Val66Met BDNF Polymorphism Implies a Different Way to Recover From Stroke Rather Than a Worse Overall Recoverability

Abstract

In search for individualized predictors of stroke recovery, the Val66Met polymorphism of the brain-derived neurotrophic factor (BDNF) is attracting great interest, because it has a negative impact on neurotrophin function. Since stroke recovery relies on brain plastic processes, on which BDNF is permissive, the dominant thought is in favor of a worse recovery in Met carriers. Conversely, we suggest that Met carriers do not differ in terms of absolute ability to recover from stroke, but they do differ on the way they recover. In particular, Met carriers rely more on subcortical plasticity, while ValVal patients more on intracortical plastic processes. Indeed, the direct evidence of impaired Met carrier recovery is inconsistent, as a high worldwide diffusion of the polymorphism suggests. The plasticity taking place in cortex, which is the one targeted by noninvasive brain stimulation strategies aimed at enhancing recovery, is less pronounced in Met carrier stroke patients, who have instead spared global recovery potential. Enhanced subcortical plasticity sustains better stroke recovery of Met carrier mice: this may also happen in humans, explaining the weaker interhemispheric cortical excitability imbalance recently described in Met carriers. Thus, BDNF haplotype determines mechanisms and structures involved in stroke recovery. The less pronounced cortical plasticity of Met carrier implies that plastic changes induced by interventional neurophysiological protocols would be better predictors of ValVal chronic outcome and those protocols would be more effective to boost their recovery. Other strategies, more focused on subcortical mechanisms, should be used in Met carriers.

Introduction

The burden of stroke is exceedingly high, especially in industrialized countries where it is among the causes of death with higher incidence and the main cause of permanent disability.^1,2 Stroke outcome is extremely heterogeneous; an early and individualized estimation of patients’ potentialities to recover can help in determining more realistic goals to aspire and the congruent allocation of resources by patients and rehabilitators. ³ Overall, better patients’ stratification can guide patient management more efficiently, in terms of hospital discharge and territorial follow-up. This has pushed forward the enquiry for precocious individualized predictors of stroke recovery.

To this aim, besides the well-known outcome predictors such as lifestyle, comorbidity, age, and initial impairment, ⁴ transcranial magnetic stimulation (TMS) may be of help. For instance, the early presence of the motor-evoked potential elicited stimulating the affected hemisphere anticipates a better hand recovery, ⁵ the increased excitability of the unaffected hemisphere correlates with poor outcome,^6,7 and beyond the basal excitability of both hemispheres, the amount of cortical plasticity that is induced by repetitive noninvasive stimulation of the brain is informative of better future outcome. ⁸ For a complete review on TMS-related predictors of stroke recovery, see Di Pino et al. ⁹

In search for individualized predictors of recovery, genetic factors are recently attracting great interest, ¹⁰ in particular, the ones related to the brain-derived neurotrophic factor (BDNF), a neurotrophin involved in synaptic plasticity^11,12 and in activity-dependent learning. ¹³

The substitution of a valine (Val) with a methionine (Met) in the precursor leads to a common haplotype—Val66Met (rs6265) ¹⁴ —characterized by impaired protein function and, to some extent, less pronounced motor cortex plasticity.^15,16 BDNF is present and active under basal conditions, but its release is boosted by activity. Indeed, physical exercise, motor cortex activation, learning, and sensory stimulation induce an increase of BDNF and of tropomyosin-related tyrosine kinase type B (TrK B) receptors. ¹⁷ Accordingly, stroke represents a potent inductor of the activity-dependent release of BDNF, which has been shown to be necessary for stroke recovery. ¹⁸ In this regard, BDNF Val66Met polymorphism results also in an impaired BDNF/TrKB signaling, responsible for a weaker stroke-induced release of BDNF. ¹⁴ This body of evidence triggered the multiplication of a large amount of studies that investigated the relationship between Val66Met BDNF polymorphism and stroke recovery and, to a lesser extent, the implication of haplotypes in the possibility to foster the recovery with noninvasive brain stimulation strategies.

Although the mechanisms responsible for the aftereffects produced by the noninvasive stimulation of the brain still need to be completely clarified, pharmacological studies suggest that they are strictly linked with synaptic plasticity. More in detail, TMS delivered with repetitive patterns (rTMS) produces its effect by acting at the cortical level on the function of interneurons projecting on pyramidal cells, ¹⁹ while transcranial direct current stimulation (tDCS), which consists in the application of weak constant currents through the scalp, seems to mainly act by modulating neuron resting membrane potential ²⁰ and requires BDNF secretion and TrkB activation. ²¹ While NMDA receptor antagonist abolishes the effects of both rTMS and tDCS, NMDA receptor agonist enhances LTP-like plasticity induced by anodal tDCS. ²² Thus, NMDA-mediated plasticity of neuronal circuits and synapses seems to be a key element in the induction of noninvasive stimulation dependent plasticity. Since NMDA-mediated plasticity is known to be significantly influenced by BDNF, ²³ depicting whether and how the Val66Met haplotype impacts neuromodulation will allow tailoring neuromodulatory interventions on different patients’ genotype. This may represent one of the possible solutions to fill the gap existing between the initial promising applications of these techniques to promote stroke recovery and the inconsistent clinical evidence achieved so far.^24,25

A multitude of other factors are at play in stroke recovery and influence long-term outcome. Among the others, genetic not BDNF-related factors (eg, Bcl-2, hypoxia inducible factors, apoptosis-inducing factors, COX-2, ApoE e4 ²⁶ ), epigenetic (eg, DNA methylation, histone acetylation/deacetylation, microRNA modulation ²⁷ ) are of relevance. This article specifically focuses on the impact of BDNF Val66Met haplotype on stroke recovery.

The Scientific Problem: The Impact of Val66met BDNF Haplotype on Stroke Recovery

To date, whether subjects heterozygote or homozygote for the Val66Met substitution (from now on Met carriers) have a worse outcome after stroke is matter of debate.^28,29 However, because of the role played by BDNF in healthy subjects and of the dependence of stroke recovery on long-term potentiation (LTP), ⁹ on which BDNF is permissive, ³⁰ the dominant thought is in favor of a disadvantage for Met carriers. The origin of the putative disadvantage given by the polymorphism has been ascribed to an impaired brain plasticity, but the existence and the possible mechanisms of this correlation are still unknown.

Notwithstanding the common thought, we suggest that (a) ValVal and Met carrier stroke patients do not differ in terms of absolute ability to recover after stroke, but (b) they differ in the way they recover and, in particular, (c) motor recovery relies more on subcortical plasticity in Met carriers and on intracortical plasticity in ValVal patients.

Heterozygote and homozygote Val66Met substitutions would probably have different impacts on stroke recovery; however, in agreement with many previous investigations,^23,31 we consider all of them as Met carriers since we think that the evidence achieved so far does not allow us to frame those genotypes independently.

To build our interpretation of the evidence, we exploited data coming from retrospective studies, epidemiologic considerations, neurophysiological evidence and stroke animal models. For the sake of clarity, those data are clustered in this order and presented, step by step, in the following sections.

ValVal and Met Carrier Stroke Patients Do Not Differ in Terms of Absolute Ability to Recover From Stroke

A large body of evidence is in favor of an impaired plasticity in Met carriers. For instance, they have poorer episodic memory, abnormal hippocampal activation, ¹⁴ and lack the training-induced expansion of cortical motor maps. ³² However, the reported direct evidence in favor of an impaired stroke recovery is not as much as consistent.

The analysis of retrospective studies does not lead to a sole result. If Kim and colleagues found that Val66Met polymorphism was associated with acute and long-term poor outcomes, and with worsening of physical disability and cognitive function, ²⁸ accordingly to Cramer and Procaccio, a poorer outcome of Met carrier ischemic stroke patients was present only at 1 month poststroke, but disappeared later. ³³ Moreover, poorer outcome and worse memory have been widely reported in Met carriers with subarachnoid hemorrhage^34-36 but not in hemorrhagic patients who developed ischemic stroke. ³⁷ Finally, Manso and colleagues, in a very consistent sample size of 546 patients, did not find a significant correlation between BDNF haplotype and outcome but only a significant interaction of BDNF with 3 other growth factors. ²⁹

A further clue may come from epidemiologic considerations: on top of all that, a spared absolute ability of Met carriers to recover is strongly sustained by the extremely high prevalence of the Val66Met substitution in the population, which ranges from 30% in Caucasians up to 70% in Asians. ³⁸ In our opinion, a polymorphism that would give a consistent disadvantage in recovery after brain injury would have hardly reached such worldwide diffusion.

ValVal and Met Carrier Stroke Patients Differ on the Way They Recover

What the revision of the available neurophysiological evidence suggests? TMS studies report that the acute stage of stroke is often characterized by an interhemispheric cortical excitability unbalance due to a decrease of excitability of the affected hemisphere ³⁹ paired to a relative increase of excitability of the unaffected one. ⁷

Investigating the impact of BDNF haplotype on this poststroke rearrangement, we recently described that, compared to ValVal stroke patients, Met carriers display a 9-fold weaker interhemispheric unbalance of cortical excitability in the acute stage. ³¹ According to an influential current model of stroke recovery, known as the interhemispheric competition model, ⁴⁰ the increased excitability of the unaffected hemisphere is particularly deleterious for recovery. Indeed, as this model determines, it is the result of the lack of inhibition from the hemisphere damaged by stroke, and in turn, it is responsible for an additional hampering inhibition toward an already damaged hemisphere. Thus, the lower interhemispheric imbalance observed in Met carriers should have led to an advantage in terms of a better, rather than worse, recovery.^40-42 However, the picture seems to be likely more complicated than the one this model is able to predict,^9,43 while it could be better outlined by a model that correlates recovery and interhemispheric balancing depending on the specificity of the single patient as the type, location, and extent of the damage (the balance-recovery bimodal model). This correlation might be negative for mild damages but positive for more severe damages. ⁹ Moreover, the interhemispheric competition model does not differentiate patients neither on the basis of BDNF haplotype nor on other individual genetic features.

TMS is able to probe ongoing plastic phenomena providing measures of brain excitability, interhemispheric excitability balance, and evaluating the effectiveness of neuromodulatory techniques on brain excitability itself. In fact, rTMS allows to actively interfere with brain plasticity. Along this line, Chang and colleagues applied 10 sessions of excitability enhancing rTMS over the hand cortex of the affected hemisphere aiming at boosting clinical recovery and found that Met carrier subacute stroke patients improved less than ValVal control patients. ⁴⁴ At a first look, this finding would suggest a general disadvantage for Met carriers. However, the minor improvement was confined to the limb direct pertinence of the stimulated brain area (upper limb), while no differences were found on the recovery of the affected lower limb. This can be explained by speculating that, in those patients, 2 different ways to recover were together at play: a BDNF haplotype-dependent rTMS-correlated recovery and a BDNF haplotype-independent recovery not influenced by rTMS.

Motor Recovery Relies More on Subcortical Plasticity in Met Carriers and on Intracortical Plasticity in ValVal Patients

Data from animal models are extremely precious to widen our point of view on the topic. Indeed, novel findings have been recently reported suggesting that Met carriers’ recovery plasticity takes place in different brain structures. In a relevant work, it has been shown that after middle cerebral artery occlusion, unexpectedly for the authors who designed the study, Met carrier mice had better chronic motor recovery than wild type, because of an enhanced structural and functional plasticity in the controlesional striatum, ⁴⁵ independently of the infarct size. ⁴⁶ Although the translation of findings from animal studies to humans has to be approached with utmost caution, it might be hypothesized that, also in humans, Met carrier chronic recovery relies more on subcortical plasticity, while ValVal patients’ recovery might rely more on cortical plastic changes.

This thesis perfectly matches with our recent findings on the presence of different interhemispheric cortical excitability balance in different BDNF haplotype. ³¹ Indeed, if in ValVal stroke patients intracortical plasticity could be the main determinant of recovery due to the functional cortical changes, a higher interhemispheric cortical excitability imbalance appears in ValVal as compared to Met carriers. ³¹

In summary, we suggest that Met carriers are not deficient in plasticity in general but only in the plastic processes that take place in the cortex: these processes are also the ones more easily evaluated and modulated by TMS and rTMS. The recovery potential would be overall spared in this population possibly due to different subcortical plastic processes that are able to compensate for the lack of pronounced cortical changes.

Implications and Recommendations

The propensity of brain cortices to undergo plastic changes can be evaluated by measuring the response to rTMS protocols, such as the intermittent theta burst stimulation (iTBS). iTBS is a subthreshold rTMS paradigm that, by modulating the activity of cortical interneurons, produces prolonged LTP-like changes in the stimulated cortex ⁴⁷ as well as long-term depression (LTD)-like phenomena in the cortex of the contralateral hemisphere. ⁴⁸ In a previous study, in a not-genotyped acute stroke population, we demonstrated that the LTP-like changes of the affected hemisphere induced by iTBS, and the LTD-like changes induced in the unaffected hemisphere, are reliable predictors of long-term outcome. ⁸ This has likely reflected the positive contribution of perilesional LTP to recovery ⁴⁹ and the reduction in the inhibitory drive from the unaffected hemisphere to the affected hemisphere. ⁴⁰

Our present analysis implies that different BDNF haplotypes would determine different susceptibility to iTBS-induced plasticity. Indeed, whether in the recovery of ValVal stroke patients, but not in Met carrier patients, intracortical plasticity would play the lead role, then the response to iTBS would be a better predictor of chronic outcome for ValVal stroke patients than it would be for Met carriers. Hence, despite Met carriers being less unbalanced, their reacquisition of the balance would be less predictive of their recovery.

In this regard, we envisage to evaluate the ability of the cortical plastic changes induced in the acute state to predict chronic outcome in different BDNF haplotypes; this could confirm the present interpretation of the evidence and, by restricting the correlation to only ValVal subjects, could lead to achieve an even more accurate prediction than previously reported. ⁸

The knowledge of the mechanisms and structures differently involved in stroke recovery depending on different BDNF haplotype could have indeed relevant impact. For instance, it could pave the way for a more rational, genotype-driven, choice of the neuromodulatory protocol to enhance the recovery and of the brain area that could represent the best target. ⁹ Noninvasive brain stimulation of cerebral cortex would be more effective in ValVal subjects, whereas alternative strategies targeting the intact subcortical drive to the cortex (ie, with dopamine agonists^50,51) might turn to be useful to foster the recovery in Met carrier stroke patients.

Conclusions

In a recent work we highlighted the limitations of the prediction of recovery based only on the evaluation of single electrophysiological parameters (eg, the cortical excitability interhemispheric balance), without taking into consideration the specificities characterizing individual patients. ⁹ Here again, the more we investigate the mechanisms of recovery after stroke, the more we understand that the prediction of recovery, as well as the choice of the better therapy to apply, should go through a careful patient stratification. This work suggests alternative, but luckily equivalent, mechanisms through which the recovery is achieved in subjects carrying different BDNF haplotypes, supporting the inclusion of BDNF genotype among the parameters that should be routinely evaluated in stroke patient stratification, as part of a multimodal diagnostic approach that takes into account also complex interactions between factors.