Early considerations of genetics in aphasia rehabilitation: a narrative review

ABSTRACT

Background

Early investigations linking language and genetics were focused on the evolution of human communication in populations with developmental speech and language disorders. Recently, studies suggest that genes may also modulate recovery from post-stroke aphasia.

Aims

Our goal is to review current literature related to the influence of genetics on post-stroke recovery, and the implications for aphasia rehabilitation. We describe candidate genes implicated by empirical findings and address additional clinical considerations.

Main Contribution

We describe existing evidence and mechanisms supporting future investigations into how genetic factors may modulate aphasia recovery and propose that two candidate genes, brain derived neurotrophic factor (BDNF) and apolipoprotein E (APOE), may be important considerations for future research assessing response to aphasia treatment. Evidence suggests that BDNF is important for learning, memory, and neuroplasticity. APOE influences cognitive functioning and memory in older individuals and has also been implicated in neural repair. Moreover, recent data suggest an interaction between specific alleles of the BDNF and APOE genes in influencing episodic memory.

Conclusions

Genetic influences on recovery from aphasia have been largely unexplored in the literature despite evidence that genetic factors influence behaviour and recovery from brain injury. As researchers continue to explore prognostic factors that may influence response to aphasia treatment, it is time for genetic factors to be considered as a source of variability. As the field moves in the direction of personalized medicine, eventually allied health professionals may utilize genetic profiles to inform treatment decisions and education for patients and care partners.

KEYWORDS:

• Aphasia
• treatment
• genetics
• rehabilitation

Introduction

Genetic links to speech and language development have been hypothesized since the early nineties (Atkinson et al., 2018; Enard et al., 2002; Fisher et al., 1998), but researchers have only recently begun to investigate how genetics can influence recovery from acquired communication disorders, such as aphasia. As evidence builds and our collective knowledge of the genetic influence on aphasia recovery grows, understanding genetic implications will be important not only for researchers but clinicians as well. All human illnesses have a hereditary component (Collins, 2001), making it pertinent that all health professionals must begin incorporating new collective knowledge of genetics into their work with the aim of developing a personalized medicine approach to client evaluation and treatment. Although genomic data are not currently accessible for the routine diagnostic purposes of allied health professions (Ashley et al., 2010; Curtis et al., 2016), it may be important for clinicians to begin considering the ways genetic testing may impact their work in the near future. The cost of genomic assays, or laboratory procedures to determine the presence of genetic variables of interest, continues to fall (Marshall et al., 2017) and for the last ten years serious discussions about the future of genomic medicine has been progressing (Kullo et al., 2013). Recently, we have seen achievements in the merger of available genomic assays with electronic health records to redefine past diagnoses (James et al., 2020). Considering other areas of medicine where genomics is more widely applied today, clinicians may be involved in counselling family members on what genetic testing might reveal medically and including the family in discussions of choosing appropriate interventions based on specific genetic profiles from a genomic dashboard that presents genomic reports to clinicians (Swaminathan et al., 2016). For example, using either a blood or saliva sample, diagnostic tests for genes implicated in cardiovascular disease could be used to guide future care of patients and families at risk for, or recovering from, stroke (Ganesh Santhi et al., 2013).

Although stroke is the most common cause of aphasia and stroke risk is influenced by genetic factors, it can be difficult to draw conclusions about the relationship between these genetic factors and the incidence of aphasia because not all individuals who survive a stroke experience aphasia. However, there is emerging evidence to suggest that genetic factors may also play a role in an individual’s ability to recover from aphasia (Fridriksson et al., 2018) and that knowing a patient’s genetic profile has the potential to inform vital treatment decisions. Moreover, genetics impact learning and memory (De Blasi et al., 2009; Egan et al., 2003; Gorbach et al., 2020), which can further influence recovery. Thus, the purpose of this manuscript is to provide an overview of the research to date on the influence of genetics on post-stroke recovery from aphasia. Although the body of research is limited, this manuscript will serve as an introduction to the topic and foundation for future work, highlighting the importance of considering genetics in aphasia rehabilitation.

We begin with a genetics primer and a brief description of early research investigating the genetic bases of speech and language, then proceed to discussing genetic factors influencing acute stroke outcomes and neuroplasticity, and candidate genes influencing neuroplasticity in the language system. We follow up with a mention of additional genes to consider, pharmacogenomic considerations, protection of genetic information, and future directions. Of note, epigenetic mechanisms, transcriptional regulation, and an exhaustive analysis of all genes involved in stroke recovery are beyond the scope of this review.

Genetics primer

Genes can influence both typical and atypical behavioural presentations, by influencing the expression of characteristics such as cognitive and linguistic skills, such as learning, memory, attention, lexical processing, phonological processing and semantic processing (Greenwood & Parasuraman, 2003). One mechanism for variation in genetics is the single nucleotide polymorphism (SNP, colloquially pronounced “snip”): a substitution of a single DNA base pair location for a different base pair that makes the DNA sequence differ from others in the population. For example, if most people have a “C” base pair (cytosine) at a particular location on a chromosome, a subset of the population may instead have an “A” base pair (adenine) at that same location. Accordingly, SNPs are a common genetic mechanism to confer unique traits and influence differences in disease risk across the population (Curtis et al., 2016; O’Donnell et al., 2011). When one or more SNPs can be reliably associated (a type of correlation) with a given disorder in a sample, that provides preliminary evidence that the genetic factor may be responsible for the disorder. Moreover, SNPs can be reliability associated with individual differences in cognition in unimpaired populations (Greenwood & Parasuraman, 2003). For example, SNPs in the Brain Derived Neurotropic Factor (BDNF) gene are associated with performance on immediate and delayed recall of a narrative in healthy adults (Egan et al., 2003). A polymorphism of the DRD4 is associated with the normal personality trait of “novelty seeking” (Benjamin et al., 1996) and efficiency of executive attention in healthy adults (Fossella et al., 2002), as well as attention deficit hyperactivity disorder (ADHD). Although SNPs can be associated with normal variations in cognition and language abilities, much of our knowledge addressing the relationship between genetics and language functioning comes from studies investigating individuals with developmental disorders who possess particular SNPs (Bartlett et al., 2004, 2002, 2014).

General screening of the whole human genome can be achieved through either 1) linkage studies, which correlate the transmission of particular genetic regions (i.e., sequences of DNA) with language-related phenotypes among family members, or 2) association studies, which correlate particular SNP alleles (i.e., a specific base pair of the two possible base pairs in an SNP) with disease vs disease-free status in samples of unrelated individuals (Li & Bartlett, 2012). For example, linkage studies have found evidence for genes within particular regions of chromosome 13 to be associated with specific language impairment (SLI; Bartlett et al., 2004, 2002) and sections of chromosomes 13, 15, and 16 to relate to the incidence of SLI and autism spectrum disorder (ASD; Bartlett et al., 2014). Association studies may be especially useful in investigating the role of particular SNPs on recovery from aphasia.

Early research investigating the genetic bases of speech and langauge

Historically, some genes have been thought to specifically influence speech and language development, whereas others may have broader implications on cognitive and linguistic functions as a result of their role in neurophysiology and recovery mechanisms after stroke. The latter will be discussed in subsequent sections of this manuscript. Regarding the former, the first major gene specifically implicated in speech and language development unique to humans was the FOXP2 gene (Enard et al., 2002), which was shown to be linked with a familial speech and language disorder involving severe articulation, linguistic, and grammatical impairments (Fisher et al., 1998). Enard et al., proposed that FOXP2 could be responsible for the enhanced linguistic capabilities of humans compared to non-human primates based on evidence that the gene was recently selected in our evolutionary history. However, the role of the FOXP2 gene in the evolution of human communication has recently been called into question, as Atkinson et al. (2018) demonstrated its presence in archaic human ancestors. That said, other genes have been linked to the evolution of human language, such as ROBO1, ROBO2, and CNTNAP2, which play a role in either reading or language disorders and were shown to change after modern humans’ evolutionary separation from archaic hominins (Mozzi et al., 2016). Despite uncertainty surrounding the evolution of human language, it is still widely recognized that genetics has a role to play in how we communicate. For a review of the genomic and epigenetic factors involved in various speech and language disorders, see, Guerra and Cacabelos (2019).

Genetic factors influencing acute stroke outcomes and neuroplasticity

The potential for recovery from acquired neurological injury, such as stroke, is influenced by a host of factors, including the size and location of the lesion in the brain (Chen et al., 2000; Moon et al., 2017; Watila & Balarabe, 2015); however, genotypes have only recently been considered in rehabilitation research, with most of the work focusing on motor recovery (Kim et al., 2016; Kleim et al., 2006; Di Lazzaro et al., 2015; McHughen et al., 2010; Qin et al., 2014; Shiner et al., 2016) and few studies investigating language rehabilitation (De Boer et al., 2017; Fridriksson et al., 2018). Broadly, genetics play a major role in risk for and presentation of neurological diseases (Arne & Jane, 2016; Ganesh Santhi et al., 2013; Gromadzka et al., 2005). There is evidence that genetic factors influence the occurrence and severity of stroke (Gromadzka et al., 2005), such as predisposition to risk factors (e.g., cerebrovascular disease), vascular outcomes, mortality, recurrence of cerebrovascular accident (CVA; Arne & Jane, 2016), and neuroplastic responses associated with functional recovery of a skillset, either spontaneously or after intervention (Fridriksson et al., 2018; Di Lazzaro et al., 2015).

Neural responses to acute cerebrovascular injury include recruitment of white blood cells, inflammatory mediators, and elevation in C reactive protein levels (Frizzell, 2005). Considering the different neurophysiological responses to acute ischemic and haemorrhagic CVA, there may be differences in genetic influences based on type of stroke. For ischemic strokes, ion pump disruption leads to increased intracellular calcium levels and overstimulation of membrane channels resulting in cell death (Frosch et al., 2005; Gannong, 2005; Messing, 2003). Genetic factors may influence cellular adaptation to ischemic injury in a variety of ways, including increasing inflammatory markers (Gromadzka et al., 2007; Marousi et al., 2011) and increasing clotting and thrombosis (He et al., 2015; Rocio et al., 2006). A meta-analysis by Math et al. (2019) found that individuals who had an ischemic stroke and Cytochrome P450 2C19, or CYP2C19, polymorphisms (specifically named CYP2C19 *2, CYP2C19 *3, and CYP2C19 *17) were 2.4 times more likely to have poor outcomes, potentially because they are poor metabolism variants that result in decreased metabolization of antiplatelet medications, such as clopidogrel (Qiu et al., 2015; Yi et al., 2016), rendering the medications less effective. The same review found that individuals with a polymorphism in the brain-derived neurotrophic factor (BDNF) gene were 1.6 times more likely to have poor outcomes after an ischemic CVA, but that alleles of the apolipoprotein E (APOE) and interleukin 6 (IL6) genes did not influence ischemic stroke outcomes.

For haemorrhagic strokes, blood compresses surrounding tissue resulting in anoxia and eventually necrosis of neurons. Macrophages clear the blood and tissue, forming a CSF-filled cavity in the long term. The APOE ε2 allele has been associated with increased mortality and overall poorer outcomes in intracerebral haemorrhagic strokes possibly because the ε2 variant increases vessel damage caused by β-amyloid deposits, resulting in larger hematoma volumes (Biffi et al., 2011). Furthermore, Math et al. (2019) found that individuals who had an intracerebral haemorrhage and were carriers of the APOE ε4 allele were 2.6 times more likely to have poor functional outcomes. Biffi et al. (2011) hypothesized that the APOE ε4 allele tends to influence intracerebral haemorrhagic strokes but not ischemic strokes because of the relationship between the polymorphism and coagulating ability. ε4 carriers have higher partial thromboplastin time (PTT) ratios, meaning they have a propensity for bleeding. Thus, the pathophysiology of the CVA seems to be directly influenced by genetics.

Neuroplasticity associated with post-stroke recovery occurs at both the cellular level (i.e., synaptic plasticity) and the systems level (i.e., representational plasticity; Pearson‐Fuhrhop & Cramer, 2010). Examples of neuroplastic changes include dendritic spine formation, pruning, remodelling, calcium channel regulation, changes in NMDA receptors, changes in AMPA receptor trafficking, (Pearson‐Fuhrhop & Cramer, 2010), and increases in long-term potentiation (i.e., the strength of connection between neurons as a result of excitatory stimulation in the presynaptic pathway; Bliss & Collingridge, 1993). Moreover, these recovery mechanisms may vary depending on lesion location (e.g., cortical versus subcortical), which could be differentially impacted by specific genetic polymorphisms (Arne & Jane, 2016). Stroke severity may also interact with genetics to influence recovery. For example, Di Lazzaro et al. (2015) suggest that less severe strokes may benefit from one polymorphism of the BDNF gene when using rehabilitative repetitive transcranial magnetic stimulation (rTMS), whereas more severe strokes may be negatively affected by this same polymorphism.

Kohen et al. (2008) found that the following polymorphisms in the SLC6A4 gene, STin2 VNTR and 5-HTTLPR, were associated with poststroke depression. Yang et al. (2011) found that depressed patients also have lower serum BDNF levels, which reflect the amount of BDNF circulating in the body based on a blood sample. Lower levels of serum BDNF in the acute phase of stroke rehabilitation have been associated with worse functional outcomes (Stanne et al., 2016). Thus, there seems to be a relationship between serum BDNF levels and stroke rehabilitation; however, the degree to which polymorphisms of the BDNF gene impact levels of circulating BDNF is still unclear (Terracciano et al., 2013). It is also unclear if stroke rehabilitation responses to SSRIs in post-stroke individuals are influenced by genotype (Arne & Jane, 2016).

Aerobic exercise and noninvasive brain stimulation techniques have been investigated as adjuvants to aphasia therapy, with serum BDNF levels as a hypothesized mechanism of effect. Harnish et al. (2018) found decreases in serum BDNF levels during the first six weeks of an exercise regimen, followed by an increase toward or at baseline during the next six weeks of exercise. The authors hypothesize a possible stress response upon beginning an exercise regimen, decreasing serum BDNF levels, with rising levels once the participants acclimate to the exercise program. It is unclear whether continuing an exercise intervention beyond the 12-week program would yield increases beyond baseline levels. Three of five participants showed better response to word-retrieval therapy after aerobic exercise than they did with no exercise, but this response was also present in a control participant after a stretching intervention. Since genetic analysis was not completed in this study, it is unclear whether polymorphisms of the BDNF gene may have affected serum BDNF levels.

Noninvasive brain stimulation techniques have shown promise in assisting with post-stroke aphasia recovery (Fridriksson et al., 2018; Marangolo et al., 2014), with BDNF concentrations as a possible mechanism of effect (Chervyakov et al., 2015). However, Marongolo (2014) found that bihemispheric transcranial direct current stimulation (tDCS) (i.e., exciting perilesional cortex while suppressing contralesional cortex that may exert inhibitory influences) improved language outcomes with therapy, but did not increase BDNF. Thus, future work is necessary to determine the relationship between serum BDNF levels and language rehabilitation, and the extent to which aerobic exercise or noninvasive brain stimulation techniques may modulate this effect depending on BDNF genotype.

Genetic differences are responsible for structural and functional variance in the brain (e.g., Thompson et al., 2002) and play a role in post-stroke motor rehabilitation (e.g., Shiner et al., 2016), but genetics are rarely considered in post-stroke clinical trials for language rehabilitation or clinical practice because of the lack of foundational research. One reason for the paucity of empirical data is that whole-genome sequencing is expensive and analysis of SNPs genome-wide requires much larger sample sizes than rehabilitation studies can typically accommodate, typically in the thousands of patients and controls. Alternatively, when candidate genes are identified, they can be investigated with much smaller sample sizes. Although many studies of stroke outcomes have investigated candidate genes, many have not been replicated (Arne & Jane, 2016), which is problematic. In addition, there is some evidence that genes may interact to influence stroke recovery (Arne & Jane, 2016; Ward et al., 2014), suggesting future work investigating multiple gene interactions may be warranted.

Candidate genes influencing neuroplasticity in the language system

Brain-derived neurotrophic factor (BDNF)

The BDNF gene encodes the BDNF protein, a neurotrophin that is important for learning and memory. BDNF regulates neuroplasticity molecules, including CREB and Synapsin-I, which play a role in synapse formation and axonal elongation (Chytrova et al., 2008; Vaynman & Gomez‐Pinilla, 2006; Vaynman et al., 2004). Variations, or polymorphisms of this gene, may account for differences in how it impacts experience-dependent neuroplasticity. BDNF has two common isoforms; the Val66Met polymorphism is a valine (Val) to methionine (Met) substitution at codon 66 of the amino acid (protein) chain. The Met isoform is less common than the Val isoform of the protein. It is estimated that approximately 30% of the population have the BDNF Met isoform, which is associated with reduced BDNF protein secretion (Egan et al., 2003), increasing risk of developmental language impairment (Simmons et al., 2010), decreased experience-dependent motor map reorganisation (Kleim et al., 2006), differences in motor system function, and greater error in short-term motor learning in healthy individuals (McHughen et al., 2010). There is also recent evidence to support that the Val66Met polymorphism is related to decreased motor system recovery (Shiner et al., 2016) and decreased brain activation patterns (Kim et al., 2016) after stroke. BDNF Met carriers showed worse post-stroke motor recovery with therapy, specifically for individuals with mild-to-moderate impairment who may have still had sufficient cortex to support cortical neuroplasticity (Shiner et al., 2016). Thus, there is evidence building that the BDNF Met polymorphism may negatively impact experience-dependent neuroplasticity after stroke and may impact long-term stroke outcomes (Stanne et al., 2014).

Acute Language Recovery. In terms of language recovery, to our knowledge there has only been one published study investigating the impact of BDNF genotype on acute post-stroke language recovery. De Boer et al. (2017) studied the BDNF Val66Met polymorphism in the initial stages of stroke recovery in a sample of 53 patients in an inpatient rehabilitation program. All participants were less than three months post-stroke. Findings showed large variability across all participants in changes on the Amsterdam Nijmegen Everyday Language Test (ANELT), which measures communication in daily life situations (Blomert et al., 1994), and the Boston Naming Test (BNT; Kaplan et al., 2001), but no significant difference between Met carriers and noncarriers. Several methodological considerations may have played a role in this finding.

First, the outcome measures, the ANELT and BNT, were not tied to therapy in a controlled, task-specific way and may not have been specific enough to capture subtle changes. There were large variations in improvement scores for both Met carriers and noncarriers in this early phase after stroke. De Boer and colleagues note that this variability decreased the chances of detecting genotype-related differences, whereby they report that the presence of the Met allele did not correlate with significantly worse therapy outcomes. However, the authors also note that in these early stages post-stroke, spontaneous recovery interacts with treatment, and thus, there is a need to investigate how BDNF may interact with language therapy in the chronic phase. Investigating the relationship between BDNF genotype and experience-dependent neuroplasticity in the chronic phase of aphasia recovery holds promise precisely because the impact of widely variable patterns of spontaneous recovery is decreased.

Another consideration that may have impacted the lack of correlation between BDNF genotype and treatment outcomes is that the amount of therapy and type of therapy for each participant varied. Participants underwent 2–5 hours of individually tailored therapy per week, with the first weeks to months focusing on cognitive-linguistic therapy to optimize language processing at the affected linguistic levels (i.e., semantics, phonology, syntax), and later stages focusing on communicative strategies. Therefore, differences in treatment dosage and treatment ingredients may have confounded the ability to detect differences between groups. Finally, severity of stroke was not considered in the analysis. The volume of the stroke may have impacted neuroplasticity due to the degree of available cortex for language recovery.

In sum, De Boer et al. (2017) took the first steps into investigating the potential effects of BDNF on aphasia recovery in the weeks and months following stroke. Although they found no significant effect of BDNF genotype on outcomes of aphasia therapy, it is important to note that methodological considerations and variability in spontaneous recovery may have played a role in the ability to detect differences during this early phase in rehabilitation.

Chronic Language Recovery. To our knowledge, there has also only been one published study examining the role of BDNF in treatment of chronic aphasia recovery. In an effort to better understand the relationship between the SNP variants of the BDNF gene that underlie the Val66Met polymorphism and by extension, neuroplasticity in recovery, Fridriksson et al. (2018) conducted a randomized controlled trial among 66 participants with aphasia using anodal transcranial direct current stimulation (A-tDCS). It is believed that A-tDCS enhances long-term synaptic plasticity, which is dependent upon BDNF. The authors used a 2 × 2 factorial design, crossing genotype (Met carriers vs non-carriers) and non-invasive brain stimulation (A-tDCS vs sham stimulation). Participants underwent genotyping to determine their genetic profile and were stratified into two groups based on whether or not they were Met carriers for the BDNF gene. Within each group, approximately half of the participants received A-tDCS and the other half received sham stimulation for the first 20 minutes of the behavioural treatment sessions. Behavioural treatment consisted of 45 min of computer-based naming therapy and sessions were conducted five times per week for three weeks. Fridriksson et al. hypothesized that individuals with the typical BDNF isoform (Val/Val) who underwent A-tDCS would respond better to the aphasia treatment than any other group: typical non-Met-carriers who received sham stimulation, atypical Met-carrying participants (Val/Met or Met/Met) who received sham stimulation, and Met-carriers who received A-tDCS. A significant interaction effect was found between genotype and stimulation condition, supporting their hypothesis. That is, those with the typical BDNF gene who received A-tDCS during treatment demonstrated significantly greater improvement in confrontation naming than individuals in the three remaining groups. Importantly, those with the typical BDNF genotype did have higher baseline language performance than the atypical BDNF group, although this difference was no longer significant after controlling for time post-stroke and lesion size.

Based on their findings, the authors speculate that BDNF production may have a greater influence on the chronic stage of recovery from aphasia than the acute phase. Moreover, this difference in BDNF production between those with and without the Met allele could contribute to the differences seen in baseline language performance between chronically aphasic individuals with typical and atypical variants of the BDNF gene. In fact, Kristinsson et al. (2019) have shown that Met allele carriers presented with more severe aphasia and demonstrated less fMRI activation (i.e., smaller number of active voxels) during an in-scanner naming task compared to non-carriers.

Although these results suggest that Met carriers may have less potential for neuroplastic recovery than non-carriers, Di Pino et al. (2016) suggest an alternative possibility. They first corroborate that non-invasive brain stimulation of the cerebral cortex is less effective in Met carriers, but secondarily propose that Met carriers may benefit more from enhanced subcortical plasticity, based on studies of mice (Qin et al., 2014, 2011). Similarly, human Met carriers show greater balance in interhemispheric cortical excitability (Di Lazzaro et al., 2015). In other words, Met carriers may not have worse absolute ability for motor recovery after stroke; rather, they may rely more on unimpaired subcortical plasticity because cortical plasticity is disrupted. Furthermore, since non-carriers may rely on intracortical plasticity more than Met carriers, lumping them together in clinical trials for treatments targeting cortical plasticity (e.g., non-invasive brain stimulation) may confound results. However, with greater information about how the Met allele differentially impacts intracortical and subcortical plasticity, we may be able to use genotyping to better match patients with appropriate treatments targeting the most promising neural substrates. For example, Di Pino and colleagues suggest that the use of dopamine agonists that act at the subcortical level may enhance plasticity for Met carriers, especially in rehabilitation of motor deficits. Thus, genetic information from a patient, including BDNF genotype, may eventually help with treatment planning decisions.

Given the promise of BDNF as a modulator of motor map reorganisation in neurotypical adults (Kleim et al., 2006) and treatment responsiveness in the motor domain after stroke (Shiner et al., 2016), as well as preliminary evidence supporting BDNF as a modulator of chronic aphasia recovery, additional research on the contribution of BDNF polymorphisms on chronic language recovery with therapy is warranted.

Apolipoprotein E (APOE)

Based on studies of mice with the APOE ε4 allele that demonstrated synaptic loss and impaired neuronal plasticity (Cambon et al., 2000; White et al., 2001), we can infer that APOE plays a role in neural repair processes. APOE has additionally demonstrated an influence on neuronal remodeling and protection (Cedazo-Minguez, 2007; Mahley & Rall, 2000). Evidence from human studies further indicate that the APOE ε4 allele, of which 28% of the general population are estimated to be carriers (McCarthy et al. 2016), influences neuroanatomy and cognitive functioning in older populations. For example, APOE ε4, like the BDNF Met carriers, has been associated with decreased hippocampal volume in isolation from cognitive performance (Bath & Lee, 2006; Burggren et al., 2008; Plassman et al., 1997) as well as in the setting of declining memory (Gorbach et al., 2020) and age-related cognitive decline (Caselli et al., 1999). In a large sample (N = 620) of elderly people without dementia, APOE ε4 was shown to impact episodic memory, both registration and recall (De Blasi et al., 2009). Hence, performance on a delayed word recall task was significantly worse for ε4 carriers than non-carriers. One explanation for this difference was that episodic memory deficits could be related to a reduction of hippocampal volume, found to be more pronounced in ε4 carriers (Cohen et al., 2001; Lind et al., 2006). The hippocampus is in part responsible for learning lexical, semantic and syntactic knowledge, specifically in the initial stages of language learning (Maguire & Frith, 2004; Opitz & Friederici, 2003). Thus, the presence of the APOE ε4 allele may contribute to reducing the volume of hippocampi, a structure known to be important to memory and language, particularly in recovery from aphasia (Meinzer et al., 2010).

The APOE ε4 allele has been linked with an increased risk of cerebrovascular disease (Curtis et al., 2016; Wilson Peter et al., 1996) and poorer behavioural recovery from acute stroke, as demonstrated by change in the NIH Stroke Scale Score at one-month post-stroke (Cramer & Procaccio, 2012). Arne and Jane (2016) also cite evidence of earlier death (Gromadzka et al., 2005) and poorer functional recovery (Cramer & Procaccio, 2012) post-stroke in ε4 carriers. Another study (Wagle et al., 2010) found that the presence of the ε4 allele was a significant risk factor for cognitive impairment at 13 months post-stroke (N = 104). Carriers demonstrated significantly worse verbal learning and memory than non-carriers at follow-up. Using estimates of pre-stroke verbal learning and memory via questionnaires, no difference was seen between carriers and non-carriers prior to stroke. Of course, one limitation of this work was the inability to assess verbal learning and memory pre-stroke, necessitating the use of questionnaires to estimate this information. Interestingly, verbal learning, as well as immediate and delayed memory were impaired, but visuospatial/constructional function was preserved in the ε4 carriers. To explain the association between ε4 and both memory and verbal learning impairment, the authors suggest perhaps an underlying mix of undiagnosed neurodegenerative disease (i.e., Alzheimer’s) and vascular pathology, since ε4 has been established as a risk factor for Alzheimer’s disease (Corder et al., 1993; Saunders et al., 1993). An alternative possibility is that the hippocampi were malformed pre-stroke, therefore after stroke these impairments manifested as stroke-related structural and functional changes of the hippocampi (Jak et al., 2007; Qiu et al., 2009). Additional evidence suggests that the integrity of the hippocampus is indeed associated with language training success in individuals with aphasia (Meinzer et al., 2010).

In sum, there is evidence to warrant further investigation of the influence of APOE genotypes on language rehabilitation post-stroke. Moreover, Ward et al. (2014) suggest that BDNF and APOE genotypes may interact to influence cognitive functions in healthy older adults, such that the presence of BDNF Met isoform may modulate whether APOE polymorphisms affect episodic memory. Therefore, in addition to investigating the influence of BDNF and APOE polymorphisms on stroke recovery individually, the literature supports the need for investigation of the contribution of a possible interaction effect between these two genes on post-stroke aphasia rehabilitation outcomes.

Additional genes to consider

1. It is outside the scope of this review to perform an exhaustive analysis of all genes that may contribute to vascular, neurological, and linguistic variables; however, we feel it is important to mention that additional genes described by Arne and Jane (2016) may play a role in aphasia rehabilitation For example, the COX-1, NINJ2, TLR4, COL3A1, and GPIIb, have been associated with neurological and physical deficits post-stroke, such as prognosis, stroke recurrence, and mortality. Genes including IGF-1, MPO, SIGMAR1, APOD, GPIIIa, COX-2, CYPC19, CRP, and COMT have been associated with functional outcomes from stroke either in the short-term or at three months into recovery. Finally, at least one gene, the serotonin transporter (officially called the solute carrier family 6 member 4 gene or SLC6A4), has been linked with social participation and depression after stroke. Ultimately, it is important to recognize that stroke outcomes are highly variable and influenced by many factors outside of genetics, including environmental factors. This makes it challenging to draw conclusions related to the specific role each gene may play. However, better understanding of the contributions of each of these genes to stroke recovery and how they interact has the potential to improve risk and recovery predictions for patients and their families.

Pharmacogenomic considerations

In addition to the direct influence of genetics on post-stroke aphasia recovery, there is also mounting evidence that genotype can influence response to various prescription medications used to treat conditions in this population. Medications used to treat cardiovascular disease, such as warfarin, antiplatelet drugs (e.g., clopidogrel), beta-blockers, and statins have demonstrated differences in effectiveness and safety based on genetic factors (O’Donnell et al., 2011). Certainly, the genetic influence on responsiveness to such drugs is relevant to stroke patients, who are likely to have symptoms of cardiovascular disease and may be prescribed these same medications.

Recent studies have investigated the use of pharmacological interventions in treatment specifically for post-stroke aphasia. A variety of drugs have been investigated, including catecholaminergic (i.e., bromocriptine, levodopa, dextroamphetamine, and amantadine), cholinergic (i.e., ameridin, donepezil, bimeflane, aniracetam, galantamine, and physostigmine), nootropic (i.e., piracetam), and serotonergic (i.e., selective serotonin reuptake inhibitors [SSRIs] such as escitalopram and fluvoxamine) compounds, among others (Saxena & Hillis, 2017). SSRIs have been especially promising when used to treat individuals in the subacute phase of aphasia who had damage to posterior superior temporal gyrus and the superior longitudinal fasciculus/arcuate fasciculus (Hillis et al., 2018). At least one combination SSRI-antipsychotic medication, fluoxetine-olanzapine, which can be used in treatment-resistant depression for post-stroke patients (Goldberg et al., 2015), has demonstrated an influence by various SNPs (Houston et al., 2012), indicating that genetics may influence response to some of these candidate pharmacological interventions.

Considering the evidence, it is plausible that additional pharmacological agents used to treat speech and language deficits in aphasia are also modulated by genetic factors, perhaps similar to the way BDNF expression has demonstrated an influence on response to tDCS when used in combination with behavioral speech therapy to treat aphasia (Fridriksson et al., 2018). Therefore, future studies investigating treatment of aphasia with pharmaceuticals should include genotyping among other potentially modulating variables of interest.

Protection of genetic information

Given the sensitive nature of genetic information, especially when it comes to health-related conditions, and the potential for stigmatization and discrimination, legal considerations are also crucial. For these reasons, legal safeguards were put in place under the 2008 Genetic Information Nondiscrimination Act (GINA) and the 2010 Patient Protection and Affordable Care Act (ACA) to ensure confidentiality and deter insurance companies, employers, or others from using discriminatory practices (Curtis et al., 2016). However, there may still be gaps in these laws and protections. Thus, it will additionally be important for SLPs and audiologists, as well as other allied health professionals, to become familiar with the legal implications of genetic testing and how results from such testing might impact their practice.

Conclusions and future directions

The field of aphasiology is just now beginning to investigate the role that genetics may play in the variability among patients undergoing treatment from post-stroke aphasia. Whole-genome sequencing studies require very large sample sizes, a feat that is quite difficult to achieve in language rehabilitation research. Two genes, BDNF and APOE, have demonstrated relationships to language, cognition, and neuroplasticity. Therefore, as a first step to expanding our understanding of genetic predispositions in the pathophysiology of post-stroke language recovery, BDNF and APOE are a rational starting point for genetic-based inquiry due to their biological relevance.

Preliminary data from one study supports that BDNF modulates neuroplasticity associated with A-tDCS and aphasia therapy in chronic aphasia (Fridriksson et al., 2018), but no such support was found in an acute language rehabilitation trial (De Boer et al., 2017). Future investigations into the relationship of BDNF on aphasia recovery should examine the effects of the Val66Met polymorphism on treatments that primarily target cortical plasticity (e.g., non-invasive brain stimulation) versus treatments that may target intracortical plasticity (e.g., dopamine agonist medications) to test the hypothesis by Di Lazzaro et al. (2015) that Met carriers are more reliant on intracortical plasticity.

No studies to date have investigated the relationship between APOE genotype and post-stroke language recovery. We believe that APOE is a good candidate to investigate because the ε4 allele is implicated in cognitive decline and impaired memory in aging; decreased hippocampal volume, which plays a role in language learning and potentially relearning (Meinzer et al., 2010); and poorer behavioural recovery on the NIH Stroke Scale one month after stroke. There is also reason to believe that polymorphisms of BDNF and APOE may have the potential to interact to influence behavioural outcomes (Ward et al., 2014).

As the field of aphasia rehabilitation makes the inevitable move toward more personalized medicine, genetic information is a source of treatment variability for which we must begin to account. Once sufficient evidence is available, systematic, or scoping reviews should be undertaken to better clarify the relationship between specific genes and language recovery after stroke. Clinically, individuals who are less likely to respond to intensive, restorative therapy, as determined in part by their genetic profiles, may choose to dedicate their time and resources toward compensatory approaches. Identification of genetic factors that impact treatment responsiveness will allow for better estimation of prognosis and improved triage of individuals into appropriate therapy regimens. Ultimately, the goal is to determine if genetics may be one influential factor among many in matching patients to appropriate treatments.

Acknowledgments

The authors on this review article were supported by the National Institute on Deafness and Other Communication Disorders of the National Institutes of Health under award number R01DC017711. The authors have no other competing interests.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the the National Institute on Deafness and Other Communication Disorders of the National Institutes of Health [R01DC017711].