Neuroplasticity — Exercise-Induced Response of Peripheral Brain-Derived Neurotrophic Factor
- 3.1k Downloads
Exercise is known to induce a cascade of molecular and cellular processes that support brain plasticity. Brain-derived neurotrophic factor (BDNF) is an essential neurotrophin that is also intimately connected with central and peripheral molecular processes of energy metabolism and homeostasis, and could play a crucial role in these induced mechanisms.
This review provides an overview of the current knowledge on the effects of acute exercise and/or training on BDNF in healthy subjects and in persons with a chronic disease or disability. A systematic and critical literature search was conducted. Articles were considered for inclusion in the review if they were human studies, assessed peripheral (serum and/or plasma) BDNF and evaluated an acute exercise or training intervention. Nine RCTs, one randomized trial, five non-randomized controlled trials, five non-randomized non-controlled trials and four retrospective observational studies were analysed. Sixty-nine percent of the studies in healthy subjects and 86%of the studies in persons with a chronic disease or disability, showed a ‘mostly transient’ increase in serum or plasma BDNF concentration following an acute aerobic exercise. The two studies regarding a single acute strength exercise session could not show a significant influence on basal BDNF concentration. In studies regarding the effects of strength or aerobic training on BDNF, a difference should be made between effects on basal BDNF concentration and training-induced effects on the BDNF response following an acute exercise. Only three out of ten studies on aerobic or strength training (i.e. 30%) found a training-induced increase in basal BDNF concentration. Two out of six studies (i.e. 33%) reported a significantly higher BDNF response to acute exercise following an aerobic or strength training programme (i.e. compared with the BDNF response to an acute exercise at baseline). A few studies of low quality (i.e. retrospective observational studies) show that untrained or moderately trained healthy subjects have higher basal BDNF concentrations than highly trained subjects. Yet, strong evidence still has to come from good methodological studies.
Available results suggest that acute aerobic, but not strength exercise increases basal peripheral BDNF concentrations, although the effect is transient. From a few studies we learn that circulating BDNF originates both from central and peripheral sources. We can only speculate which central regions and peripheral sources in particular circulating BDNF originates from, where it is transported to and to what purpose it is used and/or stored at its final destination. No study could show a long-lasting BDNF response to acute exercise or training (i.e. permanently increased basal peripheral BDNF concentration) in healthy subjects or persons with a chronic disease or disability. It seems that exercise and/or training temporarily elevate basal BDNF and possibly upregulate cellular processing of BDNF (i.e. synthesis, release, absorption and degradation). From that point of view, exercise and/or training would result in a higher BDNF synthesis following an acute exercise bout (i.e. compared with untrained subjects). Subsequently, more BDNF could be released into the blood circulation which may, in turn, be absorbed more efficiently by central and/or peripheral tissues where it could induce a cascade of neurotrophic and neuroprotective effects.
Neuroplasticity refers to the ability of the brain and CNS to adapt to environmental change, respond to injury and to acquire novel information by modifying neural connectivity and function. Neurotrophins support (activity-dependent) neuroplasticity; in particular, they are capable of signalling neurons to survive, differentiate or grow.[1, 2, 3, 4, 5] Therefore, neurotrophins gain increasing attention in research for the treatment and prevention of neurodegenerative and, more recently, metabolic diseases.[5, 6, 7, 8, 9, 10] Neurotrophic factors not only play a role in neurobiology, but also in central and peripheral energy metabolism. Their effect on synaptic plasticity in the CNS involves elements of cellular energy metabolism and in the periphery they take part in metabolic processes such as enhancing lipid oxidation in the skeletal muscle via activation of AMPK (i.e. adenosine monophosphate-activated protein kinase).
Physical activity and, in particular, acute exercise and training seem to be key interventions to trigger the processes through which neurotrophins mediate energy metabolism and in turn neural plasticity.[1, 2, 3,13, 14, 15, 16, 17] Of all neurotrophins, brain-derived neurotrophic factor (BDNF) seems to be the most susceptible to regulation by exercise and physical activity.[2,3,5] BDNF is a basic protein of 252 amino acids that is coded by the BDNF gene. This gene extends over 70 kb, is located on chromosome 11, band p13 and contains 11 exons and 9 functional promoters.[19, 20, 21] As in all other neurotrophins, BDNF has a single coding exon; the 3′ exon that encodes for most of the protein. Recently, a variant in the human BDNF gene has been identified, Val66Met, a single nucleotide polymorphism (SNP) at nucleotide 196 (G/A) that encodes an amino acid substitution (i.e. a valine [Val] to methionine allele [Met]) at codon 66 in the prodomain of the BDNF gene.[22,23] This gene mutation occurs in 20–30% of the human population[24,25] and results in a decreased activity-induced response of BDNF. Casey et al. predict that carriers of the variant BDNFMet allele will have less neurotrophic support for plasticity at a certain moment in their development, whereas carriers of the BDNFVal allele will experience the inverse.[25,26]
It is generally accepted that BDNF has a wide repertoire of neurotrophic and neuroprotective properties in the CNS and the periphery; namely, neuronal protection and survival, neurite expression, axonal and dendritic growth and remodelling, neuronal differentiation and synaptic plasticity such as synaptogenesis in arborizing axon terminals, and synaptic transmission efficacy.[27, 28, 29, 30, 31] Animal studies also revealed a neuroendocrine and/or metabotrophic capacity of BDNF in the periphery, which (i) reduces food intake; (ii) increases oxidation of glucose; (iii) lowers blood glucose levels; and (iv) increases insulin sensitivity.[32, 33, 34, 35, 36] In addition, Molteni et al. found that, in animals, a high-fat diet reduces hippocampal levels of BDNF, but exercise is able to reverse this dietary decrease. Komori et al. showed a central interaction between the adipocyte-derived hormone leptin that plays a key role in regulating appetite and energy metabolism and BDNF expression in the hypothalamus of mice. A human case study revealed a clinical phenotype of impaired cognitive function, hyperactivity and severe obesity associated with a chromosomal inversion of a region encompassing the BDNF gene and a reduction of serum BDNF. Additionally, Araya et al. found that serum BDNF was increased in insulin-resistant, overweight and obese subjects after a reduced-calorie diet. These findings confirm that BDNF is not only essential in the neuronal system, but is also intimately connected with central and peripheral molecular processes of energy metabolism and homeostasis.[11,41]
In search of mechanisms underlying plasticity and brain health, exercise is known to induce a cascade of molecular and cellular processes that support (brain) plasticity. BDNF could play a crucial role in these induced mechanisms. Therefore, since the early 1990s, studies started to investigate the effects of physical activity, acute exercise and/or training on BDNF concentration, first in animals[42, 43, 44, 45, 46] and then, since 2003, in humans. The first human study examined the effect of acute exercise on peripheral BDNF in subjects with a neurodegenerative disease (i.e. multiple sclerosis [MS]) in order to explore the restorative potential of exercise. Since then, two dozen other studies on the effects of acute exercise and/or training on BDNF have been conducted of which most concern healthy subjects. The purpose of the current review is to provide an insight in the overall effect of physical activity on peripheral concentration of BDNF.
1. Literature Search Methodology
1.1 Search Strategy
1.2 Criteria for Consideration
Studies were selected using predetermined inclusion and exclusion criteria. An initial raw screening resulted in a selection of 860 articles. A more profound screening of titles, abstracts and full-text articles, based on specific criteria, resulted in a final selection of 24 studies. Figure 1 shows the progress of the literature screening and the reasons for inclusion or exclusion.
Inclusion criteria were as follows: healthy subjects; persons with a chronic disability or disease; acute aerobic and strength exercise protocols (low to high intensity); endurance/aerobic, strength/resistance training protocols (low to high intensity); randomized controlled trials; controlled trials; clinical trials; comparative and evaluation studies; assessment of peripheral (serum and plasma) BDNF concentrations; and articles written in English, French, Dutch or German. Studies were excluded when they concerned animals, no exercise/training intervention, no physical activity, behavioural studies, reviews, studies on cognitive learning, no assessment of peripheral BDNF and general studies on neuroplasticity/neurogenesis. Inclusion and exclusion criteria were selected to be able to give an answer to the question whether acute exercise or training has an effect on peripheral BDNF, in particular, in humans. This question is of interest as acute exercise and training could be a viable treatment of neurodegenerative and metabolic diseases through their possible effect on neurotrophins and, thus, neuroplasticity. Four studies with no acute exercise or training intervention were nevertheless included in this review because of their possible relevant contribution. The four studies research the relation between the level of physical fitness and basal peripheral BDNF concentration.
1.3 Data Extraction
2. Exercise and Peripheral Brain-Derived Neurotrophic Factor (BDNF)
The main purpose of this literature review is to provide an insight in the effects of exercise and/or training on peripheral concentration of BDNF. The second purpose is to review the materials and methods that were used to research the effects of exercise and/or training on BDNF. The following sections summarize the study populations, exercise protocols, biochemical analysis, basal BDNF concentrations and the effects of exercise or training on peripheral BDNF in all included studies.
2.1 Number and Type of Studies
Twenty-four studies were included; nine studies were randomized controlled trials,[50,54,56,60,66, 67, 68, 69,71] one was a randomized non-controlled trial, five were non-randomized controlled trials,[47,51,57, 58, 59] five were non-randomized, non-controlled comparative trials (the study of Rojas Vega et al. has a corrigendum that was published a year later; we always refer to this study and the corrigendum)[62, 63, 64, 65,70,75] and four studies were retrospective observational studies.[52,53,55,61]
2.2 Study Populations
The sample size of trials that were included in this review varied from 8[62, 63, 64] to 55 subjects with a mean sample size of 24 subjects. For the four retrospective observational studies, sample sizes were larger, ranging from 26 and over 44 and 75 to 85 subjects. Proof of evidence would become more solid if all studies included an a priori power analysis to determine the appropriate sample size.
Study populations were drawn from several sources; for example, general population,[47,54,57,60,66] students, athletes,[56,63,64] spinal cord injured (SCI) athletes, persons with major depression,[58,59] cognitive impairment or MS.[47,51,67] Thirteen studies examined both males and females,[47,50, 51, 52, 53, 54, 55,57, 58,60,67,69, 70] while nine studies examined only males[56,61, 62, 63, 64, 65,68,71, 72,75] and one study only females. The mean age of participants in all the included studies ranged from 20.8 ± 0.6 years to 70.0 ± 8.3 years. Three studies examined a population of the elderly (i.e. mean age ≥55.0 years)[50,55,59] and no study that included children or adolescents (i.e. mean age ≤18.0 years). Lommatzsch et al. showed that basal concentrations of BDNF significantly changes with increasing age. Katoh-Semba et al. stated that children and adolescents could be prone to changes in neurotrophines due to maturation and growth. Therefore, it might be interesting to study possible differences in effects of acute exercise and training on peripheral concentration of BDNF between young and old healthy subjects or in young and old persons with a chronic disease or disability.
In most of the included studies, it is not always clear whether it concerns untrained, moderately trained or well trained subjects. Studies should report on the level of fitness, expressed in maximal oxygen uptake (V̇O2max) or maximal power output, of their study population. It is likely that the effects of acute exercise and training on peripheral BDNF depend on the physical fitness of the subjects, as BDNF could be involved in processes of energy metabolism.[37,40,79]
2.3 Exercise Protocols
Twenty out of 24 studies applied an exercise intervention. In general, four different interventions can be distinguished as follows: an acute aerobic or strength exercise; and an aerobic or strength training programme.
2.3.1 Acute Exercise Protocols
Graded exercise tests (GXTs) should be distinguished from acute aerobic exercise protocols of long or short duration. Sixteen of 20 interventional studies carried out a GXT until exhaustion a few days prior to the intervention or as an intervention on its own. In these studies, GXTs are mainly performed to determine the intensity of an acute aerobic exercise or training protocol. In three studies, a GXT was used as an isolated intervention to study its effect on circulating concentrations of BDNF.[54,59,75] In these cases, a GXT is evaluated as a short acute exercise of high intensity (table III). In two studies a GXT was part of a prolonged acute exercise protocol of high intensity1.[55,62,63] Protocols of all GXTs can be found in table II.
Fifteen of 20 studies applied an acute aerobic exercise intervention (table III). Seven of those studies (table III) investigated the effect of both low to moderate and high-intensity aerobic exercises,[54,56,58,63, 64, 65,68,71] five studies focused only on exercises of low to moderate intensity[47,51,67,69,70] and three on the effects of an isolated high-intensity exercise[59,62,75] on concentration of BDNF. The protocols of the acute exercise interventions differ in each study, which makes it difficult to compare between studies. Nevertheless, all studies could be categorized according to their exercise intensity (i.e. based on exercise load and duration) [table III]. In four studies the acute exercise intervention was part of the test protocol before and after an aerobic training programme. The effect of an aerobic training programme on the BDNF response from rest to the end of a standardized acute exercise of low, moderate or high intensity was studied.[51,67,68,75]
Recently, the relation between an acute strength exercise session and concentration of BDNF was researched in two studies.[57,72] Goekint et al. and Yarrow et al. used an acute strength exercise session to analyse the change in BDNF from rest to immediately post-exercise and this was repeated at the end of a strength training programme.
Table III shows that the moments of blood acquisition for analysis of BDNF are similar in most of the 16 studies on acute exercise: (i) at baseline; (ii) immediately following a low-, moderate- or high-intensity strength or aerobic exercise; and (iii) 15-60 minutes following the acute exercise. In two cases blood was not collected immediately following the acute exercise[51,70] and only Castellano and White collected blood more than 60 minutes following the acute exercise.
2.3.2 Exercise Training Protocols
2.4 Blood Sampling and Biochemical Analysis
For the analysis of free circulating peripheral BDNF, blood serum (16 studies) is preferred to that of blood plasma (eight studies). This could be due to the fact that blood serum has been the conventional standard for most biochemical analysis although, generally, the choice between blood serum and plasma is determined by the requirements of the individual laboratory. In some studies, preference is given to blood serum because the addition of anticoagulants (e.g. heparin or EDTA) in blood plasma can activate blood platelets and change the concentration of the constituents to be measured.[84,85] Concentrations of serum BDNF are approximately 200-fold higher relative to those of plasma BDNF, indicating that low concentrations of BDNF are circulating free in the blood and higher amounts of BDNF are stored in platelets or in immune cells.[86,87] Moreover, platelets circulate for up to 11 days in peripheral blood, whereas BDNF protein circulates in plasma for <1 hour, indicating that platelets could be a storage compartment and its BDNF could represent a long-term marker of varying plasma BDNF concentrations.[87,88] To finally unravel the link between plasma and serum BDNF, measurement of both plasma and serum BDNF could be interesting in future studies.
When plasma is being used, the type of anticoagulant that is added to the collection tubes is not clarified and only one study corrected plasma BDNF for platelet reactivity. However, Schneider et al. pointed out that some anticoagulants (i.e. EDTA) may activate blood platelets and thus influence the concentration of plasma BDNF ex vivo. Rasmussen et al. and Seifert et al. centrifuged blood plasma a second time to ensure that platelets were spun down and thus removed from the surfactant. Nevertheless, to ascertain plasma BDNF is not influenced by BDNF stored in platelets, a correction for platelet reactivity is recommended. For both plasma as well as serum, details of centrifugation are lacking in 12 studies and details of storage temperature after centrifugation are missing in eight studies. In the studies of Gold et al. and Schulz et al., blood serum was analysed for BDNF; nevertheless, they report on the use of heparinized tubes for the collection of blood samples. As a result, it is not clear whether they analysed serum or plasma BDNF. Additionally, only two of the 24 studies[71,72] reported on corrections of BDNF concentrations for changes in serum or plasma volume following acute exercise or training. BDNF values could change due to haemoconcentration following acute exercise or pseudo anemia following training.[90,91] Kargotich et al.[92,93] pointed out that moderate to intense exercise results in a decrease of blood volume or also haemoconcentration. As a result, changes in blood solutes after exercise or training could represent an inherent change in haemoconcentration due to shifts in blood volume instead of a real exercise-induced change in BDNF concentration.[92,93] Future studies should present corrected serum and plasma BDNF concentrations (i.e. corrected for the shift in plasma volume by the formula of Van Beaumont and colleagues).
In all studies, the diagnostic biochemical technique used to detect BDNF in blood serum or plasma is the ELISA. Trajkovska et al. showed that ELISA-kits (ChemiKine™; Millipore, Billerica, MA, USA) are an accurate, valid and reproducible analysis tool for peripheral BDNF. ELISA-kits of different manufacturers were used and details on the sensitivity of the assay or intra- and interassay variations were not always given (table V). Guidelines on how to handle the samples from collection until analysis or on storage conditions are not provided in the manuals of any of the kits. With regard to the biochemical analysis of BDNF, laboratories and/or manufacturers of analysis kits should reach a uniform consensus on the peripheral assessment of BDNF from the collection of blood until the analysis with ELISA. Meanwhile, researchers should repeatedly use uniform collection and analysing techniques within their own laboratories.
2.5 Basal Concentrations of BDNF
2.5.1 Healthy Subjects
For serum, basal BDNF concentrations in healthy subjects range from 1.5 to 30.9 ng/mL throughout the 24 studies. Literature confirms that basal values of serum BDNF in healthy subjects (non-athletes) vary extremely.[77,86,95, 96, 97, 98, 99] Values could be influenced by different factors such as diurnal fluctuations, physical fitness, age, sex, bodyweight, nutrition and possible neurological, immunological or metabolic disorders.[77,100] A remarkable finding is that basal serum BDNF values in the studies of Floël et al. 1.5 ± 0.5 ng/mL, Currie et al. (7.2 ± 2.7 ng/mL), Gold et al. (4.7 ± 0.5 ng/mL) and Rojas Vega et al.[63,64] (5.8 ± 1.9 ng/mL), were low compared with values in all other studies. The study of Rojas Vega et al.[63,64] concerned recreational athletes and the study of Currie et al. included predominantly subjects engaged in some level of recreational and sport-based activity. Two other studies confirm that basal BDNF concentrations in athletes could be lower than in untrained subjects,[52,61] while the studies of Zoladz et al. and Seifert et al. disagree with this. A lower level of BDNF in trained subjects and athletes could indicate that BDNF clearance in trained subjects or athletes is more effective (i.e. a higher disappearance rate), with less stored or circulating BDNF in the periphery as a result. Alternatively, plasma volume increases by 10–20% following regular physical training, thus lower levels of BDNF could merely represent the shift in blood volume instead of a true increase in BDNF.[90,91] The study of Floël et al. indicates no correlation between the level of physical activity and basal BDNF, and five out of seven studies on strength or aerobic training in healthy subjects showed no short term effect on basal concentration of BDNF.[56,57,60, 61,72] Nevertheless, it should be noted that more studies with a longer duration of the training period and in different populations (i.e. trained versus untrained, healthy versus diseased) are necessary to elucidate whether basal plasma and serum BDNF concentrations are influenced by the level of physical fitness/activity.
Not only physical fitness, but also factors such as sex, age, bodyweight and nutrition could influence basal concentration of BDNF. For instance, the low basal serum BDNF value in the studies of Gold et al. and Floël et al. could be due to sex effects. (i.e. 6/14, 28/47 male/female, respectively). According to Lommatzsch et al., women display significantly lower concentrations of platelet BDNF levels than men (i.e. groups matched for bodyweight) because of the sex-specific differences in BDNF expression of resident cells or organs and thus experience an altered uptake of BDNF into platelets. However, Gustaffson et al., Katoh-Semba et al. and Ziegenhorn et al. found no sex-related differences regarding serum and plasma BDNF, while Trajkovska et al. and Baker et al. reported higher concentrations of whole blood and plasma BDNF, respectively, in women. The mean basal serum BDNF level (i.e. 30.5 ± 6.9 ng/mL) measured in healthy control women in the study of Laske et al. is the second highest value reported in 13 of the included studies. Basal values of serum BDNF could also be influenced by age and/or bodyweight. Katoh-Semba et al. found an increase of serum BDNF over the first several years in healthy individuals and then a slight decrease after reaching adult age (i.e. mean level in 30- to 39-year-old age group). Also, Ziegenhorn et al. observed a decreasing concentration of serum BDNF with increasing age, while Lommatzsch et al. reported the same in plasma concentration of BDNF but, on the other hand, no age-related influence on platelet concentration of BDNF. Floël et al. measured basal BDNF concentration in 75 healthy older individuals (i.e. mean age 60.5 ± 6.9 years) and reported the lowest serum BDNF concentration of all 24 included studies (i.e. 1.5 ± 0.5 ng/mL). On the other hand, Laske et al. reported very high concentrations of serum BDNF in healthy older female controls (mean age 58.9 ± 6.6 years; [BDNF]serum 30.5 ± 6.9 ng/mL). Monteleone et al. suggest that serum BDNF concentration is also increased with increasing bodyweight. The studies of Floël et al. and Castellano and White both studied healthy (control) subjects with a mean BMI > 27, yet found very different basal concentrations of serum BDNF (i.e. 1.5 vs 20.2 ng/mL). It is likely that, next to age or sex, alternations in energy balance and nutritional variables also influence peripheral BDNF concentrations, yet it is still unclear in which direction.
In plasma, basal BDNF concentrations in the 24 included studies range from 10.3 pg/mL to 2.5 ng/mL in healthy subjects. Thus, basal plasma values of BDNF vary even greater than in serum values. Three other studies in the literature also report strongly varying resting BDNF values, ranging from 77.0 over 92.5 to 1700.0 pg/mL.[77,86,96] Overall, the large variations in plasma BDNF concentration confirm the hypothesis of Lommatzsch et al. and Ziegenhorn et al. that peripheral BDNF is stored, for most of the time, in the blood platelets and varying concentrations of BDNF are released from the platelets upon agonist stimulation and circulate free in the blood plasma, depending on the specific need of BDNF in certain tissues. Moreover, not only between studies, but also within a study, large variations in basal plasma BDNF concentrations can be observed, as pointed out by the large standard deviations of plasma BDNF concentrations in five studies.[60,61,66,75,96] This clearly indicates that basal plasma BDNF concentrations are extremely fluctuating. In the six studies on plasma BDNF concentration, it is not clear which type of anticoagulant was used. Only in the studies of Rasmussen et al. and Seifert et al. was it specified that tubes containing EDTA were used. It is possible that the type of anticoagulant influences platelet activation in vitro and, thus, plasma BDNF concentration. Only Baker et al. adjusted plasma BDNF levels for the contribution of activated platelets to ensure only plasma BDNF is measured.
Peripheral BDNF is also subject to sex-related diurnal variations.[102, 103, 104] In men, plasma BDNF peaks in the morning and decreases substantial during the day similar to the cortisol circadian rhythm.[102,103] This rhythmic circadian variation and correlation with cortisol levels is less explicit in plasma BDNF of women.[103,104] In women, hormonal fluctuations blunt the diurnal rhythm related to cortisol. Researchers should take these circadian hormonal fluctuations into account when measuring plasma BDNF.
2.5.2 Persons with a Chronic Disease or Disability
Only one study investigated SCI athletes and recorded a basal serum BDNF concentration of 37.2 ng/mL. Yet, the sample size of the study was small (n = 8), it concerned athletes and there was no control group to verify this result. In persons with MS, basal serum BDNF values range from 4.4[47,67] to 10.0 ng/mL. This means that in MS, basal serum BDNF concentrations were significantly lower (i.e. ≤10.0 ng/mL) compared with the pool of BDNF data we found in the literature for healthy subjects. Nevertheless, Gold et al. found equal concentrations of BDNF in MS and healthy controls in their study. Gold et al. used stable persons with MS (i.e. persons with an acute relapse were excluded from their study). Castellano and White, on the other hand, reported significantly lower values of basal BDNF in MS versus healthy controls. According to Azoulay et al., lower concentrations of serum BDNF in MS could suggest that there is a reduction in tissue protection by BDNF or that there is an increase in the consumption of BDNF by the CNS due to damaged tissue. In the elderly and in persons with MS, the low level of BDNF at rest could also be due to a reduced production of BDNF as a result of lower levels of messenger RNA (mRNA); this can be confirmed by animal studies.[107, 108, 109]
Gustafsson et al. and Laske et al. both studied the effects of exercise on BDNF in subjects with major depressive disorder (MDD). Only Laske et al. found decreased basal serum BDNF concentrations in female patients with MDD. Gustafsson et al. included only moderately depressed patients. According to Ströhle et al. basal BDNF concentration in patients with panic disorder is also decreased. Thus, most studies on persons with a chronic disease or disability found deviating concentrations of serum BDNF compared with levels in young, healthy, untrained subjects. This is in agreement with other studies on neurodegenerative and metabolic diseases. Altered peripheral BDNF concentrations could be observed in persons with depression, anorexia nervosa (serum BDNF is decreased), and in persons with allergic asthma and obesity (serum BDNF is increased). All point values on basal BDNF concentration can be found in table V.
2.6 Exercise-Induced Response of BDNF
2.6.1 Effect of an Acute Aerobic Exercise
2.6.2 Effect of an Acute Strength Exercise
Goekint et al. and Yarrow et al. were the first to study the exercise-induced BDNF response following an acute strength exercise. Yarrow et al. reported a significant strength exercise-induced increase of 32% in serum BDNF while Goekint et al. did not find a significant change (i.e. 3.6%) in BDNF following an acute strength exercise session. Goekint et al. speculated that exercise intensity in their study was too low (i.e. six strength exercises of 3 × 10 repetitions at 80% of 1 repetition maximum (1RM) with relatively large resting periods between efforts). However, Yarrow et al. implemented two strength exercise protocols of different intensity as follows: (i) traditional resistance exercise/training (TRAD), which incorporates two strength exercises of 4 × 6 repetitions at 52.5% 1RM concentrically and eccentrically; and (ii) eccentric-enhanced resistance exercise/training (ECC+), which incorporates two strength exercises of 3 × 6 repetitions at 40% 1RM concentrically and 100% 1RM eccentrically. They found similar transient increases in BDNF in both groups, independent of training intensity. However, the groups were matched for training volume. Presumably, an acute strength exercise stimulates peripheral BDNF on the condition that the exercise load is intensive enough.
2.6.3 BDNF Response during Passive Recovery
The increase of peripheral BDNF following an acute aerobic or strength exercise is transient. In most studies BDNF concentration returned to baseline within 10–60 minutes post-exercise, showing a fast disappearance rate of circulating BDNF after cessation of an aerobic or strength exercise. Castellano and White and Yarrow et al. observed a significant decrease below baseline concentration in peripheral BDNF concentration 2- and 3-hours post-exercise, both in persons with MS and healthy (control) subjects. The exercise-induced response of peripheral BDNF seems to include an elevated release of BDNF into the blood circulation on the one hand and a greater tissue absorption on the other hand.
2.6.4 Effect of Aerobic Training
Schiffer et al., Castellano and White and Schulz et al. reported no training-induced effect on basal BDNF concentration in healthy subjects[51,66] and in persons with MS, respectively.[51,67] Only Zoladz et al. and Seifert et al. observed an increase in basal plasma BDNF5 concentration following a period of aerobic training. However, Zoladz et al. did not include a control group. Furthermore, Seifert et al. found that BDNF from the brain (vena jugularis) was increased but no differences were seen in plasma BDNF concentration in a peripheral artery. Also, they included overweight men in their study who lost weight and body fat due to aerobic training. Furthermore, the training group experienced a higher loss in adipose tissue than the control group so that elevated basal BDNF concentration in the training group could be the result of altered energy metabolism.[10, 11, 12] Baker et al. reported an elevated basal plasma BDNF concentration in men with mild cognitive impairment following aerobic training as compared with a stretching programme. In women they reported the inverse phenomenon. In both sexes plasma cortisol concentration fluctuated accordingly to plasma BDNF (i.e. plasma cortisol concentration increased in men and decreased in women relative to controls). When the change in BDNF from rest to post-acute exercise was assessed at the end of an aerobic training period, Zoladz et al., Schulz et al. and Seifert et al. reported a significant increase in peripheral BDNF concentration (307.1%, 365.2% and 100.0%, respectively) following an acute exercise compared with basal BDNF concentrations at the end of the training period. Nevertheless, Schulz et al. and Seifert et al. observed the same acute exercise-induced increase (i.e. 100.0%) at baseline. Thus, only according to Zoladz et al. an aerobic training programme elevates the BDNF response to an acute exercise; however, they did not use a control group in their study. The findings of Zoladz et al., Schulz et al. and Seifert et al. are in contradiction with those of Castellano and White, where a decrease in serum BDNF concentrations following an acute exercise was found at the end of the training period. However, also at baseline, BDNF concentration decreased following an acute exercise; assessment of peripheral BDNF was not performed immediately following the acute exercise, but 30 minutes, 2 hours and 3 hours following the acute exercise. Similar results could be found in persons with MS and healthy (control) subjects, although the disappearance rate of BDNF in persons with MS following the acute exercise differed significantly between week 4 (86%) and week 8 (59%).
Thus, results differ when it comes to the point of a BDNF response to aerobic training. Three studies6 observed an elevated basal plasma BDNF concentration following a training period.[50,68,75] Remarkably, these studies use a more intensive training protocol as compared with the other studies[51,66,67] (i.e. training ratio, respectively, 4–7 ×/week as compared with 2–3 ×/week and training intensity at a higher percentage of the heart rate reserve or V̇O2max) [table IV]. As with strength training, only one study on aerobic training observed a greater plasma BDNF response to acute exercise following a period of aerobic training (i.e. as compared with the BDNF response to acute exercise at baseline). However, this study lacked a control group. More studies are requested to unravel the benefits of aerobic training on peripheral BDNF concentration and/or cellular processing.
2.6.5 Effect of Strength Training
Four studies examined the effect of strength training on peripheral BDNF concentration in healthy subjects.[57,60,66,72] Protocols can be found in table IV. The studies of Goekint et al., Levinger et al., and Schiffer et al., used similar strength training protocols. Yarrow et al. limited strength training to two exercises and used two groups who trained the same volume but at different training intensities (i.e. traditional versus eccentric-enhanced strength training).
All studies agree that strength training has no effect on basal peripheral BDNF concentration. However, Goekint et al. and Yarrow et al., also studied the BDNF response to a single strength exercise following a strength training programme. Only Yarrow et al. reported a significant increase of serum BDNF concentration post-acute exercise, both at baseline and after completion of a strength training programme. Moreover, the change in BDNF from rest to immediately post-acute exercise was 98% greater at the completion of the 5-week strength training programme than at baseline. Except for the study of Levinger et al., who investigated middle-aged individuals with clusters of metabolic risk factors, no other study investigated the effects of strength training on peripheral BDNF concentration in persons with a chronic disease or disability. From four studies, the inquiry for whether strength training influences peripheral concentrations of BDNF remains inconclusive. It can be assumed that a strength training programme does not elevate basal BDNF concentration; therefore, maybe strength training protocols are not strenuous enough (i.e. training ratio of 3 ×/week) [table IV]. Strength training could possibly trigger a greater BDNF response to acute exercise in trained as compared with untrained subjects, although more research is necessary to support this hypothesis.
2.6.6 Basal BDNF in Trained versus Untrained Subjects
Zoladz et al. reported a lower concentration of plasma BDNF in untrained subjects compared with trained subjects. The other studies concerning this topic did not perform an experimental intervention protocol but observed the concentration of peripheral BDNF in trained and untrained subjects.[52,53,61] Moderately trained, untrained or low cardio-respiratory fit subjects seem to have higher concentrations of serum levels of BDNF than trained subjects. This was observed in athletes and untrained subjects at rest by Currie et al., Nofuji et al. and Chan et al. As suggested earlier (section 2.5.1), a lower level of BDNF in trained subjects and athletes could indicate that BDNF clearance is more effective (i.e. a higher disappearance rate) than in untrained subjects. However, no clinical trial has been performed to support this hypothesis. Moreover, Floël et al. could not find a correlation between BDNF concentration and level of physical activity. A longitudinal randomized clinical trial with trained and untrained subjects in a crossover design could give a more definite answer to whether basal BDNF concentration in trained subjects is lower than in untrained subjects. Moreover, it is also possible that lower concentrations of BDNF could represent the shift in blood volume instead of a true increase in BDNF as plasma volume increases by 10–20% following regular physical training.[90,91]
2.7 Guidelines for Future Research
An acute aerobic exercise induces an increase in peripheral BDNF concentration in healthy subjects, as well as in persons with a chronic disease or disability. Most studies found a statistically significant dose-response relationship between the intensity of an acute aerobic exercise protocol and concentration of BDNF. Although, Seifert et al. measured BDNF concentration at different exercise intensities (i.e. starting from 60% to 100% of V̇O2max) and could not observe a change in BDNF response. Independently of the intensity of the exercise, BDNF concentration generally returns back to baseline within 15–60 minutes and tends to decrease below baseline after 60 minutes.
From the results of the 24 included studies, it is difficult to determine the essential exercise parameters (i.e. intensity, duration and mode) that are necessary to induce an increase in peripheral BDNF concentration. It is most likely that exercise parameters are related to each other and that BDNF response is triggered when exercise becomes strenuous. Therefore, it would be interesting to objectivate exercise intensity and relate it to ratings of perceived exertion following exercise and to monitor mean and maximal heart rate values. When it comes to the mode of acute exercise, we can conclude from the included studies that acute aerobic exercise is more likely to elevate BDNF concentration than acute strength exercise. Presumably, in the two studies on acute strength exercise, the load of the exercise is not intensive enough for the given subjects to influence basal BDNF concentration. Moreover, in strength exercise resting periods between efforts are often implemented, which results in a decrease in heart rate, lactate, minute ventilation etc. between repetitions and between sessions. Within acute aerobic exercise cycling, hand cycling, running, rowing and stepping all bring on an exercise-induced BDNF response. To better understand the relationship between the nature of an acute exercise and peripheral BDNF concentrations, future studies should look at dose-response relationships between percentages of increase in BDNF on the one hand, and exercise parameters such as intensity, duration and mode of exercise on the other. The nature of an acute exercise and also subject characteristics, determine the BDNF response; in healthy subjects high-intensity exercise is more likely to induce a BDNF response, while in persons with a chronic disease or disability an exercise of low to moderate intensity seems already strenuous enough to trigger a BDNF response. Metabolic response to exercise or training is known to differ between healthy subjects and persons with a disease or disability and, therefore, probably also BDNF response, will be different. Thus, more studies are required on the BDNF response to different exercise intensities, both in healthy subjects and in persons with a disease or disability. Next to this, Chen et al. and Egan et al. point out that activity-dependent secretion of BDNF is diminished in subjects with the BDNF polymorphism, Val66Met. Future studies should, whenever possible, identify whether subjects are carriers of the genetic variant of the BDNF gene (Val66Met) [i.e. 20–30% of human population[24,25]. Also, for an appropriate comparison between studies, it would be better if changes in BDNF concentration would be expressed in relative units (e.g. in percentages of changes ± the standard deviations). Finally, it would not only be interesting to determine the amount of increase in BDNF, but also the disappearance rate of peripheral BDNF following acute exercise (e.g. BDNF assessment at 0, 30, 60 and 120 minutes post-exercise or by marking BDNF protein in vivo).
Three of six studies reported an aerobic training-induced response on basal BDNF while none of the four studies on strength training was capable of elevating basal BDNF concentration. Two of six training studies (i.e. one on aerobic and one on strength training) observed an elevated circulating BDNF response to an acute exercise following a training period. Because of the greater effect of strength training on insulin-like growth factor (IGF)-1 production compared with aerobic training, there could be a possible differential effect of aerobic and strength exercise and/or training on peripheral concentration of BDNF. It is well known that IGF-1 is needed to transform pro-BDNF into BDNF in the CNS and that IGF-1 easily crosses the blood-brain-barrier. Nevertheless, with regard to training-induced effects on peripheral concentration of BDNF, inconsistent findings and too few studies impede to draw a distinct conclusion whether a training period elevates basal BDNF concentration and/or upregulates the cellular processing of BDNF. Future studies on training-induced BDNF response should systematically investigate effects on basal BDNF concentration but also on BDNF response to an acute exercise. Another issue which demands some attention, is whether exercise-induced BDNF increase is limited or not and which physiological and/or environmental factors determine the ceiling effect.
2.8 Origin of Exercise-Induced BDNF Response
The cellular origin of the exercise-induced BDNF response remains partially unclear. Recently, Rasmussen et al. found evidence for a release of BDNF from the brain, as BDNF is able to cross the blood-brain barrier.[114,115] They reported that the brain has a significant BDNF production both at rest and during prolonged exercise (i.e. 2- to 3-fold increase of the production at rest) in healthy subjects. Confirming these results, Seifert et al. very recently showed that indeed BDNF is released from the brain (vena jugularis), and that aerobic training in obese subjects increases basal BDNF concentration. The brain contributes for almost 75% of circulating BDNF, suggesting that the brain is the major, but not sole, contributor to circulating BDNF in healthy subjects. Yet, it remains to be elucidated from which regions in the CNS and the brain (e.g. hippocampus, cerebral cortex, prosencephalon, cerebellum, hypothalamus) BDNF originates. Animal studies agree that exercise principally upregulates messenger RNA (mRNA) expression in the hippocampus.[4,68,116] A quarter of circulating BDNF seems to stem from a peripheral source. It has been speculated that the exercise-induced BDNF response originates partially from the contracting muscle cells. However, Matthews et al. showed that, in vitro, BDNF is indeed synthesized by skeletal muscle cells during contraction, and that muscle-derived BDNF is not released into circulation but used to enhance fat oxidation in the muscle cell. However, more studies are necessary to confirm that, in vivo, skeletal muscles do not release BDNF into the circulation following high-intensity contractions. Several other studies revealed some sources of BDNF within the blood circulation. Initially, low concentrations of BDNF in plasma suggest that BDNF is normally not present in the circulation but it is stored in the blood platelets until activation. Yamamoto and Gurney stated that platelets contain BDNF mRNA derived from the cytoplasm of the megakaryocyte and that they release BDNF protein upon agonist stimulation. Consequently, platelets might also synthesize BDNF protein. However, platelets have limited protein synthesis capacity and Fujimura et al. found extremely low or no concentration of BDNF mRNA in blood platelets. Therefore, it is more likely that BDNF is sequestered from the blood circulation, and originates for a major part in the brain and for some part elsewhere.[62,96,119] Other sources of human BDNF circulating in blood plasma and serum or stored in blood platelets could be peripheral cells or endocrine organs such as vascular endothelial cells, immune or peripheral blood mononuclear cells (e.g. T and B lymphocytes,[120, 121, 122] eosinophils,[123,124] monocytes,[121,122,125] vascular smooth muscle cells, the pituitary gland, salivary glands such as the submandibular glands).[128,129]
Synthesis and release of BDNF into the blood circulation increases as a result of a physical stimulus in a dose-response manner. The more intense the acute stimulus or (positive) stress is, the greater the BDNF response. Normalization of the BDNF concentration occurs when the stressor disappears, indicating that BDNF is used or stored elsewhere or/and that elevated BDNF secretion has ceased. Following exercise, peripheral BDNF clearance could be elevated indicating that circulating BDNF is used in the periphery or that BDNF is transported via the blood circulation to the brain where it crosses the blood-brain barrier to enhance neural health. When the same stimulus or stressor is repeatedly administered, for example an exercise training programme, it is possible that a ceiling effect occurs because subjects become accustomed to the stimulus or enriched condition and homeostatic mechanisms take over again. To verify this hypothesis, it would be interesting to perform a longitudinal study comparing training-induced BDNF response in well trained subjects to sedentary subjects. Recently, Berchtold et al. showed that both daily exercise and alternating days of exercise increased rat hippocampal BDNF protein, and concentrations progressively increased with longer running duration. They found that hippocampal BDNF protein remained elevated for several days after exercise ceased and could easily be induced again by another brief exercise exposure. These findings could indicate that peripheral BDNF is indeed retrogradely transported and used in the brain following exercise, and that long lasting effects of exercise on BDNF are only traceable in the CNS.
This review gives a summary of the current knowledge on the exercise-induced response of BDNF in healthy subjects and persons with a chronic disease or disability. When interpreting the results of this review, the selection process must be kept in mind. Taken together, it is difficult to compare studies because of varying study populations, sample sizes, different acute exercise and training protocols and different biochemical analysis techniques. Research concerning acute exercise or training interventions should clearly define protocol parameters that result in a functional benefit regarding peripheral BDNF concentration. Also, exercise-induced changes in BDNF should be expressed in relative units (for example in percentages of changes ± the standard deviations).
Overall, an acute aerobic exercise unmistakably influences circulating BDNF concentration, although the effect is transient. In healthy subjects it is rather unlikely that regular exercise (i.e. training) results in an elevated basal BDNF concentration, although the current amount of studies is insufficient to be able to exclude any training-induced effect on basal BDNF. However, the BDNF response to exercise is most probably an epiphenomenon of what happens centrally, and exercising regularly could induce central effects without elevating peripheral basal BDNF concentration. Circulating BDNF probably originates for a large part in the brain; we can only speculate where the other part of circulating BDNF originates from, where it is transported to and for what purpose it is used or stored at its final destination. Furthermore, future research has to show whether repeated administration of an enriched condition, stimulus or stressor (i.e. acute exercise, training and/or reduced-calorie diet) influences the efficiency of the cellular processing of BDNF or basal BDNF concentration (i.e. increase of basal BDNF and disappearance rate). Instead of a training programme consisting of the same exercise protocol during each training session, a protocol with a new exercise stimulus for every few sessions could be implemented.
Whether the use of acute exercise and training is viable for the treatment of neurodegenerative and metabolic diseases (seen from its effects on peripheral BDNF), seems plausible. Recent studies show that BDNF plays a role in regulating central (i.e. through interaction with leptin and through the hypothalamic pathway that controls bodyweight and energy homeostasis) and peripheral energy metabolism (i.e. BDNF as a contraction-inducible protein in skeletal muscle). As a result, central and/or peripheral BDNF could possibly mediate some of the health benefits of exercise in metabolic disorders. On the other hand, effects of repeated exercise on peripheral BDNF concentration could be important with regard to the treatment and prevention of neurological diseases and impairments such as MS, Parkinson’s disease and SCI. Long-term effects of exercise on symptoms of neurodegenerative diseases and neurological disorders and its relation with BDNF, have not yet been investigated. The synergistic effect of a combination of BDNF-stimulating factors such as acute exercise or training, changes in the nutrient content of a diet, and key pharmaceuticals could be a next step to study.
It should be noted that in the studies of Rojas Vega et al.[63,64] and Gustafsson et al., an acute exercise of low to moderate intensity preceded the GXT. This could influence the effect of a GXT on peripheral BDNF levels. The preceding exercise of low to moderate intensity, together with the GXT, has also been evaluated as a prolonged acute exercise protocol of high intensity and will be discussed in section 2.6.1.
In the study of Laske et al., BDNF concentration in healthy control subjects did not increase following an acute exercise of high intensity.
In the study of Gustafsson et al., a significant increase in [BDNF]p following an acute exercise of high intensity was only found in male control subjects.
In the study of Gustafsson et al., a significant increase in [BDNF]p following an acute exercise of low to moderate intensity was only found in male persons with MDD.
Seifert et al. found an increase in [BDNF]p measured in the vena jugularis but not in arterial [BDNF]p following an aerobic training programme.
Baker et al. only found an increase in men with mild cognitive impairment.
The preparation of this article was funded by the Vrije Universiteit Brussel and by the Research Foundation Flanders. The authors have no conflicts of interest that are directly relevant to the content of this review.
- 26.Gratacos M, Gonzalez JR, Mercader JM, et al. Brain derived neurotrophic factor Val66Met and psychiatric disorders: meta-analysis of case-control studies confirm association to substance-related disorders, eating disorders and schizophrenia. Biol Psychiatry 2007; 61 (7): 911–22PubMedCrossRefGoogle Scholar
- 57.Goekint M, De Pauw K, Roelands B, et al. Strength training does not influence serum brain-derived neurotrophic factor. Eur J Appl Physiol. Epub 2010 May 14Google Scholar
- 58.Gustafsson G, Lira CM, Johansson J, et al. The acute response of plasma brain-derived neurotrophic factor as a result of exercise in major depression. Psychiatry Res 2009; 94 (12): 1159–60Google Scholar
- 84.Banfi G, Bauer K, Brand W, et al. Use of anticoagulants in diagnostic laboratory investigations and stability of blood, plasma and serum samples [report no. WHO/DIL/LAB/99.1 rev. 2]. Geneva: World Health Organization, 2002Google Scholar
- 96.Fuijimura H, Altar CA, Chen R, et al. Brain-derived neurotrophic factor is stored in human platelets and released by agonist stimulation. J Thromb Haemost 2002; 87: 728–34Google Scholar