activating protein phosphatase PP2A, the enzyme responsible for dephosphorylating and inactivating TH.^{77, 78} While TH is involved in both dopamine and melanin biosynthetic pathways, TYR is considered the key regulatory enzyme in melanogenesis.⁷⁹ Pan et al. showed that over-expression of α-syn in pigmented A375 cells resulted in an apparent decrease in TYR activity, estimated by measuring L-DOPA oxidation in vitro, and knocking down α-syn expression in hypopigmented SK-MEL 28 cells resulted in an apparent increase in TYR activity.⁷² While this does show an apparent correlation between α-syn expression and TYR activity, work from Watabe et al. has shown that in SK-MEL 28 cells, TYR is not trafficked to the site of melanin biosynthesis, the melanosome, but rather is rapidly degraded by the proteosome due to impaired glycosylation.⁸⁰ This led us to question if other mechanisms may also be at play, namely whether α-syn impacts another important pigmentation gene, PMEL.

α-Syn modulates Pmel17 functional amyloid formation

The PMEL gene encodes the pre-melanosomal protein (Pmel17, also known as gp100), which forms a fibrillar matrix that serves as a scaffold for melanin polymerization within early-stage melanosomes.⁸¹ It is now recognized that these Pmel17 fibrils are amyloidogenic in nature, making them a so-called “functional amyloid” because they serve a biological purpose.^{82, 83} Given that amyloid fibrils share similar morphological and structural features, interplay between amyloidogenic proteins is proposed to occur when there is overlap of two amyloid-associated pathologies, such as with Alzheimer’s disease and type-II diabetes^84–90 and other examples.^91–98 Thus, we reasoned that the disruption of melanin biosynthesis by α-syn in melanoma cells could involve the two amyloid-forming proteins, α-syn and Pmel17.

To test this hypothesis, we first determine whether α-syn localizes to the melanosome, where Pmel17 resides and the site of melanin biosynthesis. Using SK-MEL 28 cells and confocal immunofluorescence, we showed localization of endogenous α-syn to melanosomes, which were represented by Pmel17-positive cytoplasmic punctates (Fig. 1a).⁶ This was confirmed by immunoblotting for α-syn in a subcellular fraction enriched in melanosomal proteins (Fig. 1b, M). For the first time, our results revealed that α-syn localizes to the melanosome, the site of Pmel17 functional amyloid formation.

Figure 1: α-Syn localizes to the melanosome and modulates Pmel17-RPT aggregation.

(a) Immunogenic detection of α-syn and Pmel17 in SK-MEL 28 cells. Co-localized pixels are shown in white, individual channels not shown. (b) Immunoblots of SK-MEL 28 total cell lysate (T), and fractions enriched in cytoplasmic (C), nuclear (N), and melanosomal (M) proteins. (c) Schematic representation of in vitro experiments using the amyloidogenic repeat domain (RPT) of Pmel17, where α-syn fibrils (gray rectangles) stimulate RPT (black circles) aggregation toward the formation of α-syn-like twisted filaments (black squares). In the absence of α-syn and presence of RPT fibrils (gray rounded rectangles), rod-like filaments with no observable twists are observed (black rounded squares). Representative negatively-stained transmission electron microscopy images of cross-seeded (left) and self-seeded (right) fibrils are also shown, scale bars represent 10 nm. Data were first reported in Dean and Lee.⁶

Next, we investigated whether α-syn influenced Pmel17 amyloid formation in vitro using an amyloidogenic fragment of Pmel17 spanning the repeat region (RPT, Pmel17 residues 315–444), which has been previously characterized extensively by our group.^99–106 For a full discussion on RPT amyloid formation, we refer the reader to two recently published review articles.^{82, 83} Given the aforementioned data supporting the presence of α-syn amyloid deposits in the skin of PD patients, cross-seeding experiments were performed to evaluate the effect of pre-formed α-syn fibrils on RPT amyloid formation. We found that α-syn fibrils stimulated RPT aggregation to form fibrils with a twisted architecture similar to that formed by α-syn (Fig. 1c, left). The twisted fibrils formed by cross-seeding are unique from that formed by RPT on its own, which have a rod-like architecture (Fig. 1c, right) similar to that observed in early-stage melanosomes.¹⁰⁷ This suggests that RPT can adopt an α-syn-like amyloid fold and aggregate by templating from the ends of α-syn fibril seeds, which should be explored in future work using structural biology tools. Importantly, RPT fibrils had no influence on α-syn aggregation, revealing that the process is uni-directional. Additional characterization of cross-seeded fibrils revealed changes to the secondary structure akin to α-syn, as well as changes to the overall protease accessibility suggesting an altered fibril conformation. Altogether, our work reveals how the presence of α-syn in the melanosome has the potential to modulate Pmel17 functional amyloid formation (Fig. 2), an important step in melanogenesis, which in turn may alter melanin production leading to the observed amelanotic phenotype in melanoma cells and tissue.

Figure 2: α-Syn disrupts melanogenesis in human melanoma cells.

Melanocytes synthesize melanin within melanosomes, which undergo four stages of maturation. In stage II, Pmel17 forms amyloid fibrils which serve as a scaffold for melanin biosynthesis in stages III and IV. The prevailing hypothesis suggests that α-syn disrupts melanin biosynthesis by inhibiting the enzymes involved in melanin production, and herein described is an alternative hypothesis that involves the modulation of Pmel17 aggregation by α-syn.

The question remains whether the modulation of Pmel17 aggregation by α-syn translates to an impact on melanin synthesis. Work by McGlinchey et al. suggests that the specific fibril architecture formed may not matter, as amyloid fibrils formed by HET-s and Sup35NM enhanced melanin biosynthesis similar to RPT fibrils.¹⁰⁸ However, these studies were conducted in vitro using TYR isolated from mushrooms, and thus may not best reflect conditions found within the melanosome of human melanocytes. Additionally, while we explored the impact of α-syn fibrils on RPT aggregation, it remains to be determined as to whether α-syn is in an aggregated state within the melanosome, which warrants further study of the impact of soluble α-syn on RPT aggregation. Thus, while our initial work is promising and provides a unique perspective on this topic, clearly more work is needed to fully address the relationship between α-syn and Pmel17 in regards to pigmentation in melanoma cells.

Future Perspectives

A large number of studies, including ours, utilizes SK-MEL 28 cells to investigate the role of α-syn in melanoma. While we used this cell model to simply address α-syn localization, there is an inherent flaw in using this cell line because TYR trafficking to the melanosome is disrupted,⁸⁰ as discussed earlier. We therefore question whether the observed amelanotic phenotype in SK-MEL 28 cells has any relation to α-syn despite its wide utilization in the field and thus, raise the concern that more relevant systems are needed to address these biological questions regarding α-syn and its impact on melanogenesis. Additionally, while numerous transgenic animal models are available for both PD and melanoma, no efforts have been made to study melanoma among PD animal models and vice versa. For example, would neonatal transgenic mice which overexpress α-syn be more susceptible to UV-induced melanoma¹⁰⁹ compared to non-transgenic or SNCA knockout mice? Developing such cellular and animal models will be essential in advancing our understanding of how α-syn impacts the molecular events of melanogenesis.

There is also a major need to more broadly investigate how α-syn may promote melanoma cell survival. Specifically, work should be done to address whether α-syn is ubiquitously expressed in healthy melanocytes or only found in the cancerous state. Some groups have reported detection of α-syn within the skin of healthy individuals,^{51, 54, 56, 58, 59, 68, 110–112} while others have not.^{48, 70, 113} Turriani et al. reported a basal level of SNCA expression within healthy skin tissue, which was increased in vertical growth phase and metastatic growth phase melanoma tissue, the most aggressive stages of melanoma growth and progression.⁶⁹ Additionally, work from Shekoohi et al. has shown that knocking out SNCA in SK-MEL 28 cells suppressed tumor growth.¹¹⁴ Does this suggest that α-syn is ubiquitously expressed in melanocytes, where it normally acts to promote cell survival but when upregulated leads to uncontrolled growth i.e., melanoma?

Along these lines, we reported endogenous detection of α-syn within nuclei of SK-MEL 28 cells (Fig. 1b), an observation supported by other studies also showing nuclear localization of α-syn in melanocytes/melanoma.^{68, 69, 74, 110, 115} While not in the context of melanoma, α-syn has been shown to localize to nuclei of neuronal-like cells and modulate gene expression by interacting with nucleosomes,¹¹⁶ histones,^117–120 and regulatory regions of genes.^{121, 122} In particular, Pinho et al. recently reported the differential expression of cell cycle genes as a result of nuclear α-syn localization.¹¹⁶ Given that both melanocytes and dopaminergic neurons are derived from the neural crest,¹²³ this leads us to ask if α-syn could behave similarly in melanocytes and melanoma. In addition, future studies should also focus on understanding how autophagy is impacted by α-syn, as it has been suggested to promote cell survival in melanoma^{69, 115} yet promote cell death in PD.¹²⁴

To conclude, we have described the unique epidemiological and molecular connections between PD and melanoma, detailing how α-syn could act to disrupt melanogenesis in melanoma cells by a multi-pronged approach of disrupting enzymes involved in melanin biosynthesis (TH and TYR) as well as modulating Pmel17 functional amyloid formation (Fig. 2), both of which warrant further investigation. Moreover, we have discussed how future work should investigate how α-syn influences melanoma cell survival, as this may reveal new insights into α-syn function which may be utilized to prevent dopaminergic cell death and subsequently PD.