Surgical and Radiologic Anatomy

, Volume 28, Issue 2, pp 150–156

White matter damage of patients with Alzheimer’s disease correlated with the decreased cognitive function

Authors

  • Jin-Hai Duan
    • Department of Anatomy and Neurobiology, Preclinical Medicine SchoolSun Yat-sen University
  • Hua-Qiao Wang
    • Department of Anatomy and Neurobiology, Preclinical Medicine SchoolSun Yat-sen University
  • Jie Xu
    • Department of Anatomy and Neurobiology, Preclinical Medicine SchoolSun Yat-sen University
  • Xian Lin
    • Department of Anatomy and Neurobiology, Preclinical Medicine SchoolSun Yat-sen University
  • Shao-Qiong Chen
    • Department of Radiology, Third Affiliated HospitalSun Yat-sen University
  • Zhuang Kang
    • Department of Radiology, Third Affiliated HospitalSun Yat-sen University
    • Department of Anatomy and Neurobiology, Preclinical Medicine SchoolSun Yat-sen University
Original Article

DOI: 10.1007/s00276-006-0111-2

Cite this article as:
Duan, J., Wang, H., Xu, J. et al. Surg Radiol Anat (2006) 28: 150. doi:10.1007/s00276-006-0111-2

Abstract

Increasing evidence demonstrates that there is marked damage and dysfunction in the white matter in Alzheimer’s disease (AD). The present study investigates the nature of white matter damage of patients with Alzheimer’s disease with diffusion tensor magnetic resonance imaging (DTI) and analyses the relationship between the white matter damage and the cognition function. DTI, as well as T1 fluid attenuated inversion recovery (FLAIR) and T2-FLAIR, was performed on probable patients of Alzheimer’s disease, and sex and age matched healthy volunteers to measure the fractional anisotropy (FA) and mean diffusivity (MD) in the genu and splenium of the corpus callosum, anterior and posterior limbs of the internal capsule, and the white matter of frontal, temporal, parietal, and occipital lobes. FA was lower in the splenium of corpus callosum, as well as in the white matter of the frontal, temporal, and parietal lobes from patients with Alzheimer’s disease than in the corresponding region from healthy controls and was strongly positive correlated with MMSE scores, whereas FA appeared no different in the anterior and posterior limbs of internal capsule, occipital lobes white matter, and the genu of corpus callosum between the patients and healthy controls. MD was significantly higher in the splenium of corpus callosum and parietal lobes white matter from patients than in that those from healthy controls and was strongly negative correlated with MMSE scores, whereas MD in the anterior and posterior limbs of internal capsule, as well as in frontal, temporal, occipital lobes white matter and the genu of corpus callosum, was not different between the patients and healthy controls. The most prominent alteration of FA and MD was in the splenium of corpus callosum. Our results suggested that white matter of patients with Alzheimer’s disease was selectively impaired and the extent of damage had a strong correlation with the cognitive function, and that selective impairment reflected the cortico–cortical and cortico–subcortical disconnections in the pathomechanism of Alzheimer’s disease. The values of FA and MD in white matter, especially in the splenium of corpus callosum in AD patients, might be a more appropriate surrogate marker for monitoring the disease progression.

Keywords

Alzheimer’s diseaseWhite matter damageDiffusion tensor imagingFractional anisotropyMean diffusivity

Introduction

Alzheimer’s disease (AD) is characterized pathologically by amyloid plaques (SPs), neurofibrillary tangles (NFTs), neuronal loss, degeneration, and atrophy of cerebral cortex. These pathological changes originate in the medial temporal region, entorhinal cortex and hippocampus, subsequently spread over the entire limbic cortex, and then into the neocortical association cortex [4], which result in progressive memory and cognitive decline accompanied by daily living and behavioral disturbance. Although they are generally considered to affect gray matter, histological study shows the pathological changes, such as loss of axons, oligodendrocytes, and lipid components in the white matter, together with reactive astrocytosis [5, 12, 13]. Moreover, accumulating MRI studies have shown in vivo that there is a considerable amount of microscopic white matter impairment [6, 7, 14, 28, 33]. However, whether the white matter damage is correlated with cognitive function is still controversial [6, 33].

Conventional MR imaging can show overall brain atrophy or atrophy of specific brain structures in patients with AD, but has a limited capability to quantify changes in the subcortical white matter. Diffusion tensor magnetic resonance imaging (DTI) is a non-invasive technique and can be used for quantitatively in vivo measuring the degree and directionality of the displacement distribution of water molecules to provide information about the size, shape, orientation, and geometry of brain structures [19], as well as to demonstrate the white matter abnormalities, including subtle changes not visualized on conventional MRI [2, 21]. Restricted by many factors including cell membranes, axonal membranes, cytoskeletal structures, such as neurofilaments and microtubules [26], the diffusion of water molecules in biological tissues may not be same in all direction. The preference of diffusion in fiber direction is called anisotropy. Fractional anisotropy (FA), one of the most robust measures of anisotropy, is useful for mapping the functional integrity and specific organization of white matter fibers. Regional differences in fiber packing density, degree of myelination, fiber diameter, and packing density of neuroglial cells can contribute to the variation of FA. Mean diffusivity (MD) is a measure for randomized mean water diffusion. It can be used as a measure of alterations of brain tissue [20].

In healthy subjects, increased age is associated with changes in underlying tissue composition, for example, water, protein, and mineral content of tissue [9]. Aging can induce a decline of FA and an increase of MD in the white matter [25, 32] and exhibits a roughly anterior-to-posterior gradient effect on anisotropy and diffusivity [16]. But in AD patients, white matter microstructure changes are more severe as demonstrated in previous DTI studies [6, 28, 33]. The deposition of plaques and tangles in the brain of AD patients can influence the signal properties of affected tissue [3] and cause the changes of DTI parameters [30]. In detail, a decline in FA and an increase in MD in the corpus callosum, as well as in the white matter of frontal, temporal, and parietal lobes in AD patients has been demonstrated, and strong correlations were shown between the MMSE score and the average overall white matter MD and FA [6]. On the other hand, Takahashi et al. [33] found that FA significantly reduced in the temporal white matter, the corpus callosum, and in the anterior and posterior cingulum of AD patients, and that the reduction of FA was not correlated with MMSE scores in each studied area. These discrepancies of relationship between white matter damage and cognitive function encourage us to make further investigation. In the present study, we measured the FA and MD in the white matter region with DTI technique to identify region-by-region the relationship between the white matter damage and cognitive function decline.

Materials and methods

Subjects

From April 2004 to July 2005, 16 mild and moderate probable AD patients (6 men and 10 women; mean age 65.22±8.58 years, range 60–80 years; disease duration 2–5 years; education 7.19±3.39 years) who met the National Institute of Neurological and Communicative Disease and Stroke and the Alzheimer’s disease (NINCDS-ADRDA) criteria [22] for a diagnosis of clinically probable AD were recruited from patients hospitalized in the Nanhai welfare center, Guangdong province,China. Twelve sex and age matched healthy controls (5 men and 7 women; mean age 64.35±9.15 years, range 60–80 years; education 8.25±2.90 years) without complaints of cognitive and memory deficits were recruited. The clinical test procedures included examination of neurologists and psychiatrists, neuropsychological examination, brain MRI scans, chest X-ray film, electrocardiography, complete blood count, sedimentation rate, serum electrolytes, glucose, urea nitrogen, creatinin, liver-associated enzymes, cholesterol, high-density-lipoprotein cholesterol, triglycerid, vitamin B12, folate, and urinalysis. Major systemic, psychiatric, other neurological illness, hypertension, mellitus diabetics, tumor, drug or alcohol abuse, auto-immunologic disease were carefully excluded for both groups. All patients received no special treatment for dementia and had scores of less than 4 on the Hachinski Ischemic scale [11]. Mean Mini-Mental State Examination (MMSE) score (corrected for age and level of education) was 13 (range, 8–22) for patients with AD and 28 (range, 27–30) for control subjects. Subjects were excluded from study if they had either one hyperintense area with a diameter equal or greater than 5 mm or more than four hyperintense areas smaller than 5 mm in diameter on T2-FLAIR MRI [6]. Local ethics committee approval and written informed consent from all subjects or their guardians were obtained before study.

MRI acquisition and postprocessing

All patients were imaged using a 1.5T clinical MR scanner (GE SIGNA EXCITE, US) with a standard head coil. Prior to DT imaging, all patients underwent T1-FLAIR (TR 2,000 ms, TE 10.9 ms, TI 750 ms, FOV 240 mm×240 mm , Matrix 320×256, 5 mm section thickness and 1 mm section separation) and T2-FLAIR(TR 8,800 ms, TE 120 ms, TI 2,200 ms, Matrix 320×192, FOV 240 mm×240 mm, 5 mm section thickness and 1 mm section separation) imaging to obtain axial and coronal images. A spin-echo type echo-planar imaging sequence with diffusion gradients was used for DTI sequence to obtain axial and coronal images with 25 directional diffusion-encodings (the b value was 1,000 s/mm2 for each direction) as well as with no diffusion encoding (b value=0 s/mm2), TR 8,000 ms, TE 79.3 ms, Matrix 128×128, FOV 240 mm×240 mm, axial slices number 28, coronal slices number 28, 5 mm thickness , no section gap. The axial slices were positioned to run parallel to a line that joins the most inferoanterior and inferoposterior parts of the corpus callosum and the coronal slices were acquired with orientation perpendicular to the axial slices. This set of slices allowed us to cover a relatively large portion of the cerebral hemispheres where white matter was highly represented.

All images were post-processed on a workstation connected to the scanner, using the Functool 2 image analysis software (General Electric Medical Systems, Buc, France). As previously described [10], FA and MD were measured to evaluate the anisotropy and diffusion degree of water molecules in the white matter. The regions of interest (ROIs, round areas of 20–40 mm2) were placed bilaterally in the white matter of the following areas: the genu and splenium of corpus callosum, the posterior and anterior limb of internal capsule, and the white matter of four lobes. The genu and splenium of corpus callosum were sampled at the slices of the optic chiasm and anterior part of the inferior colliculus, respectively. The anterior and posterior limbs of the internal capsule, respectively, indexed as the regions bounded by the corner between the head of the caudate nucleus and the pallidum, and by the pallidum and the thalamus, were selected on three continuous slices. Frontal lobe white matter was placed on the three continuous slices of middle frontal gyrus. Parietal lobe white matter was sampled on the three continuous slices of the angular gyrus. Temporal lobe ROIs were positioned on the white matter medial to the temporal lobe and marked on three continuous slices [15]. Occipital lobe ROIs were placed on the three continuous slices of the occipital apex. Care was taken to avoid the partial volume averaging from the CSF and gray matter. The average FA an MD were calculated for each of the selected white matter areas. Figure 1 illustrates the location of all the white mater ROIs.
https://static-content.springer.com/image/art%3A10.1007%2Fs00276-006-0111-2/MediaObjects/276_2006_111_Fig1_HTML.jpg
Fig. 1

Location of the white matter regions of interest: 1 whiter matter of parietal lobe; 2 white matter of frontal lobe; 3 anterior limb of internal capsule; 4 posterior limb of internal capsule; 5 white matter of occipital lobe; 6 genu of corpus callosum; 7 white matter of temporal lobe; 8 splenium of corpus callosum

Statistical analysis

An independent-samples t test for paired samples was used to compare FA and MD values of the white matter from AD patients and healthy controls. The correlations between average MD and FA in each region studied and MMSE scores were investigated by Pearson’s correlation coefficient. The alpha level for statistical significance was established at 0.05. Statistics were processed with SPSS software package version 11.5.

Results

Compared with the healthy controls, patients with AD demonstrated a significant lower FA in the splenium of corpus callosum, as well as in the white matter of temporal, parietal, and frontal lobes, but presented no differences in occipital lobe, the anterior and posterior limbs of internal capsule, and the genu of corpus callosum. (Table 1, Fig. 2).
Table 1

Fractional anisotropy values of the selected white matter areas from patients and controls \( (\ifmmode\expandafter\bar\else\expandafter\=\fi{x} \pm s)\)

Regions

AD

Controls

P value

Genu of corpus callosum

0.585±0.026

0.610±0.046

0.201

Splenium of corpus callosum

0.602±0.043

0.733±0.042

<0.001

Anterior limb of internal capsule

0.525±0.079

0.514±0.027

0.895

Posterior limb of internal capsule

0.629±0.036

0.614±0.027

0.827

Frontal lobe

0.270±0.034

0.305±0.037

0.005

Temporal lobe

0.302±0.032

0.340±0.027

0.012

Parietal lobe

0.294±0.043

0.341±0.020

0.001

Occipital lobe

0.348±0.032

0.349±0.012

0.385

https://static-content.springer.com/image/art%3A10.1007%2Fs00276-006-0111-2/MediaObjects/276_2006_111_Fig2_HTML.jpg
Fig. 2

Compared with healthy controls (a), patients with Alzheimer’s disease (b) showed a significantly decreased anisotropy in the splenium of corpus callosum (arrows)

In contrast to controls, MD values in patients with AD were significantly higher in the splenium of corpus callosum and parietal lobe, but appeared as no difference in other regions (Table 2).
Table 2

Regional MD values in patients with AD and controls \( (\ifmmode\expandafter\bar\else\expandafter\=\fi{x} \pm s,\,\hbox{m}^{2}/\hbox{s} \times 10^{{ - 9}} )\)

Regions

AD

Controls

P value

Genu of corpus callosum

1.085±0.073

1.053±0.071

0.245

Splenium of corpus callosum

0.918±0.029

0.727±0.035

<0.001

Anterior limb of internal capsule

0.839±0.026

0.815±0.029

0.631

Posterior limb of internal capsule

0.859±0.044

0.801±0.034

0.781

Frontal lobe

0.813±0.028

0.785±0.075

0.188

Temporal lobe

0.917±0.018

0.829±0.056

0.259

Parietal lobe

0.826±0.015

0.726±0.072

0.004

Occipital lobe

0.825±0.048

0.818±0.015

0.607

Analysis on correlation between values of FA and MD and MMSE scores in various parts of white matter indicated that FA in the splenium of corpus callosum, white matter of frontal, temporal, and parietal lobes presented a linear positive correlation with MMSE scores, while FA in the anterior and posterior limbs of internal capsule, occipital lobes, and the genu of corpus callosum was not correlated with MMSE scores. MD in the splenium of corpus callosum and parietal lobes was negatively correlated with MMSE scores (Table 3).
Table 3

The correlations between FA and MD values in specific white matter region and MMSE scores

Regions

FA

MD

r

P

r

P

Genu of corpus callosum

0.163

0.547

−0.165

0.542

Splenium of corpus callosum

0.809

0.001

−0.867

<0.001

Anterior limb of internal capsule

−0.411

0.114

−0.228

0.395

Posterior limb of internal capsule

0.156

0.565

0.115

0.672

Frontal lobe

0.721

0.002

0.254

0.343

Temporal lobe

0.711

0.002

−0.060

0.825

Parietal lobe

0.700

0.003

−0.808

<0.001

Occipital lobe

−0.413

0.112

−0.185

0.493

Discussion

The present study demonstrated that FA in AD patients was significantly lower in the splenium of corpus callosum, as well as in the white matter of temporal, parietal and frontal lobes, but it presented no much difference in occipital lobe, the anterior and posterior limbs of internal capsule, and the genu of corpus callosum. Meanwhile, MD was remarkably increased in the splenium of corpus callosum and parietal lobe. The decrement of white matter anisotropy suggested that white matter of AD patients were selectively impaired and the integrity of white matter tracts in these regions were damaged. The selective impairment of white matter was probably concerned with the pathologically proved distribution of NFTs and SPs in the cortex and interconnection among white matter fibers. These pathological hallmarks of AD predominantly affect the posterior brain and the temporal and parietal association cortices [1]. NFTs are mainly in the supra- and infragranular layers, particularly in the association cortex layers III and V [24] and primarily impair the large glutamatergic projection neurons [17]. SPs occur predominantly in the layers II and III in the association areas of neocortex and affect those laminae characterized by large pyramidal cells subserving input/output functions [27]. SPs are also significantly numerous in cortical laminae dominated by their role in corticocortical associative relations [27]. It has been proved that deposition of amyloid β can lead to white matter injury that can be detected by DTI [30]. Moreover, Our results showed that the most remarkable reduction of FA and elevation of MD was in the splenium of corpus callosum (Fig. 2), but not in the genu which suggested an anterio-posterior gradient in anisotropy alteration in AD patients. These findings were also supported by other research [16]. One possible explanation for this is maybe that the fibers in the splenium of corpus callosum originate from the temporoparietal cortex which are characteristically affected in AD patients [5, 8, 18]. The high prevalence in microstructural damage of fiber tracts connecting cortical associative areas reflected the cortico–cortical and cortico–subcortical disconnections and suggested that wallerian degeneration of white matter fiber tracts was secondary to neuronal loss in the associative cortex.

Increasing evidence found that normal aging was associated with demyelination and changes in water, protein, and mineral content of white matter [9] and potentially lead to the alteration of anisotropy and diffusivity in the white matter region, but aging mainly resulted in the significant changes in white matter in anterior region, while the dementia status of AD patients was associated with alteration of anisotropy and diffusivity in posterior fiber tracts [16, 29]. These findings were in agreement with our data. Furthermore, vascular damage can have the potential to lead to the changes of FA and MD [23, 31]. But considering the criteria used to select patients and no or mild changes on T1-FLAIR and T2-FLAIR images, their role is likely to be minor. The present findings may likely represent white matter damage caused purely by AD pathology. The selective FA reduction and MD elevation indicated the loss of barriers restricting water molecular diffusion and tissue anisotropy of white matter in AD patients, which supports the histopathological proof showing loss of myelin, axons, and oligodendrial cells in the white matter [5].

The correlation between changes of FA and MD in some specific white matter region and MMSE scores suggested that white matter damage in AD patients was related to decline of cognitive function. However, the relationship between white matter impairment and cognitive function was strongly augmented. Takahashi et al. [33] did not find that there was a correlation between the FA and MMSE scores. Two possible reasons might lead to the discrepancy between findings in this and our study. First, we selected more ROIs in each region to detect FA and MD compared to a previous study. Second, patients and control subjects in our study were younger (mean age: 65 and 64 years) than those in earlier investigation (mean age: 70 years in both groups). Thus, aging effect on anisotropy and diffusivity was more prominent in former research. Consequently, present study could more precisely indicate the alteration of anisotropy and diffusivity and reflect the damage information in white matter tracts. Our findings were also consistent with prior investigations demonstrating a significant correlation between the lattice index of corpus callosum splenium and MMSE scores [28].

Although the sample size of this study was small, the correlation studied encouraged the recognition of importance of white matter damage in the occurrence and development of AD. DTI technique provided a more direct assessment of the integrity of white matter fibers and was a very powerful tool used to explore the pathological mechanism of AD. In contrast to the conventional MRI and neurophysiological assessment, the detection of FA and MD in white matter, especially in the splenium of corpus collasum in AD patients might be a more appropriate surrogate marker for monitoring the disease progression.

Acknowledgment

This study was supported by the National Natural Science Foundation of China (no. 3040052), the Natural Science of Foundation of Guangdong Province, China (no.20013137), and the Social Development Projects of Guangdong Province (no. 2005B10401047). We thank Ms. Mo Si-jie and Dr. Hu Tao for recruiting the subjects from Nanhai welfare centre, Guangdong Province, China. We also gratefully acknowledged Dr. Shan Hong of Department of Radiology, Third Affiliated Hospital, Sun Yat-sen University and Dr. Zhang Zhong-Wei and Dr. Wang Jian-Bo of Department of Radiology, First Affiliated Hospital, Sun Yat-sen University for their excellent technical assistance in this project. The present experiment complies with the relative law of the People’s Republic of China.

Copyright information

© Springer-Verlag 2006