Brain-derived neurotrophic factor (BDNF) is widely expressed in the central nervous system. This protein binds to tropomyosin-related kinase B (TrkB) receptor to promote neuronal growth, survival, differentiation, and synaptic plasticity.
2,3 Studies have proposed that BDNF has an important role in PD. BDNF is necessary for the establishment of dopaminergic neurons in the SNpc, and 70% of these neurons express BDNF protein and TrkB.
2,3 Furthermore, BDNF protects dopaminergic neurons against neurotoxin-induced neuronal death in animal models of PD.
4 Therefore, failure or decrease in neurotrophy supported by BDNF could be an etiologic factor for PD.
1Studies have investigated a relationship between the development of PD and the presence of a common polymorphism in the BDNF gene, Val66Met (rs6265). BDNF Val66Met is a single nucleotide polymorphism (G to A) that results in the substitution of valine (Val) for methionine (Met) at position 66 in the prodomain region. This alteration decreases the interaction of the precursor form of BDNF (pro-BDNF) with sortilin (which mediates the transport of pro-BDNF). This leads to decrease in BDNF dendritic distribution, reduced BDNF transport to secretory granules, and impairment of the activity-dependent secretory pathway of BDNF.
2,5 Several studies investigated a relationship between the Met variant of BDNF Val66Met polymorphism and depression, anxiety, and cognitive performance in healthy individuals.
6,7 Also, a positive association with cognitive impairment,
8 planning ability
9 and early development of dyskinesia
10 was observed in PD patients.
Nevertheless, only one case-control study observed homozygosity of A allele for the BDNF Val66Met polymorphism as a risk factor for PD, in a Japanese population.
11 Other studies failed to report such association
12,13 and even described the opposite association (G allele more frequent in PD).
14 Association studies of this polymorphism are relevant because frequency and effects of the polymorphism can be different between distinct populations.
7 In Brazil, an effect of this polymorphism on cognition in PD patients has been observed.
15 However, other analyses, such as genetic frequency comparisons with control individuals, or evaluations of other clinical features were not performed. The present study investigates this polymorphism in control subjects and PD patients in a Brazilian population. We hypothesized that BDNF polymorphism Val66Met is associated with PD and that this association is relevant to specific symptoms of the disease, namely anxiety and depression, given previous studies that associated BDNF polymorphisms with these disorders in non-PD individuals.
2,16,17Patients and Methods
PD patients (N=104) were recruited from Onofre Lopes University Hospital in Natal, Rio Grande do Norte (Brazil). Diagnosis of PD was based on the United Kingdom Parkinson’s Disease Society Brain Bank diagnostic criteria
18 and confirmed by a certified neurologist specialized in PD. Control subjects (N=96) with no kinship with patients and no history of neurological diseases were recruited in the hospital. The groups were matched by age and sex. This study was approved by the Ethics Research Committee of the Onofre Lopes University Hospital. All participants signed the written informed consent.
The BDNF Val66Met polymorphism was genotyped by polymerase chain reaction amplification and restriction enzyme digestion.
17 Genomic DNA was extracted from EDTA-containing peripheral blood samples (commercial kit FlexiGene DNA kit, QIAGEN, Germany). DNA integrity was assessed on a 1% ethidium bromide stained agarose gel, and quantitation was performed in SpectraMax 190 spectrophotometer ultraviolet-Vis Microplate Reader (Molecular Devices, United States). The following primers were used: BDNF Forward 5′-AAA GAA GCA AAC ATC CGA GGA CAA G-3′ and BDNF Reverse 5′-ATT CCT CCA GCA GAA AGA GAA GAG G-3′, which resulted in a 274 base pair (bp) PCR product. The thermocycler Esco’s Swift Maxi Thermal Cycle (United States) was used for DNA amplification. Amplification reactions were performed in a total volume of 25μL (5pmol of each primer, 200μM deoxyribonucleotide triphosphate (dNTP), 1×NHSO
4 buffer, 1.5mM MgCl
2, 0.75U Taq DNA polymerase; Fermentas-Thermo Fisher Scientific, United Kingdom) and 2 μl of pure DNA sample. The PCR cycling conditions were an initial denaturation for 3 minutes at 95°C, 40 cycles of 95°C for 30 seconds, 55°C for 30 seconds, and 72°C for 2 minutes, and a final extension at 72°C for 5 minutes. The BDNF V66M polymorphism was differentiated using the NlaIII (10 units/μl) (Uniscience-Bio Labs, New England, Ipswich, Mass.) restriction enzyme. Amplicons and restriction fragments were visualized after electrophoresis on a 3% ethidium bromide agarose gel (
Figure 1).
The following clinical evaluations were performed: age at PD onset, first symptom, side of symptom initiation, disease progression, Unified Parkinson´s Disease Rating Scale (UPDRS), Hoehn & Yahr (HY) scale, Schwab & England Activities Daily Living Scale (SE), Beck Depression Inventory (BDI), and Beck Anxiety Inventory (BAI).
The data were presented as mean±standard deviation [minimum value; maximum value]. Clinical outcomes between groups were compared with Mann-Whitney test given the nonparametric nature of the data. Chi-square test was used to compare the allele and genotype frequencies between control and PD patients. Adjusted residues with corrected Bonferroni p values were used as post hoc analysis.
19 The analyses of the Hardy-Weinberg Equilibrium (HWE) and polymorphism association with PD were performed in the SNPstats program.
20 Binary logistic regression model was used to estimate the effect of genotype in association with other factors on the probability of development of specific symptoms of PD. Power analysis was performed using the G*Power statistical program.
21 The analysis was considered significant when p<0.05.
Results
The 104 patients (73 men and 31 women) recruited for the PD group presented a mean age of 64.32±11.71 [37–89], mean age at onset of 55.70±11.97 [30–87], and 8.76±5.79 [1–32] years of progression of the disease. The 96 subjects (67 men and 29 women) of the control group presented a mean age of 63.15±10.18 [38–88]. No significant difference in age was observed between groups [U=4434.000, p=0.359]. Thirty-seven percent (37%) of PD patients had early Parkinson’s onset (EPDO; i.e., first symptoms before or at 50 years old), 44% reported positive family history, and 17.3% presented both conditions. Moreover, 70% described their first symptom as involuntary tremors, and 56% exhibited the disease as unilateral at first. Across the progression of the disease, all patients presented bilateral manifestations eventually.
Distribution of allele and genotype frequency was in agreement with HWE (p>0.05). Among PD patients, 75 (72.1%) had G/G genotype, while 26 (25%) had one copy and three (2.9%) had two copies of the BDNF variant allele A. The BDNF Val66Met genotype frequencies observed were similar to those previously seen in Brazilian PD patients.
16 In the control group, the frequency of homozygotes was 50 (52.1%) for the G allele, two (2.1%) for the A allele, and 44 (45.8%) were heterozygotes. The genotypes showed a statistically different frequency between the control group and PD patients (χ
2(2)=9.524, p=0.009), with the G/G (Bonferroni-adjusted residue, p<0.001) and A/G (Bonferroni-adjusted residue, p<0.001) genotypes being significantly more frequent in the PD and control groups, respectively (corrected significant, p=0.0083). Allele frequencies were not significantly different between the groups (χ
2(1)=3.712, p=0.054). Genotypes and allele frequencies are presented in
Table 1.
In order to perform binary logistic regression analysis, we pooled cases with A/G and A/A together due the small number of individuals with A/A genotype. The association analysis indicated that the A allele presented a protective factor against PD with an odds ratio (OR) equal to 0.42 (95% CI=0.22–0.72, p=0.003;
Table 2). Power analysis showed that this regression had 81.2% power to detect an association.
PD patients presented a significant reduction in functional independence (97.70±6.49 [60;100] versus 64.83±21.44 [20;90]) as measured in SE [U=296.000, p<0.001], higher UPDRS scores (4.58±5.84 [0;34] versus 45.11±19.82 [12;95]) [U=101.000, p<0.001], and more presence and severity of anxiety and depression symptoms as evaluated by BAI (9.35±8.63 [0;37] versus 16.52±9.51 [0;47]) and BDI (7.18±7.80 [0;47] versus 16.22±9.52 [2;47]) [U=2548.000, p<0.0001; 1857.500, p<0.0001], respectively, when compared with controls.
Table 3 shows the scores obtained in clinical symptoms evaluation in PD group according to genotype. There were no significant differences between the investigated genotypes. In addition, none of the clinical parameters evaluated were statistically different between carries and noncarriers of A allele in control group.
We also performed binary logistic regression analyses to verify if the BDNF genotype could exert an influence on development of depression and anxiety in PD. In association with other factors (
Table 4,
Table 5), PD group with G/G genotype presented a threefold risk (OR=3.4; IC 95%: 1.02–11.39; p=0.04) for depression and a fourfold risk for anxiety (OR=4.0; IC 95%: 1.12–14.94; p=0.03). Post hoc analysis showed a 99% power for both regressions, indicating that the size sample was adequate to analysis. This effect was only observed in the PD group. In the control group, the G/G genotype or the A allele did not influence the development of depression (G/G genotype: OR=1.139; IC 95%: 0.403–3.217; p=0.806; A allele: OR=0.878; IC 95%: 0.311–2.481; p=0.806) or anxiety (G/G genotype: OR=0.857; IC 95%: 0.304–2.416; p=0.771; A allele: OR=1.166; IC 95%: 0.414–3.286; p=0.771).
Discussion
We investigated the possible association of the BDNF Val66Met polymorphism with the development of PD in a Brazilian sample. Many studies have investigated this association, with contradictory overall results. One study observed a higher frequency of homozygosity (A/A) in PD patients in a Japanese population, but the control group was not in accordance with HWE.
11 Other studies found no association.
12,13 Furthermore, some meta-analyses performed on the subject showed no association,
22,23 and others indicated that the presence of A allele was a risk factor for PD in European populations, but with marginal confidence interval.
24Our data indicated that the A allele has a significant protective effect against PD. Although our results are contradictory with some of the previous research, other studies corroborate this finding. One study in East Asians observed that the A allele was significantly less frequent in PD patients.
14 Other research showed the AG genotype as a protective factor in Asians, but this association only reached significance among individuals with previous exposure to pesticides.
25 Furthermore, studies have shown that carriers of A allele exhibit a significantly later onset of PD.
26,27 Moreover, in two different studies in European and Asiatic populations, the G allele was considered a risk factor for Alzheimer´s disease, which shares pathological characteristics with PD.
28,29A possible mechanism for this protective role could be the fact that BDNF is synthesized as the precursor form proBDNF, which is cleaved to produce the mature form. Besides being an intermediary compound, proBDNF is a biologically active form that exerts the opposite function of the mature form. Specifically, proBDNF induces long-term depression and apoptosis through binding to p75 receptor, while the mature form promotes neuronal survival and long-term potentiation by binding to the TrkB receptor.
30 Chen and colleagues showed that activity-dependent secretion is the main form of BDNF release and that most of it is secreted as proBDNF, which is cleaved into mature BDNF.
5 Additionally, it was demonstrated that the amount of BDNF in this secretion pathway is diminished by the presence of at least one copy of the A allele, although total amount of BDNF in the brain is not altered.
2 Therefore, the reduction of proBDNF caused by the polymorphism may confer protection, particularly in a brain that is under the neurodegenerative process of PD.
31Truncated BDNF is another form of BDNF, which is obtained through proteolytic cleavage in a different site of the proBNDF. The function of this product is unknown.
30 Carlino et al. hypothesized that this isoform is an inactive form of proBDNF or that truncated BDNF inactivates proBDNF through the formation of inactive heterodimers.
30 Thus, another possibility is that extracellular processing of proBDNF in truncated or mature BDNF could be influenced by the presence of the G allele, and an alteration in the proportion of these isoforms could lead to higher susceptibility to PD.
30,31Most of genetic association studies on PD and the polymorphism BDNF Val66Met have been conducted in European and Asiatic populations, with only one study being conducted in South America.
32 It is known that the frequency of A allele varies greatly in the world population, being nearly absent in Sub-Saharan Africa and in some American indigenous populations and reaching 72% in She people, an ethnic group in China.
33 On the one hand, even though Asians show a higher percentage of the A allele, the effects associated with this polymorphism are more consistent among Caucasians.
34 This suggests that some populations could have developed a compensatory mechanism to the effects of the polymorphism. Our study found a protective association between the presence of BDNF Val66Met (allele A) polymorphism and PD. It is possible that such association is visible in the Brazilian population because it is one of the most heterogeneous populations in the world. This mixture is the result of five centuries of crosses between different ethnic groups, from European colonizers (mainly Portugal) to African slaves (west coast of Africa) and native indigenous tribes.
35 Such heterogeneity creates a particular genetic background in which the polymorphism could generate a linkage disequilibrium with another polymorphism in the same gene or with other genetic variation in a gene near BDNF, causing the association effect found in this population.
29 On the other hand, regarding race-related genetic factors, the ethnical heterogeneity of the Brazilian population is a limitation because of the lack of a coherent biological category. Nevertheless, apart from Brazil, there are a significant number of countries with a multiracial profile, which reinforces the need of genetic association studies in heterogenous populations.
Importantly, the present results also showed that the G/G genotype predisposes PD patients to more severe anxiety and depressive symptoms. According to the neurotrophic theory, impaired neurogenesis and neuronal plasticity are etiological factors for the development of psychiatric disorders induced by stress, such as depression and anxiety.
36,37 It has been shown that decreased BDNF secretion is related to major depressive disorder in humans and that antidepressant treatments normalize the levels of this neurotrophin.
37 Nonetheless, this theory has been reevaluated with the recent findings in animal and human studies. Govindarajan and colleagues observed that mice that overexpressed BDNF showed an increase in spine density in the basolateral amygdala, which is indicative of chronic stress and strongly correlated with anxiety. In contrast, BDNF overexpression also protected mice from stress-induced atrophy in hippocampal CA3 area, as well as decreased immobility in the forced swim test, suggesting that BDNF could exert different effects on different areas of the brain.
38 Recently, a paper demonstrated the relation between increased BDNF in the nucleus accumbens and reduced social interaction in mice submitted to chronic social defeat stress.
39 Moreover, Krishnan et al. showed that mice with G/G genotype for BDNF Val66Met polymorphism demonstrated depressive-like behavior, while A/A genotype mice exhibited resiliency, performing similar to control animals when submitted to the social defeat protocol.
40 Therefore, these findings support our results, in which PD patients with G/G genotype had higher risk for more severe depression and anxiety symptoms in association with other variables.
Our study has some limitations, such as the small number of participants in comparison with other polymorphism studies. However, the number of patients in this study is consistent with official diagnostic cases in the reference hospital for movement disorders where the data collection was performed. Furthermore, power analyses showed 81% and 99% power of the detection to the significant binary logistic regressions. Thus, the associations between allele A and a protective action against PD, as well as between G/G genotype and increased risk of anxiety and depression in PD patients, are reliable.
Another limitation, as mentioned above, is the fact that this polymorphism varies greatly between ethnic populations, so it would be interesting to have included this information in the analysis. Nonetheless, Brazil has a high rate of miscegenation, so that the phenotypic characterization of ethnicity would be subjective.
Finally, the PD group of the present study included a higher proportion of those with early onset and also with positive family history than values expected in a PD sample. These two observations could indicate that most of the patients are genetically predisposed to the disease. Nevertheless, the analyses did not show genetic or clinical differences among PD subgroups (early onset, family history), at least regarding the BDNF polymorphism studied here. In this respect, an interesting question would be if this pattern of earlier age onset and positive family history of PD is common in the Brazilian population. To our knowledge, however, there are no surveys that specify proportion of such PD subgroups in Brazil. This reinforces the need for more epidemiological and genetic association studies on PD in Brazil.