Genetic polymorphisms for BDNF, COMT, and APOE do not affect gait or ankle motor control in chronic stroke: a preliminary cross-sectional study

Abstract

Background:

Motor deficits after stroke are a primary cause of long-term disability. The extent of functional recovery may be influenced by genetic polymorphisms.

Objectives:

Determine the effect of genetic polymorphisms for brain-derived neurotrophic factor (BDNF), catechol-O-methyltransferase (COMT), and apolipoprotein E (APOE) on walking speed, walking symmetry, and ankle motor control in individuals with chronic stroke.

Methods:

38 participants with chronic stroke were compared based upon genetic polymorphisms for BDNF (presence [MET group] or absence [VAL group] of a Met allele), COMT (presence [MET group] or absence [VAL group] of a Met allele), and APOE (presence [ε4+ group] of absence [ε4- group] of ε4 allele). Comfortable and maximal walking speed were measured with the 10-meter walk test. Gait spatiotemporal symmetry was measured with the GAITRite electronic mat; symmetry ratios were calculated for step length, step time, swing time, and stance time. Ankle motor control was measured as the accuracy of performing an ankle tracking task.

Results:

No significant differences were detected (p≥0.11) between the BDNF, COMT, or APOE groups for any variables.

Conclusions:

In these preliminary findings, genetic polymorphisms for BDNF, COMT, and APOE do not appear to affect walking speed, walking symmetry, or ankle motor performance in chronic stroke.

Introduction

Stroke is a leading cause of long-term disability [1], which impairs independence, social engagement, ability to perform daily activities, and overall quality of life [2]. Stroke survivors show reduced walking speed compared to healthy controls, asymmetrical temporal and spatial parameters of walking [3], and perturbed ankle motor control [4]. Regaining function, especially walking ability, is one of the main self-reported goals of stroke rehabilitation.

There is much inter-individual variability in functional recovery after stroke [5, 6]. For example, motor recovery can vary from minimal to full recovery, making it difficult to predict with biomarkers. Genetic polymorphisms, frequently termed single nucleotide polymorphisms (SNP), are variations involving a change of nucleic acids in DNA sequence at one or more positions that may affect the function and expression of the gene. In other words, SNPs are DNA variations that (depending on the location) may affect gene expression, protein production, and bodily function. Many SNPs have no discernible effects, but others may explain interindividual differences in bodily functions, including functional recovery after stroke. Previous studies have suggested that polymorphisms of genes that encode brain-derived neurotrophic factor (BDNF), apolipoprotein E (APOE), and catechol-O-methyltransferase (COMT) influence functional recovery, cognitive and motor learning, and capacity for neuroplasticity in individuals with and without neurological impairment [7, 8]. It is unclear whether these polymorphisms influence recovery of walking and ankle motor control in chronic stroke.

BDNF is a growth factor that promotes neuronal growth and development [9]. The BDNF gene has a SNP called rs6265 or Val66Met that results in a valine (Val) to methionine (Met) substitution at codon 66 [10]. This substitution reduces trafficking and activity-dependent release of BDNF and affects prefrontal cortex and hippocampal volume, impairing memory and motor learning [7, 11]. It is unclear whether BDNF polymorphism affects walking and ankle movement after stroke. In one study, BDNF polymorphism was unrelated to walking speed [12], but another study found that Met allele carriers had a slower rate of step length adaptation during a locomotor task [13]. The rate of learning a visuomotor task is also decreased in Met allele carriers [14]. Given its influence on motor learning, BDNF polymorphism may potentially affect recovery of walking speed, walking symmetry, and ankle motor control after stroke.

COMT is an enzyme that regulates dopamine, contributing to intelligence, memory, and prefrontal functions [15]. The COMT gene has a SNP called rs4680 or Val158Met that results in a Val to Met substitution at codon 158; this substitution impairs dopamine degradation and causes higher dopamine levels [16]. In healthy individuals, carriers of the Val/Val genotype have better sequence switching performance than carriers of one Met allele, and faster walking speeds than carriers of two Met alleles [17, 18]. After stroke, carriers of two Met alleles have impaired functional independence and motor function [19]. Furthermore, general functional recovery measured with the Fugl-Meyer assessment after 1 week, 3 and 6 months post stroke was impaired in Met carriers [20]. However, it is unclear whether COMT polymorphism affects walking speed, walking symmetry, or ankle motor control after stroke.

APOE is a protein that influences lipid and cholesterol transport and is associated with repair and regeneration of cardiovascular and neurological systems [21]. The APOE gene has two common SNPs; rs7412 and rs429358 at codons 112 and 158 [22]. The allele APOE ε4 is defined by a cysteine to arginine substitution at both positions. APOE ε4 carriers have a higher susceptibility for Alzheimer’s disease and cognitive and executive function decline [7, 21]. In healthy older individuals, APOE ε4 carriers have slower walking speeds [23]. APOE ε4 carriers have worse motor recovery from stroke over the first month [24], but functional independence is similar after one year [25]. It is unclear what effect APOE polymorphism has on the recovery of walking parameters and ankle motor control in chronic stroke.

The purpose of this study was to determine the effect of genetic polymorphisms for BDNF, COMT, or APOE on walking speed, walking symmetry, and ankle motor control in chronic stroke survivors. Understanding the extent of the effects of genetic polymorphisms on lower limb functional recovery after stroke may help individualize rehabilitation goals and interventions for stroke survivors. To achieve this purpose, we compared these outcome measures between individuals with different polymorphisms for BDNF, COMT, and APOE. We hypothesized that carriers of BDNF Met alleles, COMT Met alleles, and APOE ε4+ alleles would have slower gait speed, greater gait asymmetry, and more impaired ankle motor control.

Materials and Methods

Subjects

For this preliminary, cross-sectional study, participants were recruited as a convenience sample from existing databases and from local hospitals and clinics. Participants were included if they were 18–80 years old, had a single mono-hemispheric stroke >6 months before enrollment, had residual walking deficits but were able to walk, and had at least 5 degrees of active ankle dorsiflexion. Presence of residual walking deficits was identified during a screening session as a comfortable walking speed <1.2 m/s and/or the presence of spatiotemporal gait abnormalities. Individuals were excluded if they had stroke lesions affecting the brainstem or cerebellum, cognitive or communication impairment (Mini-Mental State Exam <21), severe osteoporosis, contracture of lower limb, score ≥2 on the Modified Ashworth Scale, cardiorespiratory or metabolic diseases, unhealed decubitus ulcers, or persistent infection. Based on pilot data of participants with different APOE polymorphism (group ε4+: n=5, maximal walking speed=0.88 (0.35); group ε4-: n=5, maximal walking speed=1.18 (0.22); effect size d=1.03), we estimated that we would need a sample size of 36 (β=0.1, α=0.05) to detect a difference in walking speed between APOE groups with a Mann-Whitney U test. All participants included in this study provided written informed consent. This study was approved by the University of Illinois at Chicago institutional review board. All data were collected at a single laboratory site between June 2014 to October 2018. This manuscript conforms to the STROBE guidelines.

Genotype determination

To collect genomic DNA, participants provided a 2 mL saliva sample per manufacturer instructions using Oragene OGR-500 self-collection kits which include a stabilizer solution (DNA Genotek, ON, Canada). De-identified and coded samples were sent to the Functional Genomics Laboratory at the University of Maryland School of Public Health. DNA was extracted according to manufacturer instructions, and samples were checked for concentration and purity. For genotyping, primer sequences were designed for each genotype (e.g., BDNF rs6265), and typical PCR reactions were performed prior to an extension reaction that allowed for mass spectrometry analysis and allele identification [26].

As expected, allelic distribution was imbalanced for BDNF (no MET allele, n=15, one MET allele, n=7; two MET alleles, n=16), COMT (no MET allele, n=19, one MET allele, n=17; two MET alleles, n=2), and APOE (no ε4 allele, n=27, one ε4 allele, n=9, two ε4 alleles, n=2). To compare outcome measures based on genetic polymorphism, BDNF groups were formed based upon presence (MET group, n=23) or absence (VAL group, n=15) of the MET allele, COMT groups were formed based upon presence (MET group, n=19) or absence (VAL group, n=19) of the MET allele, and APOE groups were formed based upon presence (ε4+ group, n=11) or absence (ε4- group, n=27) of the ε4 allele.

Outcome measures

All outcome measures were assessed as part of baseline measurements for an ongoing randomized controlled trial evaluating motor priming and treadmill training (clinical trial registration: NCT03492229). Genotype determination was performed after collection of outcome measures, so investigators were inherently blinded to genotype during testing.

Walking speed

Participants performed the 10-meter walk test to determine walking speed, which is valid and reliable in the chronic stroke population [27]. Two trials of walking at self-selected comfortable and maximal speeds were performed. For self-selected comfortable trials, participants were instructed to walk at their normal comfortable speed. For maximal trials, participants were instructed to walk as fast as they safely could. Walking was performed without assistive aids whenever possible. There was a ~2-meter acceleration and deceleration zone. Time to complete trials was measured with a stopwatch, and the average of the two trials was calculated.

Walking symmetry

Participants performed four trials (two self-selected comfortable and two maximal trials) of walking across the GAITRite electronic walkway (classic 14’ model, CIR Systems Inc., NJ, USA), which is valid and reliable in this population [28, 29]. Using GAITRite software, we assessed step length, step time, swing time, and stance time for the paretic and non-paretic leg. Walking symmetry was expressed as the ratio (paretic/non-paretic) for each variable, where a value of 1 indicates symmetrical walking. Data from one participant was excluded because it was an outlier (more than 3 standard deviations above the mean), and data from another participant was excluded because of missing values.

Ankle motor control accuracy

Participants performed a motor tracking task with the paretic ankle using a custom ankle tracking device. As described previously [4, 30], participants were seated and the paretic ankle was secured to the device with the foot in neutral position. Ankle dorsiflexion and plantarflexion was performed to track a computer-generated sinusoidal waveform visually displayed on a monitor. Participants performed three familiarization/pre-test trials and twelve tracking trials (60-seconds in duration each). There was a one-minute break after every 4 trials. The amplitude of the sinusoid was determined individually based on maximal ankle range of motion for each participant. Ankle position was measured with Spike2 software (Cambridge Electronic Design, Milton, Cambridge, UK). To quantify ankle motor control accuracy, we calculated the root-mean-square error (RMSE) between the sine wave and the ankle position and normalized the RMSE value to the participant-specific sine wave amplitude [4, 30]. Accuracy was presented as a percentage: accuracy=(1-RMSE)*100, and averaged across the twelve tracking trials. Data from one participant was excluded as an outlier, (more than 3 standard deviations below the mean).

Statistical analyses

All data were tested for normality (Shapiro-Wilk Test) and homogeneity of variances (Levene’s Test). Outcome measures were compared between groups with different polymorphisms using independent samples t-tests and Mann-Whitney U tests. For variables with inhomogeneity of variances, the Welch-Satterthwaite method was used. SPSS Statistics software (IBM, Armonk, NY, USA) was used for all statistical analyses, and p≤0.05 was considered statistically significant.

Results

A total of 38 individuals with chronic stroke (mean age 59 years, 32 females) consented to be genotyped and participated in this study. Demographics are presented in Table 1.

Table 1:

Values are shown for all participants and for polymorphic groups for brain-derived neurotrophic factor (BDNF), catechol-O-methyltransferase (COMT), and apolipoprotein E (APOE). Values are counts, except age, which is Mean (SD). AA: African American; As: Asian; Bi: Biracial; C: Caucasian; F: Female; Hs: Hispanic; M: Male.

Demographics
	All participants	BDNF		COMT		APOE
	(n=38)	VAL (n=15)	MET (n=23)	VAL (n=19)	MET (n=19)	ε4+ (11)	ε4−(27)
Age (years)	59.3 (8.7)	57.8 (7.1)	60.0 (9.7)	58.4 (9.5)	60.0 (8.0)	59.2 (7.9)	59.5 (10.9)
Gender (M/F)	(32/6)	(12/3)	(20/3)	(15/4)	(17/2)	(9/2)	(23/4)
Race (AA,Hs,C,As,Bi)	(22,2,10,3,1)	(8,1,4,1,1)	(14,1,6,2,0)	(14,1,3,1,0)	(8,1,7,2,1)	(6,0,5,0,0)	(16,2,5,3,1)
Side affected (left/right)	(19/19)	(8/7)	(11/12)	(9/10)	(10/9)	(7/4)	(12/15)
Stroke type (hemorrhagic, ischemic)	(12/26)	(2/13)	(10/13)	(7/12)	(5/14)	(2/9)	(10/17)

Walking speed

Polymorphisms for the tested genes did not affect walking speed (Table 2). For BDNF, there was no difference between the VAL and MET groups for comfortable (t=−0.24, p=0.80, 95% CI: −0.23, 0.18) or maximal walking speed (t=−0.53, p=0.59, 95% CI: −0.33, 0.19). For COMT, there was no difference between the VAL and MET groups for comfortable (t=0.30, p=0.77, 95% CI: −0.17, 0.23) or maximal walking speed (t=−0.02, p=0.98, 95% CI: −0.27, 0.26). For APOE, there was no difference between the ε4+and ε4- groups for comfortable (t=0.31, p=0.75, 95% CI: −0.19, 0.26) or maximal walking speed (t=−0.06, p=0.95, 95% CI: −0.30, 0.28), (Fig 1).

Table 2:

Walking parameters are shown for comfortable and maximal speeds for polymorphic groups for brain-derived neurotrophic factor (BDNF), catechol-O-methyltransferase (COMT), and apolipoprotein E (APOE). Walking symmetry is presented as ratios (paretic/non-paretic). Values are Mean (SD).

Walking speed, walking symmetry, and ankle motor control accuracy
	BDNF		COMT		APOE
	VAL (n=15)	MET (n=23)	VAL (n=19)	MET (n=19)	ε4− (n=27)	ε4+ (n=11)
Comfortable speed
Walking Speed (m/s)	0.66 (0.30)	0.69 (0.31)	0.69 (0.32)	0.66 (0.28)	0.67 (0.29)	0.70 (0.35)
Step Length Ratio	1.11 (0.12)	1.06 (0.19)	1.06 (0.19)	1.10 (0.13)	1.11 (0.17)	1.02 (0.11)
Step Time Ratio	1.50 (0.26)	1.34 (0.41)	1.44 (0.43)	1.37 (0.28)	1.42 (0.37)	1.37 (0.35)
Swing Cycle Ratio	1.54 (0.30)	1.50 (0.33)	1.50 (0.35)	1.54 (0.29)	1.56 (0.35)	1.43 (0.20)
Stance Cycle Ratio	0.83 (0.63)	0.85 (0.70)	0.85 (0.07)	0.83 (0.06)	0.83 (0.07)	0.86 (0.06)
Maximal speed
Walking Speed (m/s)	0.87 (0.38)	0.94 (0.41)	0.91 (0.41)	0.91 (0.38)	0.91 (0.38)	0.90 (0.42)
Step Length Ratio	1.14 (0.10)	1.06 (0.20)	1.08 (0.18)	1.12 (0.16)	1.11 (0.15)	1.06 (0.20)
Step Time Ratio	1.37 (0.17)	1.32 (0.30)	1.38 (0.27)	1.31 (0.23)	1.33 (0.25)	1.37 (0.26)
Swing Cycle Ratio	1.40 (0.23)	1.38 (0.32)	1.41 (0.26)	1.37 (0.31)	1.39 (0.32)	1.39 (0.19)
Stance Cycle Ratio	0.85 (0.06)	0.86 (0.09)	0.86 (0.09)	0.85 (0.07)	0.86 (0.09)	0.85 (0.07)
Ankle Motor Control Accuracy
Average Accuracy	78.2 (12.79)	72.05 (12.6)	78.0 (10.5)	70.3 (12.7)	73.1 (12.1)	76.8 (15.6)

Fig 1. Comfortable and maximal walking speed.

Values are shown for the MET (presence of at least one Met allele) and VAL groups based on catechol-O-methyltransferase (COMT) polymorphism, MET (presence of at least one Me allele) and VAL groups based on brain derived neurotrophic factor (BDNF) and ε4+ group (presence of at least one ε4 allele) and ε4- group based on apolipoprotein (APOE). The Met allele for BDNF and COMT and the ε4 allele for APOE are considered the predicted detrimental alleles for the respective comparisons, values are the average and error bars represent the standard error.

Walking symmetry

For BDNF, there were no differences between the VAL and MET groups (Table 2). There were no differences for step length ratio during comfortable (U=102, z=−1.60, p=0.11) and maximal (t=1.4, p=0.15, 95% CI: −0.03, 0.18) walking, step time ratio during comfortable (t=1.3, p=0.19, 95% CI: −0.08, 0.40) and maximal (t=0.5, p=0.62, 95% CI: −0.13, 0.22) walking, swing time ratio during comfortable (t=0.42, p=0.67, 95% CI:−0.17, 0.26) and maximal (t=0.25, p=0.80, 95% CI:−0.17, 0.22) walking, or stance time ratio during comfortable (t=−1.4, p=0.3, 95% CI:−0.07, 0.20) and maximal (U=157.5, z=0.00, p>0.99) walking.

For COMT, there were no differences between the VAL and MET groups (Table 2). There were no differences for step length ratio during comfortable (U=194.0, z=1.35, p=0.18) and maximal (t=−0.7, p=0.50, 95% CI:−0.15, 0.07) walking, step time ratio during comfortable (t=0.60, p=0.55, 95% CI:−0.17, 0.31) and maximal (t=85, p=0.40, 95% CI:−0.09, 0.24) walking, swing time ratio during comfortable (t=−0.33, p=0.75, 95% CI:−0.25, 0.18) and maximal (t=0.34, p=0.74, 95% CI:−0.16, 0.22) walking, or stance time ratio during comfortable (t=0.71, p=0.50, 95% CI:−0.03, 0.06) and maximal (U=162, z=0.16, p>0.99) walking.

For APOE, there were no differences between the ε4+ and ε4- groups (Table 2). There were no differences for step length ratio during comfortable (U=158, z=1.21, p=0.24) and maximal (t=−0.10, p=0.31, 95% CI:−0.19, 0.06) walking, step time ratio during comfortable (t=−0.41, p=0.68, 95% CI:−0.32, 0.21) and maximal (t=0.41, p=0.68, 95% CI:−0.15, 0.22) walking, swing time ratio during comfortable (t=−1.37, p=0.18, 95% CI:−0.32, 0.06) and maximal (t=−0.03, p=0.97, 95% CI:−0.21, 0.21) walking, or stance time ratio during comfortable (t=1.20, p=0.24, 95% CI:−0.02, 0.08) and maximal (U=132.5, z=−0.17, p=0.87) walking.

Ankle motor control accuracy

25 participants were included in analyses of ankle motor control; 12 participants were excluded from analysis because they did not have sufficient ankle ROM or had missing values post data collection. Analyses were performed for BDNF (VAL: n=8, MET: n=17), COMT (VAL: n=12, MET: n=13) and APOE (ε4+: n=6, ε4-: n=19). Data from one outlier was excluded.

For BDNF, there was no significant difference between VAL and MET groups for ankle motor control accuracy (U= 46, z=−1.3, p=0.21). For COMT, there was no significant difference between VAL and MET groups for ankle motor control accuracy (U=51, z=−1.46, p=0.15). For APOE, there was no significant difference between ε4+ and ε4- groups for ankle motor control accuracy (U=45, z=−0.76, p=0.47), (Fig 2).

Fig 2. Ankle motor control accuracy.

Ankle motor control accuracy is a percentage of normalized RMSE value. Values are shown for the MET (presence of at least one Met allele) and VAL groups based on catechol-O-methyltransferase (COMT) polymorphism, MET (presence of at least one Me allele) and VAL groups based on brain derived neurotrophic factor (BDNF), and ε4+ (presence of at least one ε4 allele) and ε4- groups based on apolipoprotein (APOE). Values are the average across the twelve tracking trials, values are the average and error bars represent the standard error.

Discussion

In this preliminary study, we determined the effect of common missense polymorphisms for BDNF, COMT and APOE on walking speed, walking spatiotemporal symmetry, and ankle motor control. We did not find an effect of these polymorphisms on any of our outcome measures in our cohort of individuals with chronic stroke.

We found that BDNF polymorphism did not affect any measures of walking speed or walking symmetry. This is consistent with the lack of association between BDNF polymorphism and walking speed that has been reported recently [12]. The lack of association between BDNF polymorphism and measurements of lower limb function is contrary, however, to what has been found in the upper limb. In the upper limb, BDNF polymorphism was associated with motor learning and short-term plasticity in young healthy adults [31] and predicted upper limb motor recovery in stroke survivors with severe motor impairment [32]. Thus, BDNF polymorphism may differentially affect the upper and lower limbs, or other study differences may explain the disparate findings. Furthermore, walking post stroke involves compensatory mechanisms that may mask the effect of genetic polymorphism. Hence, we also examined a skilled, goal-directed movement to probe these differences, but still found no difference between BDNF groups. Our results are contrary to Joundi et al. (2012), who observed that healthy individuals with one BDNF Met allele had a significantly higher error during a visuomotor adaptation paradigm than individuals with no Met alleles [14]. We may have found an effect of BDNF polymorphism if we had used a more sensitive task. Alternatively, unequal sample sizes between the BDNF groups may have limited our ability detect a difference. It is also possible that differences in motor control based on BDNF polymorphism may not be apparent in individuals with chronic stroke; prior research has shown that BDNF polymorphism has a greater influence on stroke recovery during the acute stage (0–1 months) vs. the later stages of stroke [24].

COMT polymorphism also did not affect any measures of walking speed, walking symmetry, or ankle motor control. Previous studies have suggested that COMT polymorphism may affect functional outcomes differently depending on the number of Met alleles. In healthy individuals, carrying two Met alleles as compared to one Met allele is associated with reduced walking speed in older adults [18]. Young healthy adults with no Met alleles demonstrate better performance during a sequence learning task than individuals with one Met allele but not individuals with two Met alleles [17]. Individuals with stroke with two Met alleles demonstrated greater gross motor function impairment in the upper and lower extremity compared to those with no Met alleles [19]. In the current study, we only had 2 participants with two Met alleles and 17 participants with one Met allele. To compare to individuals with no Met allele (VAL group), we combined individuals with these genotypes into the MET group. Therefore, we were limited in our ability to isolate differences that may be related to the number of Met alleles carried by an individual. Overall, results regarding COMT and its effect on motor learning have been quite variable. Some of this variability may be explained by the theory of tonic and phasic dopamine signaling; carrying one Met allele is associated with high levels of tonic dopamine which may improve cognitive stability and help with sustained attention to a task, while carrying a Val allele is associated with phasic dopamine transmission which may improve cognitive flexibility and help with motor sequencing tasks [16]. The accuracy task used in this study may require greater cognitive stability than cognitive flexibility, and consequently, combining Val/Met heterozygotes and Met/Met homozygotes may have limited our ability to detect an influence of COMT polymorphism.

APOE polymorphism also did not affect any walking speed, walking symmetry, or ankle motor control outcome. The lack of influence of APOE polymorphism in the current study may be related to etiology or time since stroke. In individuals with stroke, APOE polymorphism may be important for those with hemorrhagic but not ischemic stroke [24]. In the current study, 9 of 11 individuals in the ε4+ group had an ischemic stroke. Additionally, the effect of APOE polymorphism is pronounced during the first 3 months of recovery [24] but not after one year of recovery [25]. In this study, our participants all had chronic stroke. These factors may have affected our ability to detect differences related to APOE polymorphism.

This study was one of the first to evaluate the effect of genetic polymorphism on walking and ankle motor control in chronic stroke. There are several limitations to this study. First, we had relatively small sample sizes and unequal distribution of genotype between groups. Second, this study involved a heterogeneous sample with respect to gender, age, comorbidities, and functional level. In the future, larger samples will allow further investigation of the impact of genetic polymorphism on function and motor learning. In conclusion, our preliminary study suggests that genetic polymorphism does not influence walking speed, walking symmetry, or ankle motor control in chronic stroke.