FMR1 Carriers Report Executive Function Changes Prior to FXTAS: A Longitudinal Study

Abstract

Background:

Men with fragile X-associated tremor/ataxia syndrome (FXTAS) often develop executive dysfunction, characterized by disinhibition, frontal dyscontrol of movement, and working memory and attention changes. Although cross-sectional studies have suggested that earlier executive function changes may precede FXTAS, the lack of longitudinal studies have made it difficult to address this hypothesis.

Objective:

To determine whether executive function deterioration experienced by premutation carriers (PC) in daily life precedes and predicts FXTAS.

Methods:

This study included 66 FMR1 PC ranging from 40–78 years (Mean=59.5) and 31 well-matched healthy controls (HC) ages 40–75 (Mean 57.7) at baseline. Eighty-four participants returned for 2–5 follow up visits over a duration of 1 to 9 years (Mean=4.6); 28 of the PC developed FXTAS. The Behavior Rating Inventory of Executive Function-Adult Version (BRIEF-A) was completed by participants and their spouses/partners at each visit.

Results:

Longitudinal mixed model regression analyses showed a greater decline with age in PC compared to HC on the Metacognition Index (MI; self-initiation, working memory, organization, task monitoring). Conversion to FXTAS was associated with worsening MI and Behavioral Regulation Index (BRI; inhibition, flexibility, emotion modulation). For spouse/partner report, FXTAS conversion was associated with worsening MI. Finally, BRIEF-A executive function problems at baseline significantly predicted later development of FXTAS.

Conclusions:

Executive function changes experienced by male PC represent a prodrome of the later movement disorder.

Introduction

The Fragile X Messenger Ribonucleoprotein 1 (FMR1) premutation is a trinucleotide (CGG) expansion that occurs in approximately 1 in 110–250 females and 1 in 400–850 males worldwide [1,2]. Premutation carriers (PC) are at increased risk of developing a neurodegenerative disease, fragile X-associated tremor ataxia syndrome (FXTAS), characterized primarily by tremor and/or cerebellar gait ataxia, and often with co-occurring conditions such as peripheral neuropathy, parkinsonism, executive dysfunction, and dementia. FXTAS is accompanied by brain white matter changes (T2 hyperintensities in the cerebellum and cerebrum), general brain atrophy, and ubiquitin-positive intranuclear inclusions seen in post-mortem neuropathology studies.

Men with FXTAS often experience executive dysfunction (ED) during the progression of the disease, which is a minor criterion for FXTAS diagnosis. ED in FXTAS often includes disinhibition, frontal dyscontrol of movement, working memory loss and/or deterioration of attention. ED in FXTAS was initially recognized through examination of a series of cases utilizing psychiatric assessment [3], but it was not until the work of Jim Grigsby and colleagues that a clearer picture of the cognitive phenotype was appreciated through standardized neuropsychological assessments and specialized tests that measure the frontal/executive control of movement [4–6]. These studies revealed that while verbal intelligence and domains of perceptual reasoning not involving motor coordination are relatively spared, measures tapping regulation of manual motor movements, verbal fluency, general mental status, information processing speed, temporal sequencing, working memory, inhibition, short-term memory, and cognitive flexibility tended to show significant deficits. This study’s early focus on male PC who were asymptomatic for FXTAS also introduced an important question for the field: Do carriers without FXTAS demonstrate executive dysfunction and if so, are they stable traits associated with the premutation, or do they represent a prodrome of the movement disorder?

Initial clues to begin to answer this question came soon in the form of a cross-sectional study of PC males ranging in age from 18–69 years and comparison men without the premutation in different decades of life, by colleagues in the UK led by Dr. Kim Cornish [7,8]. The primary conclusion of this study, effectively highlighted by graphs showing decreasing scores with advanced decades of life, is that age is correlated with increasingly worse performance on measures of inhibitory control, working memory and attention in PC (compared to controls). Then, a series of studies utilizing fMRI [9,10], gait metrics affected by cognitive load [11,12], disinhibited eye movements controlled by the cerebellum [13], and a suite of structural MRI studies, including those that correlated brain with cognitive measures [14–20] have provided converging evidence of a dysexecutive pattern in PC with and without FXTAS.

While these cross-sectional studies revealed brain and cognition differences prior to the onset of FXTAS, the ability to determine whether cognitive changes represent prodromal features of later movement disorder relies on following individuals prospectively and longitudinally. To address this, we conducted a neuropsychological study, utilizing the Cambridge Neuropsychological Automated Battery (CANTAB), which included 64 male FMR1 PC without FXTAS at study entry and 30 healthy controls (HC), aged 40 to 80 years. Fifty of the PC and 22 of the HC were re-assessed after an average of 2.33 years, and 37 PC and 20 HC were re-assessed a third time after an average of another 2.15 years. Results showed that changes in visual working memory, inhibitory control, and planning progress at a higher rate in PC than matched HC, and that worsening inhibitory control and planning tracked the onset of FXTAS [21]. Although these neuropsychological tests can reveal performance-based changes in the clinic setting, it is critically important to also understand whether executive function changes affect PC in their daily lives, and if present, whether they precede or track with the onset of FXTAS. To address these questions, the present study examined self-reported and spouse/partner-reported EF symptoms on the Behavior Rating Inventory of Executive Function-Adult Version (BRIEF-A; [22]) a standardized measure that captures an adult’s executive functions or self-regulation in his everyday environment, in this longitudinal cohort of PC and HC.

Methods

Our ongoing prospective longitudinal study examines trajectories of neuropsychological, motor, brain, and molecular changes in male fragile X PCs and HC with a focus on the FXTAS prodromal period and early neurodegenerative process. Participants in the present analysis include 66 PC ranging from 40–78 years (Mean=59.5) and 31 HC ages 40–75 (Mean 57.7) at baseline (visit 1). PC were recruited from throughout North America, from a variety of sources including patient referrals for clinical evaluation, the announcements by the National Fragile X Foundation, and flyers posted on social media websites dedicated to fragile X families. HC were recruited from flyers posted in the local community and on the university campus. Exclusion criteria included: acute renal, liver, or cardiac medical conditions; history of significant head trauma; substance abuse or dependence; presence of metal implants of any kind that would preclude MRI (reported elsewhere); and primary spoken language other than English or Spanish. The project protocol was approved by the institutional review board at the University of California Davis, and all participants provided signed consent. PC visits to the research center usually entailed compensated air travel and local lodging.

Diagnosis of FXTAS and the evaluation of disease stage was carried out by a physician using established guidelines [3]. The examination was composed of standardized movement disorder assessments from the Unified Parkinson’s Disease Rating Scale [23], International Cooperative Ataxia Rating Scale [24], and the Scale for the Assessment and Rating of Ataxia [25] and review of brain MRI to assess white matter disease associated with FXTAS.

An important goal of the recruitment process was the inclusion of PC who do not yet have FXTAS at enrollment. During phone screening, detailed interviewing with potential participants suggested than none of the enrolled PC had FXTAS based upon criterion of evidence of either tremor or ataxia with any degree of impairment in functioning. Despite the screening, 11 of the 66 PC demonstrated at least equivocal/intermittent tremor and/or ataxia and functional interference (FXTAS Stage 2 or higher) at visit 1 after neurological examination. Overall, eighty-four participants returned for 2–5 follow up visits over a total duration of 1 to 9 years (Mean=4.6 years). Among PC, 59/66 had two or more visits, 40 had three or more, 17 had four or more, 2 had five visits. Among HC, 27/31 had two or more visits, 19 had three or more, 16 had four or more, and 2 had five visits (see Figure 1). The primary reasons for lack of follow up visits were travel inconvenience or problems with ambulation preventing travel, busy work schedules, COVID pandemic university restrictions on in-person research, and inability to contact participants due to moves. Of the PC participants without evidence of FXTAS at baseline who had follow-up visits to date, 28 “phenoconverted” and developed FXTAS (tremor and/or ataxia and functional interference). One of the PC died as a result of FXTAS complications.

Figure 1.

Spaghetti plot showing the ages at assessment for each study participant by group and FXTAS status.

The Behavior Rating Inventory of Executive Function-Adult Version (BRIEF-A; [22]), a standardized measure that captures views of an adult’s executive functions or self-regulation in his everyday environment, was completed by the participants and their spouses/partners completed the BRIEF-A other-report about the participant at each visit. FXTAS stage also was evaluated at each visit to track disease emergence and progression. This analysis focuses on the two broader indices of the BRIEF-A (Behavioral Regulation, BRI; and Metacognition, MCI) and the overall Global Executive Composite (GEC). The BRI includes several EF components including Inhibit, Shift, Emotional Control, and Self-Monitor. The MCI includes EF dimensions of Initiate, Working Memory, Plan/Organize, Task Monitor, and Organization of Materials. These the BRI and MCI combine to yield the GEC.

Data Sharing

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Results

There were no significant group differences at baseline (visit 1) in age, full scale IQ, education level, BRI, MCI, or GEC, with mean BRIEF-A T-scores very close to average for age in both groups (Table 1). Pearson correlation matrices of these variables showed a modest but significant correlation between IQ and BRIEF BRI in PC’s, such that higher IQ was associated with less significant behavioral regulation concerns (Supplementary table 1). Longitudinal mixed-effects models showed a greater decline with age (both linear and quadratic) in PC compared to HC on the self-report MCI, t(134)= 2.34, p=0.02, but not on the BRI or GEC (Table 2). Although there was no group difference in IQ at baseline, because of the correlation with IQ and the BRI, the mixed effect models were repeated, covarying IQ as a conservative check. The results were similar, although the effect was somewhat strengthened for the MCI, t=2.90, p=0.005.

Table 1.

Baseline (visit 1) demographic and BRIEF-A descriptive statistics by group.

	Healthy Control (n=31)		Premutation Carrier (n=66)
	M	SD	M	SD	P
Age	57.74	9.71	59.49	8.05	Ns
FSIQ	124.63	14.09	124.38	14.22	Ns
Education (yrs)	17.00	2.35	17.08	3.54	Ns
Race (% White)	90.30		95.50		Ns
Ethnicity (% Hispanic)	9.70		3.00		Ns
FMR1 CGG repeats	30.19	4.98	85.33	18.86	<.001
BRIEF MCI (T)	52.29	12.91	52.82	11.74	Ns
BRIEF BRI (T)	49.03	8.50	50.94	9.76	Ns
BRIEF GEC (T)	50.23	9.34	51.92	10.10	Ns

Table 2.

Results of longitudinal mixed model regression examining the effect of age, group (PC vs. HC), and interaction of age and group on BRIEF self-report raw scores.

	Estimate	t	P
Self-BRI
Intercept	51.09	11.11	<.001
Age	−1.21	−2.67	0.009
Age2	0.03	2.75	0.007
Group	0.21	0.03	ns
Age*Group	0.12	0.20	ns
Age2*Group	−0.00025	−0.02	ns
Self-MI	Estimate	t	P
Intercept	73.03	11.08	<.001
Age	−2.05	−3.34	0.001
Age2	0.05	3.72	<.001
Group	−15.12	−1.66	ns
Age*Group	1.94	2.30	0.023
Age2*Group	−0.04	−2.34	0.021

In premutation carriers, conversion to FXTAS based upon advancing FXTAS Stage with age in PC was associated with worsening BRI, t(83)=2.59, p=.01, MI, t(81)=2.03, p=.04 and GEC, t(83)=2.17, p=.03 (Table 3). Because 11 PC had already developed signs of FXTAS at baseline, these analyses were repeated with them excluded in the case that they skewed these results. The total N is then reduced from 65 to 54, and the number of observations from 151 to 118. For BRI, the results remained similar, t=1.95, p=0.05. For MI and GEC, the effect became non-significant.

Table 3.

Results of longitudinal mixed model regression examining the effect of age, FXTAS Stage, and interaction of age and FXTAS stage on BRIEF self-report raw scores (PC only).

Self-BRI	Estimate	t	P
Intercept	47.17	13.37	<.001
Age	−0.25	−1.48	0.143
FXTAS Stage	−3.95	−2.02	0.047
Age*FXTAS Stage	0.21	2.59	0.011
Self-MI
Intercept	57.16	7.28	<.001
Age	0.37	0.45	Ns
Age2	−0.01	−0.60	Ns
FXTAS Stage	5.84	1.19	Ns
Age*FXTAS Stage	−0.71	−1.68	Ns
Age2*FXTAS Stage	0.02	2.03	0.046
Self-GEC	Estimate	t	P
Intercept	109.43	13.87	<.001
Age	−0.44	−1.18	0.059
FXTAS Stage	−7.40	−1.72	Ns
Age*FXTAS Stage	0.39	2.17	0.030

Survival analyses predicting the probability of converting to FXTAS (stage 2 or higher), utilizing age and BRIEF self report-revealed that baseline measures of age (OR = 1.09, SE = .046, p = .048), higher BRIEF GEC (OR = 1.03, SE = .014, p = .047), and MI (OR = 1.06, SE = .024, p = .020), but not BRI, were significantly associated with probability of later FXTAS conversion. When the sample was restricted only to those without FXTAS (below stage 2) at baseline, the effect for MI (but not BRI or GEC) remained significant (OR = 1.07, SE = .032, p = .035).

For spouse/partner report BRIEF-A, no difference in trajectory was found between groups, however FXTAS conversion was associated with worsening GEC, t(77) = 2.07, p=.04, and this effect approached significance for the MI (Table 4). Agreement between PCs and spouse/partners was positive and statistically significant, but the magnitude of agreement was modest (e.g., PC GEC Intraclass Correlation Coefficient=.55.; HC GEC Intraclass Correlation Coefficient =.79).

Table 4.

Results of longitudinal mixed model regression examining the effect of age, FXTAS Stage, and interaction of age and FXTAS stage on BRIEF spouse/partner-report raw scores (PC only).

Spouse/Partner-MI	Estimate	t	P
Intercept	68.15	11.31	<.001
Age	−0.57	−1.91	Ns
FXTAS Stage	−4.73	−1.35	Ns
Age*FXTAS Stage	0.30	1.95	0.05
Spouse/Partner-GEC
Intercept	115.64	11.22	<.001
Age	−0.70	−1.41	ns
FXTAS Stage	−8.58	−1.59	ns
Age*FXTAS Stage	0.49	2.07	0.04

A longitudinal mixed-effects model was also run to examine the potential role of FMR1 CGG size on changes in BRIEF executive function within the PC group. There were no significant effects for any of the three BRIEF composites (all p>0.15.

Discussion

This study examined the ratings of executive function problems of adult male FMR1 premutation carriers (PC) and a well-matched group of healthy men without the premutation (HC) in a prospective longitudinal study. A unique aspect of the project is the focus on enrollment of PC before the onset of FXTAS in order to study its prodrome and earliest indicators of disease onset and progression. At baseline, with a mean age of about 49 years, PC and HC reported normal rates of executive function challenges with nearly identical standard scores on the BRIEF-A. Over time, however, PC reported a greater decline in metacognition (composed of self-initiation, working memory, organization, task monitoring) with age than HC, but not a significantly worse decline in behavioral regulation. Declines in both metacognition and behavioral regulation were associated with onset and progression of FXTAS in those that developed the disease over time. Interestingly, spouses/partners were able to notice and report changes in EF among those who progressed to FXTAS but seemed less able to detect the EF changes in PC generally. The survival analysis provided preliminary evidence that the PC with relative weaknesses in metacognition are at higher risk for later neurodegeneration. The results also seem to indicate that PC are aware of changes in their metacognition before it can be recognized in outward behavior by others. It is also notable that self-report of executive function challenges was a stronger predictor of later FXTAS conversion (at least as measured by stage) than CGG repeat length.

These results lend support to prior cross-sectional studies suggesting that EF changes precede FXTAS and may be considered a prodromal feature of the movement disorder, similar to what has been observed in Parkinson’s disease (PD). In two large prospective cohort studies of community dwelling elders, men who later developed idiopathic PD showed a faster decline in executive function in the prediagnostic period, followed by a steeper decline after diagnosis [26]. Indeed, parkinsonism is a common feature of FXTAS and many individuals with FXTAS are misdiagnosed with PD [27–29]. If PC experience EF changes before FXTAS develops and are an aspect of the earliest manifestations of disease, it is possible that measures could be developed that are especially sensitive to the specific EF problems that precede the onset of the movement disorder. This is a challenge, however, given that normal aging is associated with changes in EF, and such tools would need to separate FXTAS-associated executive changes from normal aging decline. Studies are needed to compare and contrast the sensitivity and specificity of predictive clinical measures and biomarkers in PC at risk for FXTAS. This is likely to require larger samples through combined efforts of investigators in consortia and standardized measures and procedures. With some targeted treatment trials on the horizon [30–32], this work will be critical in order to identify the most optimal measures to identify PC most at risk and to track response to treatment.

What are the brain mechanisms underlying motor and executive function decline in male carriers, and how are they connected to FXTAS? A growing body of research suggests that deterioration of frontal-cerebellar circuits is especially sensitive to the FMR1 premutation and appear to impact both control of movement (ataxia and tremor) and executive control of cognition (e.g., working memory) in PC [9,19,33–35]. For example, we previously showed that compared with HC, PC with and without FXTAS performing a verbal working memory EF task showed reduced activation in the right ventral inferior frontal cortex and left premotor/dorsal inferior frontal cortex [9]. In a cross-sectional structural brain imaging study of 322 males from age 8 to 81, we discovered that compared to age-matched HC, PC without FXTAS showed significantly accelerated age-related volume decreases in the cerebellum and brainstem and increased ventricular volume [18].

One of the limitations of the study is the unequal sample sizes, with fewer HC than PC. The project was designed in this way to maximize characterization of PC trajectories of change and examine predictors of phenoconversion in PC, given limited resources including domestic travel and imaging costs. Another limitation is the exclusion of female PC, and as such, the study can only be generalized to males. The reason for the focus on males is that the project was designed expressly to study the prodrome of FXTAS and factors that predict onset of disease. Given that the PC is an X-linked condition, and that only about 8–16% of women develop FXTAS [36], we chose to focus on men who are much more likely to phenoconvert during the study period. Whether EF problems are associated with neurodegenerative or other health changes in female PC is an open question. Also, the results of this study may not generalize to the full population of male PC because this project included a predominantly white, relatively affluent and well-educated group of men who were able to travel and take time away from work. It will be essential to improve the diversity of this sample in future recruitment efforts. In the future, we plan to examine self-reported and neuropsychological test-related changes in EF in concert with structural MRI changes over time in PC vs HC to obtain a more integrated picture of how neurodegeneration unfolds and determine which metrics are most useful to collect in intervention studies including clinical trials. Nevertheless, it is clear from this and prior work that EF (and especially EF change) is a critical cognitive construct to assess in fragile X PC.