Allele and Genotype Distribution

Thirteen individuals did not participate in the MRI scanning because of quitting, and the MRI data of 7 participants was discarded because of unsuccessful DNA sequencing or MRI scanning. Therefore, 61 subjects in total were finally grouped according to their BDNF genotype, into 15 homozygous carriers of the Val allele (Val/Val), 17 homozygous carriers of the Met allele (Met/Met), and 29 heterozygous (Val/Met). The frequency of the BDNF genotypes was 24.6% for Val/Val, 47.5% for Val/Met, and 27.9% for Met/Met in our study of sample (Table 1), which was similar to previously published results based on the Chinese population [14], [27], [28]. The distribution of genotypes in healthy subjects was consistent with the Hardy–Weinberg Equilibrium (χ² = 0.141, P = 0.707). There were no significant differences between males and females with respect to genotype (χ² = 4.077, P = 0.13) and allele distribution (χ² = 3.175, P = 0.075).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Genotype and allele frequencies of the BDNF Val66Met polymorphism in healthy young Chinese adults.

https://doi.org/10.1371/journal.pone.0037777.t001

Cortical Thickness Differences between Val/Met, Met/Met and Val/Val Genotypes

In general, compared to homozygous Val carriers there were no brain regions showing thickening in homozygous Met or heterozygous Met carriers. Compared with the Val/Val group, both the Met/Met and Val/Met groups exhibited a significantly thinner cortex in the frontal (superior, middle and orbital parts) cortex, middle temporal cortex (MTC), superior parietal cortex (SPC), lateral occipital cortex (LOC), insular cortex (IC) and precuneus cortex (PC) in one or both hemispheres (FDR corrected, P<0.05; Figure 1, Table 2 and 3). Moreover, the Met/Met group was observed to have reduced thickness in other areas, including the left inferior parietal cortex (IPC), and the posterior cingulate cortex (PCC) (FDR corrected, P<0.05; Figure 1 and Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Cortical thickness differences among three genotypes based on the BDNF Val66Met polymorphism.

Left panel: areas with a significantly thinning cortex in the Met/Met group compared with the Val/Val group (Met/Met < Val/Val, corrected, P<0.05). Right panel: areas with a significantly thinning cortex in the Val/Met group compared with the Val/Val group (Val/Met < Val/Val, corrected, P<0.05). cMFC, caudal middle frontal cortex; IC, insular cortex; IPC, inferior parietal cortex; LOC, lateral occipital cortex; MTC, middle temporal cortex; OFC, orbital frontal cortex; PC, precuneus cortex; PCC, posterior cingulate cortex; rMFC, rostral middle frontal cortex; SFC, superior frontal cortex; SPC, superior parietal cortex;. See Table 2 and Table 3 for details.

https://doi.org/10.1371/journal.pone.0037777.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Areas with decreased cortical thickness in the Met/Met group compared with the Val/Val group based on the BDNF Val66Met polymorphism (FDR corrected, P<0.05).

https://doi.org/10.1371/journal.pone.0037777.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Areas with decreased cortical thickness in the Val/Met group compared with the Val/Val group based on the BDNF Val66Met polymorphism (FDR corrected, P<0.05).

https://doi.org/10.1371/journal.pone.0037777.t003

VBM Differences between Val/Met, Met/Met and Val/Val Genotypes

Although numerous studies have investigated the relevance between the BDNF Val66Met polymorphism and GM volume, no study has focused on the Chinese population. In this study we observed GM volume differences among BDNF (rs6265) genotypes in young Chinese adults.

No brain regions showed increased GM volume in the Met variant, and the total GM volume over the entire brain was not different among groups. Compared to the Val/Val group, Met/Met individuals had a smaller volume in some areas of the frontal, temporal, cingulate and insular cortices in one or both hemispheres (Figure 2, TFCE corrected, P<0.05). For Val/Met subjects, the left middle cingulate cortex (MCC) and IC showed smaller volumes relative to people who were homozygous for the Val allele (TFCE corrected, P<0.05; Figure 2).

When grouping Met/Met and Val/Met genotypes as a group of Met carriers, as performed in many prior studies, the Met carriers possessed a smaller GM volume only in the bilateral MCC (TFCE corrected, P<0.05; Figure 2).

Brain Structural Difference in Met/Met Genotype Subjects

Following comparison of Met/Met and Val/Val genotypes, consistent differences in cortical thickness and GM volume were found in the present study, particularly in the OFC, MTC, PCC, and IC which showed decreased cortical thickness and GM volume in the Met/Met group. These matching results suggested that these frontal, temporal, cingulate, and insular regions were probably the key areas in which brain structure was determined by BDNF 66Met to some extent in Chinese people.

To date, there has been no study reporting brain structural differences in the BDNF (rs6265) Met/Met genotype alone. The current study was the first to report the effect of the Met/Met genotype on brain morphology using cortical thickness analysis and VBM. Some of our VBM results were consistent with previous studies which were based on Met carriers (combining Val/Met and Met/Met) who had reduced GM volume in the frontal [7], [9], [33] and temporal cortices [11]. Although the Met carriers in these studies were mainly of the Val/Met genotype, together with our VBM and cortical thickness analysis, we believe that the Met allele of BDNF (rs6265) most likely influences brain morphology and that the Met/Met genotype causes the most significant effects, at least in young Chinese adults.

The cortical thickness and VBM analysis also demonstrated GM differences in several other regions of Met/Met subjects, including the superior frontal cortex (SFC), middle frontal cortex (MFC), IPC, SPC, PC, and LOC with reduced cortical thickness, and the IFC, ACC, MCC and ITC with decreased GM volume. The differences in cortical thickness and GM volume in most of these regions were consistent with previous VBM studies [7], [9], [11], [33]excepting the MCC which has never been reported until now. Anatomical studies have showen that the IC has extensive connections with the frontal, parietal, temporal and cingulate cortices [34], [35], [36], [37], which showed reduced GM in our present study no matter the cortical thickness or GM volume. Therefore, this interactive fibers relationship may be directly or indirectly contributed to the GM differences in the MCC, PCC, parietal lobe and IC in Met homozygotes of our present study. However, because VBM and cortical thickness results in some of these regions were not consistent in the present study, the correlation of the BDNF (rs6265) Met/Met genotype and brain anatomical differences in these areas requires further clarification.

Brain Structural Differences in Val/Met Genotype Subjects

Regarding the comparison of the Val/Met and Val/Val genotypes, matching results were found in the IC by dual-analysis, which were also shown in the comparision between Met/Met and Val/Val groups. These results suggested that this area was most likely influenced by BDNF (rs6265) polymorphism. In addition, decreased GM volume was also found in the MCC in the Val/Met group, and was also observed in the Met/Met group. In addition, reduced cortical thickness in the Val/Met group was found in other areas including the OFC, MFC, SFC, MTC, SPC, PC and LOC which were also included within the cortical thickness decreased regions of the Met/Met group when compared to the Val/Val group. Thus, it can be seen that regions showing differences in Met-heterozygotes overlapped with those of Met-homozygotes. It suggested that substitution of the Met allele for the Val allele in the two homologous chromosomes of the BDNF DNA sequence would result in more severe damage in brain GM. This was demonstracted indirectly by an animal study in which the Met/Met genotype could cause a most significant decrease in the regulated secretion of BDNF by 30% (Val/Met by 18%) [5], where secreted BDNF is known to regulate neuronal development and differentiation. Therefore, extensive decreased GM in Met homozygotes may account for by serious altered neuronal morphology in the present study.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. VBM differences among three genotypes based on the BDNF Val66Met polymorphism.

Top panel: areas with a significantly decreased GM volume in Met/Met group compared with Val/Val group (Met/Met < Val/Val, corrected, P<0.05). Bottom left panel: areas with a significantly decreased GM volume in the Val/Met group compared with the Val/Val group (Val/Met < Val/Val, corrected, P<0.05). Bottom left panel: areas with a significantly decreased GM volume in the Met carriers group (Met/Met+Val/Met) compared with the Val/Val group (Met carriers < Val/Val, corrected, P<0.05 ). ACC, anterior cingulate cortex; MCC, middle cingulated cortex; MTC, middle temporal cortex; IC, insular cortex; IFC, inferior frontal cortex; OFC, orbital frontal cortex; PCC, posterior cingulate cortex; STC, superior temporal cortex.

https://doi.org/10.1371/journal.pone.0037777.g002

The findings of some previous studies reporting smaller hippocampal volume in Met allele carriers were not replicated in either of our Val/Met or Met/Met subjects. However, our negative results were consistent with several recent reports, in which no evidence for an influence of the BDNF Val66Met genotype on MRI-derived reduced hippocampus volume was found [8], [38], [39], [40]. One reason for these conflicting results may be due to the fact that our study focused on the Chinese population, unlike previous positive studies [6], [7]; however, the genetic effect of BDNF Val66Met polymorphism on hippocampal morphology remains controversial and requires further investigation.

Another point of note is the influence of the grouping method on our results. When we merged the Met/Met and Val/Met groups to obtain the Met carriers group in the VBM analysis, as performed in previous studies [6], [8], [11], [31], the reduction in GM values in brain areas was much smaller in the Met carriers than in the Met/Met group. These results suggested that grouping method of Met carriers may hide some specific roles of Met/Met in brain morphology. To date, we do not know if such a marked discrepancy caused by grouping method also exists in similar analyses based on the Caucasian race because of the low frequency of Met/Met in Caucasians. Nevertheless, our present study has revealed that the function of the Met/Met genotype and Val/Met genotype cannot be represented entirely by a general group of Met carriers. The current findings indicated that availability of a large sample size that can be divided into three genotypes is helpful in clarifying the effect of BDNF Val66Met polymorphism, which should be given more attention in future studies.

Study Limitations

There are some limitations to our work. The first one is that we didn’t measure the behavioral differences in cognition, emotion, memory and other factors among BDNF Val66Met genotypes. These behaviors related to higher brain function have often been shown to be affected by BDNF Val66Met polymorphism [4], [8], [41]. Thus, if the genetic impact on brain structure in the current study could correlate to behaviors that were close to BDNF function, this would contribute greatly to inferences regarding structural brain differences determined by BDNF Val66Met polymorphism. The second limitation of this study is the small sample size. There were only 15 and 17 individuals in the Met/Met and Val/Val groups respecitvely, which is relatively small for an objective result. Larger sample sizes are required for futher related studies. The third limitation was this study was that comparisons of brain morphology were only carried out at one time-points. To obtain more data, brain morphology data should be collected at two different time –points. These limitations need to be addressed in future related studies.

Conclusion

In conclusion, the current study demonstrates brain morphology differences related to BNDF Val66Met polymorphism (especially the Met/Met genotype) for the first time in healthy young Chinese adults using a dual analytical method, via cortical thickness analysis and VBM. Unlike Caucasians with significant brain structural differences in Val/Met genotype individuals, Chinese brain anatomy varied mostly in the Met/Met genotype with less GM mainly in the frontal, temporal, cingulate and insular cortices. In addition, by way of analyzing Met carriers, the present study suggested that combining Val/Met and Met/Met genotypes as Met carriers would probably mask the unique role of the Met/Met genotype on brain morphology, which should be taken into account in future studies.

Subjects

The data included in this study were drawn from 81 healthy, right-handed Chinese college students (mean age: 20.5±0.88 years, range: 19–23; 48 males). For all subjects, exclusion criteria were: 1) reported history of brain injury; 2) macroscopic brain T2-visible lesions, tumor or obvious cortical/subcortical atrophy on MRI scans; 3) psychiatric disorders evaluated by the Structured Clinical Interviews for DSM-IV Axis I (SCID-I) and Axis II (SCID-II); 4) existence of a neurological disease diagnosed by clinical neurological examination; 5) pregnancy or menstrual period; 6) use of prescription medications within the last month; 7) alcohol, nicotine or drug abuse; 8) dementia. Therer were no differences in IQ among genotype groups according to the Chinese Revised Wechsler Adult Intelligence Scale (WAIS-RC).

Genotyping

Genomic DNA was isolated from 500 µl of whole blood using TIANamp genomic DNA kit according to the manufacturer’s instructions (TIANGEN BIOTECH CO., LTD., Beijing, China). Polymerase chain reaction (PCR) was carried out in a volume of 20 µLcontaining 10 ng of genomic DNA, 2.5 µL of 10×Taq buffer, 2 µL of dNTP Mixture (2.5 mM), 0.5 µL of Taq DNA Polymerase (2.5 U/µL, TIANGEN BIOTECH CO., LTD.,), 1 µL of each primer and ddH₂O. The PCRs were done using a BIO-RAD cycler system. The primers were as follows: 5′-AAACATCCGAGGACAAGGTG-3′ (forward), 5′-AGAAGAGGAGGCTCCA AA GG-3′ (reverse). The amplification conditions were initiated at 94°C for 5 min, followed by 30 cycles consisting of denaturation at 94°C for 30 s, annealing at 65°C for 30 s and extension at 72°C for 30 s, with a final extension step of 5 min at 72°C. The PCR products were confirmed by 2% (w/v) gel electrophoresis and purified using the TIANgel Midi purification kit (TIANGEN BIOTECH CO., LTD.). BDNF gene sequence was implemented in Shanghai Bio-tech Co., Ltd. (Shanghai, China). Genotypes of the BDNF Val66Met polymorphism were read by at least two researchers who were blinded to the experiment. For genotypic distribution analysis, the Hardy–Weinberg Equilibrium was tested by the goodness-of-fit χ² test.

Structural MRI Acquisition

The MRI experiment was performed using a 3.0 Tesla MRI scanner (Magnetom Trio, Siemens) with a standard head coil. A set of T1-weighted high-resolution structural images was collected (TR/TE: 2.7 s/3.39 ms, FOV: 256 mm×256 mm, matrix size: 256×256, flip angle: 12°, in-plane resolution: 1 mm×1 mm, slice thickness: 1 mm with no gaps).

Cortical Thickness Processing

FreeSurfer 5.0 (http://surfer.nmr.mgh.harvard.edu/) was employed to calculate the cortical thickness from the structural MRI according to the detailed procedures in previous studies [42], [43], [44]. Local cortical thickness was measured on the basis of the difference between the position of equivalent vertices in the pial and gray-white matter surfaces. In brief, cerebral white matter was segmented from the T1-weighted images and the gray-white matter interface was estimated [43]. Topographical defects in the gray-white estimate were fixed, which were then used as the starting point for the deformable surface algorithm search for the pial surface. The surface of the gray-white matter border was inflated, and differences between subjects in the depth of the gyri and sulci were normalized. The reconstructed brain of each subject was deformed and registered to an average spherical surface [44]. To obtain cortical thickness difference maps, the data were smoothed on the surface with a Gaussian smoothing kernel with a full-width half maximum of 10 mm. Statistical thickness difference maps were constructed with t statistics (with the inclusion of age and gender as covariates) with a threshold of P<0.05 ((false discovery rate (FDR) corrected)).

VBM Processsing

Structural data was analysed with FSL-VBM, a voxel-based morphometry style analysis [45], [46] carried out with FSL tools [46]. Structural data were analyzed using FSL-VBM, a voxel-based morphometry style anslysis [45], [46], which was carried out with FSL tools [47]. The structural images were first brain-extracted using BET [48]. Then, the resulting gray matter partial volume images were aligned to MNI152 standard space using the nonlinear registration tool FNIRT, which uses a b-spline representation of the registration warp field. The registered images were averaged to create a study-specific template, to which the native gray matter images were then non-linearly re-registered. The registered images were then modulated (to correct for local expansion or contraction) by dividing by the Jacobian of the warp field. The modulated segmented images were then smoothed with an isotropic Gaussian kernel with a sigma of 3 mm. Finally, group-difference in gray matter volume was tested using permutation-based non-parametric testing (10,000 permutations) with gender and age included as covariates. Multiple comparisons across space were corrected using the threshold-free cluster enhancement method [49]. A difference was considered to be significant at a corrected P<0.05. The minimal number voxels was 8.