Abstract
There is substantial evidence based on well designed epidemiologic studies for low cancer rates in patients with Parkinson disease (PD). This risk reduction cannot be attributed to the recognized low life-long incidence of smoking in patients with PD, as not only smoking-related cancers but also non-smoking-related ones are less common in PD. Whereas the risk for most cancers appears to be relatively low in patients with PD, breast cancer and melanomas occur more frequently in the PD population as compared with controls. The relationship between this peculiar pattern of cancer rates and PD might be related to the involvement of common genes in both diseases. Mutations in parkin gene, for example, have been reported in several types of cancer. Furthermore, genes involved in familial forms of PD appear to be abnormally expressed in cancers. Thus, parkin and PINK1 might be tumor suppressor genes, whereas DJ-1 is an oncogene. Cell survival signals may differ owing to mutated genes and represent two opposite extremes such as cell proliferation in cancer and cell death due to apoptosis in PD. Unraveling the link between PD and cancer may open a therapeutic window for both diseases.
DOES PARKINSON DISEASE PROTECT AGAINST CANCER?
The notion that Parkinson disease (PD) provides some type of “biologic protection” against certain types of cancers has been raised in the last few decades, but no scientific rationale has been offered to explain possible mechanisms for this observation. As early as 1954, Doshay noted that, for reasons as yet unclear, cancer is rare in “paralysis agitans.”1 Hoehn and Yahr studied the cause of death in 194 patients with PD and found that 24 died of malignant neoplasms as compared with an expected number of 41 using the New York population (p = 0.001).2 Jansson and Jankovic3 studied 406 patients with PD and found that only 18 developed cancer as compared with an expected number of 41.9 and only 10 skin cancers were observed as compared with an expected number of 49.9 (p < 0.0001). All cancers with the exception of thyroid cancer and melanoma had lower rates in PD patients as compared with the control population. Later, other cohort studies provided additional evidence of reduced risk of cancer in patients with PD.4,5 Moller et al.4 found an overall reduction in cancer risk in PD with a risk ratio of 0.88 (0.8 to 1.0). They noted that especially lung and bladder cancers were rare, whereas melanomas were more common.
Mortality studies have shown that the risk of dying from cancer is lower in patients with PD than in the general population,5–8 although one early postmortem study9 found doubling of mortality from cancer in PD patients (7 of the 111 PD deaths were due to carcinoma, whereas the expected number was 3.6), but the sample was too small and there were other methodologic problems that did not allow definite conclusions.
In a cohort of 10,322 PD patients identified between 1987 and 1990 in Italy via pharmacy claims for antiparkinsonian medications, the risk of cancer death, as estimated by standardized mortality ratio (SMR) was 56 (95% CI 51 to 61), 44% lower than that expected SMR in the general population.8 This risk reduction could not be attributed to the well recognized low life-long incidence of smoking among PD patients10–17 because the reduction in mortality from cancer was true for both smoking-related (SMR 51, 95% CI 42 to 60) and nonsmoking-related (SMR 58, 95% CI 52 to 65) cancers.8 In a large study of 14,088 patients with PD, overall 1,282 cancers were detected, whereas 1,464 were expected, resulting in a standard incidence ratio (SIR) of 0.88 (95% CI 0.8 to 0.9), equivalent to a 12% reduction in the risk of cancer.18 Even lower SIR (SIR = 0.81, 0.8 to 0.9) was calculated after excluding melanoma and nonmelanocytic skin cancer, both thought to occur more frequently in PD. Although the reduction in SIR was due to diminished risk of smoking-related cancers (combined SIR = 0.58, 0.5 to 0.6), such as lung, larynx, urinary bladder, buccal cavity pharynx, esophageal neoplasms, myeloid leukemia, liver and cervix uteri, stomach, pancreas, and kidney, cancers presumably not related to smoking such as prostate (SIR = 074, 0.6 to 0.9), colon (SIR = 0.84), and rectum (SIR = 0.89) also occurred at lower than expected rates. The frequency of all cancers before the diagnosis of PD was found to be low compared with the general population in another study (odds ratio [OR] = 0.79; 95% CI 0.49 to 1.27), particularly in men and in patients with younger age at onset of PD.19 The risk of nonmelanocytic skin cancer, however, significantly increased after the diagnosis of PD.20
DOES PD INCREASE RISK OF CANCER?
Although most cancers appear to be less common, a few cancer types, including malignant melanoma, thyroid, and breast carcinoma, have been reported to occur with increased rates in patients with PD (table 1).3–5,18,21 The SIR for melanoma was calculated as 1.95 (95% CI 1.4 to 2.6) and 1.25 (1.1 to 1.4) for other skin cancers.18 The risk of melanoma was particularly high during the 5 years after PD onset but may be observed even before any motor signs of PD become apparent.21 The frequency of melanoma was also found to be higher than expected in the DATATOP trial cohort.22 The DATATOP cohort included 800 patients enrolled between September 1987 and November 1988 and followed until 1994. There were 5 cases of melanoma as compared with an expected number of 1.5 after adjusting for age and gender. The standardized event ratio (SER) was 3.3 (95% CI 1.1 to 7.8). A recent multicenter study has systematically examined the prevalence of PD among patients with malignant melanoma: 862 melanoma patients and an equal number of controls were studied.23 Among patients with melanoma, 2.9% had PD as compared with 1.3% of control subject (p = 0.014). Analysis of subjects older than 60 years also showed a significantly higher proportion of PD patients in the melanoma group (7.1 vs 3.1%, p = 0.016). Bertoni et al.24 surveyed 2,106 patients with PD and found that the prevalence (5-year limited duration) of invasive melanoma in US patients with PD (n = 1,692) was 2.2-fold higher (95% CI 1.21 to 4.17) than in age- and sex-matched populations in the national Surveillance Epidemiology and End Results- US Cancer Statistics Review database. Although 85 to 90% of all melanomas are attributed to solar radiation,25 it is unlikely that patients with PD belong to a subgroup of a population with above-average outdoor activities and more lengthy exposure to sunlight and sunburns. However, rural farm life may be associated with high solar exposure. The lack or relationship between sun exposure and occurrence of melanoma in the PD population is suggested by the reports of ocular choroidal melanoma in patients with PD.26
Table 1 Studies on the association between PD and cancer
The mechanism of increased risk of melanoma in patients with PD is unknown, but treatment of PD with l-dopa has been implicated as a risk factor for the development of malignant melanoma.27 The potential theory linking l-dopa and melanoma is based on the fact that l-dopa serves as substrate for rate-limiting enzyme tyrosine hydroxylase (TH), which ultimately converts l-dopa to melanin. Melanocytes and melanoma tumor cells are rich in TH. Other possible mechanisms for potentially increased risk of melanoma with l-dopa therapy include immunosuppression by enhancing the secretion of melanocyte-stimulating hormone.28–30 Siple at al.31 reviewed the published literature and concluded that of the 34 reported cases, 9 had a diagnosis of melanoma before the onset of l-dopa therapy. Of the 23 patients in whom melanoma was diagnosed after l-dopa administration, the diagnosis was made 4 months to 17 years after the institution of l-dopa. The occurrence of a number of cases of melanoma before or relatively soon after the diagnosis of PD makes l-dopa an unlikely contributor to a significant proportion of cases of melanoma.
Fiala et al.30 analyzed the clinical characteristics of 43 cases from the literature and their own 11 cases. They also agreed that the evidence for any relationship between l-dopa and melanoma is lacking. Other studies have cast doubts on any relationship between l-dopa and increased risk of malignant melanoma,32–34 although malignant melanoma is still listed as a potential adverse effect of and contraindication to l-dopa therapy.
Several reports have suggested that the risk of breast cancer in PD appears to be greater than in the general population.3,4 A study of the risk of cancer in 228 Japanese patients under age 80 showed a significantly increased risk for breast cancer (SIR = 5.49), but the CI was wide (1.1 to 16.0).5 Olsen et al.18 found marginally higher rates breast cancer (SIR = 1.24, 95% CI 1.0 to 1.5). Detection bias might influence the risk of breast cancer in PD because these patients are regularly followed by physicians; therefore, breast lesions are likely to be detected earlier. A possible link between breast cancer5,35 and significantly elevated nocturnal peak of prolactin in PD patients36 has been suggested. Estrogen has been also suggested to play a role in increasing the risk of breast cancer in patients with PD. In ovarectomized mice, studies show that estrogen partly blocks metamphetamine-induced striatal dopaminergic depletion.37–39 Tamoxifen, a selective estrogen receptor modulator, presents a neuroprotective effect against methamphetamine and methoxy-phenyltetrahydropyridine-induced toxicity when used alone but abolishes estrogen's positive effects when combined with this hormone.40 We have observed PD patients whose initial symptom of metastatic carcinoma was marked exacerbation of parkinsonian symptoms, which became non-levodopa responsive (J. Jankovic, personal observation).
PD-RELATED GENES AND THEIR ROLE IN TUMORIGENESIS
Although underdiagnosis may be a possible explanation for the relative paucity of cancers in patients with PD, this is quite unlikely because the risk reduction is seen from the first notification of PD or even before the diagnosis of PD.19 Also, the observed excess of breast and skin cancers in PD patients is against the notion of underdiagnosis. However, it is possible that other cancers that usually require more extensive evaluation, which is less likely to be performed in patients who are disabled by PD, are detected less frequently.
The discovery of gene mutations or variations associated with parkinsonian disorders (table 2) and understanding their role in cell survival and cell death may unravel the relationship between PD and cancer. It may also lead to presymptomatic diagnosis and more effective genetic testing and counseling.41
Table 2 Parkinson disease (PD)–related genes and putative role in cancer
Parkin.
Mutations in parkin, located at 6q25.2-6q27 (PARK2, OMIM 602544),42 are the most common cause of inherited PD accounting for up to 49% of familial recessive early-onset PD cases.43–45
Parkin mutations have been also implicated in tumorigenesis, a multi-step process resulting from genetic alterations that drive the progressive transformation of normal cells into malignant derivatives.46 Tumor suppressor genes (TSGs) are defined as genetic elements whose loss or mutational inactivation allows cells to acquire a neoplastic phenotype. Loss of heterozygosity (LOH) within genetically defective chromosomal regions is considered as indicative of the presence of a putative TSG.47 Several studies support the existence of TSGs or a senescence-related gene on chromosome 6q, and deletions at 6q27, a locus for the parkin gene, have been found in patients with benign ovarian tumors.48 Cesari et al. found high frequency of LOH within the parkin genomic structure.49 They genotyped 20 malignant breast tumors and 20 ovarian tumors and found deletions in 6q25-q27 locus in 55% of analyzed cases. They identified two tumor markers both located within the introns 2 and 7 of the parkin locus. Subsequent analysis of parkin gene expression in a variety of human carcinomas found transcript levels to be reduced or absent in 70% of the samples examined. Thus, expression of parkin gene appeared to be down-regulated or absent in biopsies and tumor cell lines. They suggested that the LOH observed at chromosome 6q25-q26 may contribute to the initiation and/or progression of carcinoma by interacting with or reducing the expression of the parkin gene. Aberrant transcripts were found in 15% of ovarian tumor cDNAs. This region involves parkin exons 2 to 10, suggesting that tumor-specific transcripts are a result of genomic deletions. Translation of each of these altered transcripts will result in a prematurely terminated parkin protein. Interestingly, the chromosomal region 6q25-q27 shows frequent deletions in a wide spectrum of human neoplasms such as melanoma, ovarian cancer, breast cancer, lymphomas, and others.49 Deletions in parkin gene have also been found in hepatocellular carcinoma and in breast, ovary, and non-small-cell lung carcinoma. Denison et al.50 found that 66.7% of ovarian cancer cell lines and 18.2% of primary ovarian tumors were heterozygous for the duplication or deletion of one or more parkin exons. Additionally, 13% of non–ovarian tumor–derived cell lines were found to have a duplication or deletion of one or more parkin exons. Diminished or absent parkin expression was observed in most of the ovarian cancer cell lines when studies with antibodies were performed. Although this finding suggests that parkin is a TSG, it is not clear whether mutations in the gene, found in patients with PARK2, results in increased risk of cancer.
Picchio et al.51 examined 20 paired normal and non–small cell lung cancer samples for the presence of parkin alterations in the coding sequence and changes in gene expression. They found a common region of loss in the parkin/FRA6E locus and observed parkin expression in three of nine (33%) lung tumors. They then restored gene expression in the parkin-deficient lung carcinoma cell line by use of a recombinant lentivirus containing the wild-type parkin cDNA. Ectopic parkin expression had the ability to reduce in vivo tumorigenicity in nude mice.
Wang et al.52 observed that parkin protein expression was significantly decreased or absent in all 11 hepatocellular carcinoma cell lines compared with that in normal liver tissue. Parkin gene-transfected cells grew more slowly than vector-only transfectants and also showed increased sensitivity to apoptosis induced by cell-cycle inhibitors.
Other PD-related genes.
Kim and Mak53 proposed that cancer and PD are two pathologic processes signaling by one of two sets of opposing forces, namely, those that drive cell death in PD and those that promote cell survival in neoplasms. There are two major pathways responsible for the cell survival: One is mitogen-activated protein kinase (MAPK; Erk1/2) signaling pathway and the other is PI3K/Akt-dependent pathway. Several PD-related genes may have functional relationship to these pathways. The leucine-rich repeat kinase 2 gene (LRRK2) whose mutations were found responsible for PARK8 (OMIM 607060)54–58 encodes an MAPKKK protein,59 and unraveling the MAPK signal pathway in dopamine neurons and cancer cells may help in understanding relationship between PD and cancer.
Heat shock proteins (HSPs), a group of proteins acting as molecular chaperones, bind to denatured and misfolded proteins and participate in the removal of the wasted proteins. HSP70 indirectly inhibits cyt c/ATP-dependent activation of caspase-3 through its effect on apoptotic protease-activating factor 1 (Apaf-1)-mediated activation of caspase-9, the apoptotic pathway involved in PD.60 Many studies have provided evidence that HSP90 is at the core of the so-called cytosolic molecular chaperone complex, required for the proper function of many proteins. As a number of proteins relevant to tumor growth such as v-Src, Akt, Raf-1, Bcr-Abl, ErbB2, mutant p53, and hypoxia-inducible factor-1α are among its clients, HSP90 inhibitors such as radicicol promise to complement cancer therapies (figure).61 The tumor suppressor p53 inhibits cell cycle progression and mediates apoptosis. Several studies suggest that p53 expression may correlate with neuronal death in neurodegenerative diseases. As p53 regulates many mitochondria-related genes and oxidative stress–related genes,62 it may participate in the mitochondrial dysfunctions associated with PD. By contrast, inhibition of the tumor suppressor protein p53, which inhibits cell cycle progression and mediates apoptosis, may be a useful neuroprotective strategy in neurodegenerative diseases such as PD.63
Figure Possible signal pathway involved in cell survival and apoptosis
Nigral dopaminergic neuron degeneration in Parkinson disease is mediated via apoptotic pathways. HSP90 inhibitors such as radicicol have been considered as a complement of cancer therapies, while at the same time, they are considered as HSP70 inducers, through which inhibit mitochondria-related apoptosis. PTEN, a tumor suppressor gene, modulates PI3K signaling pathway and plays a key role in regulating cellular functions associated with proliferation/cell cycle, programmed cell death (PCD), angiogenesis, and migration. DJ1 is an oncogene and negatively regulates PTEN. PINK1 gene is PTEN-induced kinase 1, which is responsible for the cell survival. The tumor suppressor p53 inhibits cell cycle progression and mediates apoptosis. LRRK2 is involved in MAPK signal pathway, which responsible for the cell survival.
Other genes responsible for familial PD (PARK loci) are also implicated in tumorigenesis. SNCA, α-synuclein (PARK1, OMIM 168601),64 is expressed in ovarian carcinomas but not in normal breast or ovarian tissue.65 Ubiquitin C-terminal hydrolase is overexpressed in esophageal and squamous cell carcinomas and colon cancer.66 UCHL-1 is one of the most abundant proteins in the brain and immunofluorescence studies of Lewy bodies are positive for UCHL-1 protein, which possibly implicates it either directly or indirectly with the development of PD. The protein is also functionally involved in the ubiquitin-dependent proteolytic pathway; hence, UCHL-1 is a good candidate gene for PD (UCHL-1, PARK5, OMIM 191342).67 Ubiquination is important for protein degradation and crucial to regulating levels of important signal transduction proteins, sorting of proteins to different intracellular compartments, and other cellular functions that may be relevant to tumirogenesis.
As these genes are involved in the regulation of cell cycle or cell death, malfunction of a biochemical pathways related to these genes may either be related to PD or cancer. During tumorigenesis, the degradation of some enzymes can be driven by UCHL-1 or phosphatase and tension homolog (PTEN).53PTEN is a tumor-suppressing gene.53 Mutations and deletions in the PTEN locus have been found to be associated with a broad range of human cancers.53 The tumor suppressor PTEN gene encodes a multifunctional phosphatase that plays an important role in inhibiting the PI3K/Akt pathway and downstream functions that include activation of Akt/protein kinase B, cell survival, and cell proliferation. Enforced expression of PTEN in various cancer cell lines decreases cell proliferation through arrest of the cell cycle, accompanied in some cases by induction of apoptosis. PTEN deleted on chromosome 10 up-regulates another important gene involved in PD pathogenesis: the PTEN-induced kinase 1 (PINK1) gene. PINK1 coded by the PINK1 gene on chromosome 1p36 is down-regulated in the absence of PTEN and has been found to be responsible for another type of familial PD (PARK6, OMIM 605909).68 Deng et al.59 showed that PINK1 protects against rotenone and 1-methyl-4-phenyl-pyridinium ion+ -induced decrease in cell viability in SHSY5Y cells. The inhibition of PI3K/Akt pathway and the up-regulation of PINK1 by PTEN suggest the involvement of PTEN in both cancer and PD (figure).
A gene that contributes to the production of a cancer/or a gene that causes normal cells to change into cancerous tumor cells is called oncogene. Oncogenes are generally mutated forms of normal cellular genes (proto-oncogenes). DJ-1, another mitochondrial protein coded by the DJ-1 gene on chromosome 1p36, is likely an oncogene because of its negative regulatory effects on PTEN and is overexpressed in carcinomas,69 whereas mutations in the DJ-1 gene have been shown in PD families (PARK7, OMIM 606324).70
In patients with cancer or with PD, these cell survival signals may be different owing to the mutation of some genes related to the cell proliferation and cell death. Breast cancer specific gene 1 (BCSG1) or persyn (OMIM 602998)71 belongs to the γ-synuclein family. It shares the some structural features with α- and β-synuclein. As other synucleins, persyn is believed to be involved in the pathogenesis of human neurodegenerative diseases. However, in contrast to other synucleins, high levels of persyn mRNA expression are found in advanced breast carcinomas, suggesting the involvement of the encoded protein in tumor progression.72
Some polymorphisms have been analyzed in connection with PD genetics. A study has shown that in Taiwanese individuals, PD risk is associated with MAOB G intron 13 polymorphism and that this association is augmented in the presence of the COMT(L) genotype.73 COMT polymorphism was implicated as a genetic trait affecting the susceptibility of an individual to breast cancer in some populations in postmenopausal women.74
CIRCULATING MELATONIN
Melatonin plays an oncostatic role in a variety of tumor cells. There is evidence (yet unproven) that the addition of melatonin it has a beneficial effect in the treatment of human cancers.75,76 Changes in melatonin levels have also been linked to PD. A prospective study on 85,138 nurses working with night shifts with 12 years' follow-up found that nurses with more years of rotating night shift had significantly lower risk of PD. Conversely, nurses who slept more hours per night had a significantly higher risk for PD. Melatonin deficiency resulting from work associated light exposure at night has been implicated as a risk factor for breast cancer.77 Working in night shifts has been found to be associated with a higher risk of breast carcinoma.78 Scherhammer hypothesized that elevated melatonin levels among PD patients could contribute to the risk reduction for cancer among patients with PD.79,80
CONCLUSION
The relatively reduced risk of most common cancers and possibly increased risk of melanoma and breast cancer in PD is probably not an incidental observation. Understanding of the relationship between PD and cancer may provide clues to the pathogenesis of both diseases.81 It remains to be determined whether some types of cancer are more or less common in PD patients who carry known mutations of PD genes that might also be TSGs such as parkin or oncogenes such as DJ-1. Unraveling the role of these genes in cell survival or cell death may improve our understanding of the link between PD and cancer and open a therapeutic window for cancer as well as PD therapy.
ACKNOWLEDGMENT
The authors thank Tianhong Pan, MD, PhD, PD Center and Movement Disorders Clinic, Department of Neurology, Baylor College of Medicine, Houston, TX, for her helpful suggestions and preparation of the figure.
Footnotes
-
e-Pub ahead of print on August 15, 2007, at www.neurology.org.
Disclosure: The authors report no conflicts of interest.
Received November 27, 2006. Accepted in final form February 27, 2007.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.
- 7.
- 8.↵
- 9.↵
- 10.↵
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.
- 30.↵
- 31.↵
- 32.↵
- 33.
- 34.
- 35.
- 36.↵
- 37.↵
- 38.
- 39.
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.
- 45.
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.
- 56.
- 57.
- 58.
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.
- 77.↵
- 78.↵
- 79.↵
- 80.
- 81.↵
- 82.
- 83.
- 84.
- 85.
- 86.
- 87.
- 88.
- 89.
Disputes & Debates: Rapid online correspondence
- Are Parkinson disease patients protected from some but not all cancers?
- Reply from the authors
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.