Ilir Agalliu MD, Sc.D Rachel Saunders-Pullman MD Albert Einstein - PowerPoint PPT Presentation

Higher Frequency of Certain Cancers in LRRK2 G2019S Mutation Carriers with Parkinson's disease: A pooled analysis Ilir Agalliu MD, Sc.D Rachel Saunders-Pullman MD Albert Einstein College of Medicine Mount Sinai Beth Israel Medical Center New York, NY

Parkinson Disease (PD) and Cancer • Opposite biological mechanisms – PD: apoptosis & premature neuronal degeneration – Cancer: uncontrolled cell proliferation & decreased apoptosis • A link between PD and cancer was suspected when higher rates of melanoma were observed among PD patients • PD patients have otherwise lower risks for non-skin cancers – RR = 0.61 (95% CI 0.58 - 0.65) for smoking-related cancers – RR = 0.76 (95% CI 0.65 - 0.89) for others cancers • Potential explanations – Prevalence of smoking and other lifestyle risk factors are usually lower in PD patients – Differences in genetic susceptibility could play a role

LRRK2 mutations and cancer in PD patients Prior Studies • PD susceptibility genes could potentially link cancer and PD • Encode proteins with biologic mechanisms that increase cell growth or decrease cell death • LRRK2 gene encodes multiple domains – A kinase domain – A Ras-oncogene-like GTPase domain, with similar structure as B-RAF kinase associated with melanoma • LRRK2 mutations have been linked to cancer – G2019S associated with increased risk of non-skin cancers, breast cancer and kidney cancer in prior studies – R1441G/C associated with colon cancer • Results have been inconsistent across studies

Objective of Current Study • Perform a pooled data analysis to further examine association between LRRK2 G2019S mutation & cancer among PD patients – 1,549 PD patients recruited across five Movement Disorders clinics located in Europe, Israel, and US • Explore factors that could explain discrepancies between prior studies Agalliu et al. JAMA Neurology 2015

Data and Methods • 1,549 PD patients from 5 Movement Disorders clinics – Europe (Norway and Spain), Israel, and the US • Clinical, demographic, and genotyping data ( LRRK2 G2019S) • Cancer Outcomes – All cancers, non-skin cancers, smoking-related cancers, hormone- related cancers, and other types of cancer • Associations of LRRK2 G2019S with cancer outcomes were examined using mixed effects logistic regression models – Adjustment for age & ethnicity (fixed) & study center (random effect) • Explored whether associations varied by – Ethnicity: Ashkenazi Jewish (AJ) vs. others – Study Center: sensitivity analysis by excluding one center at a time

Characteristics of PD patients included in the pooled analysis Total Mutation Carriers Non-Carriers Characteristics (N=1,549) (N=177) (N=1,372) P Age at exam (y); mean SD 70.9 10.8 69.9 11.1 71.0 10.8 0.21 Age at PD Dx (y); mean SD 61.9 11.7 57.8 11.8 62.4 11.6 <0.0001 Duration of PD (y); mean SD 9.8 7.0 11.2 8.7 9.5 6.7 0.05 Study Center; n % <0.0001 Israel (Tel Aviv, Sheba) 459 29.6 49 27.7 410 29.9 Israel (Tel Aviv, Sourasky) 140 9.0 68 38.4 72 5.3 Norway 25 1.6 4 2.3 21 1.5 Spain (San Sebastian) 762 49.2 25 14.1 737 53.7 US (Beth Israel, NYC) 163 10.5 31 17.5 132 9.6 Women; n % 680 43.9 94 53.1 586 42.7 0.009 Ethnicity; n % <0.0001 Ashkenazi Jews 589 38.0 136 76.8 453 33.0 Sephardic Jews 136 8.8 7 3.4 129 9.4 Other ethnicity / white 824 53.2 34 19.2 790 57.6 Smoking status ascertained; N 304 100 204 0.97 Never smoker; n % 187 61.5 61 61.0 126 61.8 Former smoker; n % 103 33.9 34 34.0 69 33.8 Current smoker; n % 14 4.6 5 5.0 9 4.4

Associations of LRRK2 G2019S mutation with Cancer Model 1 Model 2 Non-Carriers Carriers Adjusted for age & study Adjusted for age, ethnicity Cancer Outcomes (n=1,372) (n=177) center and study center N % N % OR 95% CI p OR 95% CI p All Cancers Combined 210 15.3 40 22.6 1.49 0.99 - 2.24 0.06 1.37 0.92 - 2.04 0.13 All Non-Skin Cancers 169 12.3 32 18.1 1.57 1.04 - 2.38 0.03 1.62 1.04 - 2.52 0.03 Smoking-Related Cancers 18 1.3 2 1.1 1.04 0.22 - 4.94 0.96 1.20 0.25 – 5.76 0.82 Hormone-Related Cancers* 77 5.6 20 11.3 2.06 1.22 - 3.47 0.01 1.87 1.07 - 3.26 0.03 Breast Cancer 27 4.6 12 12.8 2.88 1.39 - 5.98 0.004 2.34 1.05 - 5.22 0.04 Prostate Cancer 40 5.1 8 9.6 2.05 0.92 - 4.55 0.08 2.21 0.95 – 5.18 0.07 Colon Cancer 29 2.1 6 3.4 1.68 0.69 - 4.11 0.26 1.92 0.74 – 5.00 0.18 Kidney Cancer 8 0.6 2 1.1 1.93 0.40 - 9.17 0.41 1.93 0.40 - 9.17 0.41 Hematologic 16 1.2 1 0.6 0.48 0.06 - 3.68 0.48 0.48 0.06 - 3.68 0.48 Meningioma 10 0.7 3 1.7 2.38 0.61 - 9.21 0.21 2.38 0.61 - 9.21 0.21 *Hormone-related cancers included prostate cancer in men and breast and ovarian cancers in women (there were no endometrial cancers reported). The percentages are gender-specific for breast and prostate cancers

Influence of Study Center on Associations between LRRK2 G2019S mutation and Cancer Outcomes Sensitivity All Cancers Non-Skin Hormone- Breast Cancer Prostate Cancer Analysis Combined Cancers Related Cancers (women) (men) All Study Centers OR 1.49 1.57 2.06 2.88 2.05 p-value 0.057 0.034 0.007 0.004 0.08 Excluding Sheba / Israel OR 1.59 1.51 2.27 3.39 2.08 p-value 0.06 0.1 0.012 0.004 0.12 Excluding Sourasky / Israel OR 1.84 2.13 2.31 2.54 3.06 p- value 0.011 0.002 0.008 0.04 0.011 Excluding Norway OR 1.35 1.44 1.74 2.37 1.84 p-value 0.17 0.09 0.05 0.03 0.16 Excluding Spain OR 1.4 1.65 2.21 2.84 1.97 p-value 0.14 0.04 0.016 0.013 0.14 Excluding US (NYC) OR 1.48 1.33 1.88 3.19 1.56 p-value 0.08 0.24 0.04 0.006 0.35

Conclusions • LRRK2 G2019S mutation-carriers vs. non-carriers had statistically significant increased risks for – Non-skin cancers: OR=1.62 (95%CI 1.04-2.52) – Hormone-related cancers: OR=1.87 (95%CI 1.07-3.26) – Breast cancer: OR=2.34 (95%CI 1.05-5.22) • Biological mechanisms for hormone-related cancers – Currently unknown – Potential activation of MAPK signaling pathway • No major differences when data stratified by AJ ethnicity • Evidence of some heterogeneity across centers • Larger studies are warranted to better delineate genetic susceptibility of PD and hormone-related cancers Agalliu et al. JAMA Neurology 2015

Potential Future Work (WHI) • Gaps in current research – There are no prospective data to determine if PD patients with LRRK2 mutations have increased risk of incident breast cancer – It’s not clear whether the association between LRRK2 mutations and breast cancer is independent or not of PD • Aims: – Genotype PD patients in WHI OS & CT (n=1,323) for mutations in LRRK2 (not only G2019S, but also other mutations through targeted sequencing) – Examine whether there is an association between LRRK2 G2019S mutations and potential other mutations (R1441G/C) and risks of • All non-skin cancers • Hormone-related cancers (breast, ovary, endometrium) • Overall risk of breast cancer and risk of subtypes (potentially ER/PR status)