Abstract
INTRODUCTION
The etiology of PSP is currently unknown, however, it is believed that genetic, environmental, oxidative, and inflammatory factors may all contribute [1, 2]. A recent genome-wide association study identified a genetic risk factor, MAPT, which is comparable to the risk associated with the APOE ɛ3/ɛ4 genotype for AD [2]. In recent years there have been numerous studies examining the role of stress in the etiology of diverse medical conditions [3–7]. Thus far, the role of stress as a potential risk factor for PSP has not been studied.
Stress has been shown to have a role in the development of Alzheimer’s Disease (AD). Studies have demonstrated that prolonged exposure to psychological stressors in older non-demented individuals with the presence of a known genetic risk factor, leads to accelerated memory decline [8]. An association between stress and the pathogenesis of AD has also been demonstrated in mouse models, in which AD transgenic mice exposed to glucocorticoids, the hormonal response to stress, increase tau accumulation and therefore, demonstrate accelerated development of neurofibrillary tangles [9]. Additionally, a study involving healthy, wild-type mice, showed that the experience of chronic stress and increased glucocorticoid release induced hyperphosphorylation of tau [10]. This supports the hypothesis that even in individuals without genetic risk factors, chronic stress may be a risk factor for developing a tauopathy.
Pathologically, PSP is the most frequent primary “tauopathy”, a diseases in which tau is hyperphosphorylated and aggregates to form neurofibrillary tangles in neurons and tufted astrocytes [11]. AD is the most common tauopathy [12].
Given the pathologic similarities between PSP and AD, and in light of the findings implicating stress in the development of AD and in other disorders, this study investigated the role of stress as a risk factor for PSP. Based on the literature review and our clinical practice, we hypothesized that PSP patients would report a greater number of high severity life stressors in the preclinical stage when compared to controls.
MATERIALS AND METHODS
Participants
PSP patients (i.e. cases) were recruited from 9 enrollment sites: (Baylor University, University of Colorado, Case Western Reserve, Emory University, University of Louisville, University of Alabama Birmingham, University of California at Los Angeles, Mayo Clinic Jacksonville, and University of Maryland) as a sub-study of a case-control study to determine environmental risk factors for this disorder [13]. Internal Review Board approval was obtained at each participating site. All participants completed a consent form for enrollment in this study. The final sample consisted of 143 participants (76 cases, 68 controls).
The majority of PSP patients met the National Institute of Neurological Disorder and Stroke Society for PSP Inc. (NINDS-SPSP) criteria for clinically probable or definite PSP, the remaining 10% of the patients (N = 8) met criteria for clinically possible PSP [14]. The majority of enrolled patients scored >24 on the Mini Mental Status Examination (MMSE) [15]. An exception of the MMSE cutoff was made for 5 participants, who the primary site PI did not believe met the clinical criteria of cognitive impairment. Potential participants with other central nervous system disorders were excluded. PSP patients were asked to identify an age (±5 years) and gender matched non-blood related control. Spouses were not included. Controls were screened to rule out dementia using the Telephone Interview for Cognitive Status (TICS) and were also screened via telephone interview to rule outparkinsonism [16].
Measures
Life stressors questionnaire
All participants were administered a life stressor questionnaire by trained study personnel. The major life events in this study were derived from a 61-item questionnaire developed by Paykel [17]. Some of the events from the original questionnaire were omitted from the study (n = 25 events) because they were outdated or uncommon (e.g., child going off to war) or they were not thought to represent an event that would cause “prolonged stress” (“take an important exam” or “academic failure”). Prolonged stress is defined as lasting for two weeks or longer [8]. Some life events were combined into a single event (e.g. “move within the same city” was combined with “move to a new city within the same country”). Moreover, additional relevant questions (n = 6) were added to the questionnaire from other life event [18, 19]. The final version of the questionnaire had a total of 36 questions and is available in the Supplemental Materials. Participants were asked to rate the level of stress impact of each event they as low, mild, or high; to state whether the event was life-changing, and to report the specific year the event occurred.
All reported events were recorded; however, only events reported as occurring prior to the “reference year” were included in the statistical analyses. The reference year for each case, and associated control, was defined as 10 years prior to the date of first reported PSP symptom. The purpose of this time frame was to account for the hypothesized lag time for preclinical PSP to manifest clinically. Though there is no published literature on the length of this lag time, we chose 10 years to be on the conservative side to avoid including possible preclinical and prodromal disease periods.
Quantification of measure
Data from the life stressors questionnaire were quantified for each participant as follows: 1) total number of events, 2) number of life-changing events, and 3) total number of events self-rated as “high severity”.
Statistical analysis
Data were assessed for normal distribution graphically using QQ plots. All tests of significance were two-tailed and alpha was 0.05. Statistical analysis was performed using “R”, and the packages “gmodels”, “epitools”, and “car”.
Demographic data
For continuous variables the mean and standard deviation are reported for cases and controls; differences were assessed using the Student’s t-test. For categorical variables, frequencies and proportions are reported for cases and controls; differences were assessed using Fisher’s exact test.
Life stressors data
For each of the different measures of the life stressor questionnaire, mean and standard deviation was determined for cases and controls. The measures of life stressors were found to have a non-normal distribution. Differences between life stressors were assessed using the Wilcoxon-rank sum test.
Participants were classified into “life stressor exposure groups” for each method of life stressor quantification. To classify exposure level, the quartiles for each life stressor quantification method were determined. Participants who reported a greater number of events than the cut-off between the third and fourth quartile of the measure, were classified as having experienced “high exposure”. Participants who reported a less than or equal to number of events as the cut-off between the third and fourth quartile were defined as having experienced “normal exposure”. Conditional odds ratios (OR) with 95% confidence intervals were calculated. Significance for unadjusted ORs was assessed usingChi-Square test.
Adjusted ORs for high exposure to each measure were determined by performing logistic regression controlling for age and gender. The adjusted ORs are reported with 95% confidence intervals. Significance for adjusted ORs was determined using Chi-Square test.
Cognitive functioning and life stressor quantification methods
In order to assess if cognitive function in participants affected scores in the various methods of life stressor quantification, correlation testing between measures of cognitive function and life stressor quantification scores was performed. Measures of cognitive function in cases were: Dementia Rating Scale (DRS) [20], Frontal Assessment Battery (FAB) [21], and MMSE [15]. The measure of cognitive function in controls was the TICS [16]. Pearson’s correlation coefficient and p-value were reported.
RESULTS
Demographics
Demographic data for the 76 cases and 67 controls are presented in Table 1. There were no significant between-group differences in baseline demographics. There was an expected statistically significant difference between cases and controls current level of independence.
Demographics of Cases and Controls
Life stressors data
Table 2 shows there was no between-group statistically significant differences in number of reported total events (cases: 5.4, controls: 5.5), and life-changing events (cases: 2.9, controls: 3.1). Similarly, there was no difference when participants were categorized based on exposure to either of these measures. Additionally, no differences existed when adjusting for age and gender for either of these measures.
Life Stressor Quantification Methods Results
**Participants with values greater than this cut-off are defined as having experienced high exposure. Participants with values less than or equal to this cut-off are defined as having experienced normal exposure.
There was no statistical between-group difference in the number of self-reported highly stressful events (cases: 3.4, controls: 2.6). However, there was a statistical difference when participants were categorized based on exposure level, 24.4% of cases (N = 11) and 9.1% of controls (N = 5) were defined as high exposure, yielding an estimated OR of 3.2 (p = 0.04). When adjusting for age and gender the cases remained at increased odds of high exposure to total number of highly stressful events with an adjusted OR of 3.4 (p = 0.04).
Correlation of cognitive measures and life stressor measures
There was no significant correlation between the DRS, FAB, MMSE, or TICS and any of the life stressor quantification methods. Results are presented in Table 3.
Cognitive Functioning and Life Stress Quantification Methods - Pearson’s Correlation Coefficient (r)
DISCUSSION
We used a case-control study design to examine the potential role of stress as a risk factor for PSP. We hypothesized that the experience of high levels of stress prior to the preclinical onset of disease would occur more often in the cases than controls. Our hypothesis was based on the fact that: [1] in vivo studies show that glucocorticoids, hormones released in response to stress, accelerate the development of neurofibrillary tangles [2, 9] stress has a role in AD, a related tauopathy [3, 8] in our practice many PSP patients report stress preceding the onset of their symptoms.
We found that [1] there was no significant difference between the total number of events reported by cases and controls; however, as predicted [2] there was a significant association between high exposure to high-severity life stressors and PSP. PSP patients were found to be three times more likely to report high exposure to highly stressful events than controls. This suggests that these highly stressful events may have a role in the etiopathogenesis of PSP. Future studies should address the role of highly stressful events in the development of PSP to clarify thisrelationship.
This study has several limitations. One limitation is that, like many retrospective surveys, there is concern that as events may have occurred years earlier the events may not be recalled accurately. Due to the possible effect of PSP on memory, the exclusion criteria for cases was an MMSE score of less than 24, which is the established cut-off for dementia. To further address concerns about mental status influencing recall bias, we performed correlation analyses and found no significant correlations between various measures of cognitive functioning and the life stressor variables. Additionally, there is the concern that cases may be more likely to describe an event as “highly stressful” or “life-changing” as they may be seeking an understanding of why theydeveloped PSP.
A major strength of this study is that it is the first one to investigate the role of experienced life stress as a risk factor for PSP. By better understanding the potential role of high exposure to highly stressful events, we could better understand its biology. An additional strength of the study is that we incorporated the concept of “lag time”, the hypothesized time it takes for preclinical PSP to manifest clinically. Other studies investigating stress as a risk factor for disease have also incorporated “lag time” [4, 7]. This serves to minimize the reporting of events that occurred as a result of the presence of disease, and additionally, leads to only events that occurred at a biologically plausible time to be analyzed as potential risk factors for PSP. Finally, an additional strength of this study is that a relatively large number of patients were recruited for a rare disease. Our study included a comparable sample size to similar studies investigating the role of stress in the etiopathogenesis of disease [3–5].
In summary, our study found a significant association between high exposure to high severity life stressors and PSP. As this study is the first to investigate this potential risk factor, further investigations are needed to confirm results and understand this relationship. Future studies should focus specifically on these highly stressful events, as these appear to be the most significant in their association with PSP. Additionally, we hope that future studies will be able to investigate for a potential interaction between the recently identified genetic risk factor for PSP, MAPT, and the occurrence of life stressors. This is of particular interest as in many of the AD studies the most significant relationship between experienced stress and occurrence of AD was found in those with identified genetic risk factors. Finally, it would be important as well to determine whether highly stressful events may accelerate the progression of symptoms associated with PSP.
CONFLICT OF INTERESTS
Dr. Litvan is also an investigator in studies funded by 5P50 AG005131-31, 5T35HL007491, 1U01NS086659, 1U54NS092089-01, Parkinson Study Group, Michael J Fox Foundation, CBD Solutions-CurePSP, AVID Pharmaceuticals, C2N Diagnostics and Bristol-Myers and Squibb. She receives her salary from the University of California San Diego. She was member of the Pfizer, Abbvie Northera and Bristol-Myers and Squibb Advisories Boards. Dr. Riley has received honoraria from Allergan, Lundbeck and Teva. Dr. Bordelon received speaker honoraria from Teva and Lundbeck. Dr. Standaert is also an investigator in studies funded by Abbvie Laboratories, the American Parkinson Disease Association, the Bachmann-Strauss Foundation, the Dystonia Medical Research Foundation, Acerta Pharmaceuticals, Ceregen, Inc, The MGH XDP Consortium, the Michael J. Fox Foundation for Parkinson Research, and NIH grants P20NS092530, K23NS080912, R25NS079188, K01NS069614, 1F31NS084722, 1F31NS084722, F31NS081963, R01NS064934 and P50NS037409. In addition, in the last three years he has served as a consultant for or received honoraria from Auspex Inc., Teva Neurosciences, Serina Therapeutics, Lundbeck, Abbvie, Bradley Arrant Boult Cummings, Banner Sun Health, Viropharma (now Shire), and he has received royalties for publications from McGraw Hill, Inc. Dr. Reich received research funding from NIH-NINDS/ORDR-NCATS. He received royalties from Informa and from being a reviewer of UpToDate.
Footnotes
ACKNOWLEDGMENTS
The authors want to thank the patients and controls who generously gave their time to participate in this study. They also want to thank Drs. Jorge Juncos, Deborah A. Hall, Ryan Uitti and Joseph Jankovic for referral of a few patients who participated in this study.
Kristen D. Kelley MAS1, Harvey Checkoway PhD2, Jorge Juncos MD3, Deborah A. Hall MD PhD4, Stephen G. Reich MD5, Chris Cunningham PhD6, Irene Litvan MD7.
Drs. Litvan, Riley, Bordelon, Standaertand Reich as well as Drs. Juncos, Hall, Uitti, and Jankovic, were funded by NIA R01AG024040. This study was in part funded by the NIA R01AG005131. Additionally, the project described was partially supported by the National Institutes of Health, Grant TL1TR00098 of CTSA funding prior to August 13, 2015 and Grant TL1TR001443 of CTSA funding beginning August 13, 2015 and beyond. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.