Elevated neurofilament light chain in plasma is associated with reduced right hippocampal and amygdala volume in Alzheimer's patients

Background Plasma neurofilament light (NfL) levels have been considered as an especially promising biomarker for dementia, however, the mechanism of NfL regulating cognition is not very clear. Methods 43 amnesic mild cognitive impairment(aMCI), 35 Alzheimer’s disease (AD) and 30 cognitively normal subjects were recruited. Plasma NfL levels were examined by the Single Molecule array (Simoa) technique; the volumes of the hippocampus and amygdala were calculated and compared by T1-weighted MRI; and cognitive function was assessed by the Beijing version of the Montreal Cognitive Assessment (MoCA) Results Our results showed significantly increased plasma NfL levels in AD group (29.42 pg/ml) compared to aMCI(15.92 pg/ml) group and normal (12.85 pg/ml) group (both p < 0.001), while there was no statistical difference (p>0.05) between aMCI group and normal group. And the results of partial correlation analysis showed that plasma NfL levels were negatively correlated (p<0.05)with MoCA total score (r=-0.415, p=0.013), right hippocampal volume(r=-0.335,p=0.036) and right amygdala volume(r=-0.337, p=0.048).


Introduction
Neurofilament light chain (NfL) plays an important role in axon transmission and function maintenance, and is the most abundant intermediate filament protein in myelinated subcortical axons an effective way to distingusih FTLD from early-onset AD given that NfL levels in AD are lower in early onset compared to those in late onset presentations 5,6 . However, NfL in cerebrospinal fluid is not applicable and difficult to accept for many elderly people, Therefore, blood-based measurement of NfL might be more desirable, since the collection of blood sample is relatively less invasive and more applicable 7 .
So far, there have been a few studies on NfL expression patterns in the blood of AD patients in China.
For example, Liu S 8 et al found that gastric cancer subjects expressed lower plasma NfL levels but AD subjects expressed higher plasma NfL levels than normal controls. Hu H 9 et al found that plasma NfL concentration and its rate of change had already increased abnormally in the preclinical phase of AD.
And Lin YS 10 et al also found that plasma NfL was significantly increased in the AD group, compared with the control, mild cognitive impairment (MCI), non-demented Parkinson's disease (PD), and Parkinson's disease dementia groups. These conclusions suggest NFL in plasma may represent a biomarker of cognitive decline in AD, and it is possible to mark the onset of neurodegeneration in subjects at risk for AD familial disease 1 . However, the mechanism of NfL regulating cognition is not very clear.
Neuroimaging using Magnetic Resonance Imaging (MRI) has been widely used to describe the atrophy pattern of cognitive related brain regions in AD and FTLD as well as to find differential trajectories along the different disease stages [11][12][13] . And structural MRI of medial temporal atrophy (MTA) is considered to be a biomarker for an early diagnosis of MCI and AD 14,15 , specifically speaking, volume reduction of medial temporal lobe, including amygdala and hippocampus has been proved to be an early manifestation of AD 16 . The relationship between classical AD biochemical markers AD and neuroimaging features and their reciprocal influence have been studied during both the preclinical phases and clinical of the disease 3 . For example,Mattsson N 17 et al pointed out that high plasma NfL levels in the MCI and AD cohort were associated with smaller hippocampal volumes, thinner cortices and larger ventricular. However, it is not known whether the above conclusions are applicable to Chinese people. Therefore, in this context, our goals were (a) to provide a descriptive analysis of plasma NfL levels and structural patterns (amygdala volume and hippocampal volume) in early-onset AD, amnestic mild cognitive impairment (aMCI) and normal control(NC); (b) to study the relationship between earlyonset AD, aMCI and NC brain structural measures/cognitive function and plasma NfL levels.

Data base
Data were obtained from the China Longitudinal Aging Study (CLAS) 18 database, which was launched in 2013 as a large scale longitudinal cohort study in China and was led by principal investigator Shifu Xiao, MD and PhD. The CLAS participants were recruited from 20 target communities (ie, 2 rural and 18 urban) located in the eastern, mid, and western parts of China, and all of them were permanent residents aged 60 years or older 19 .

Participants
Our study enrolled 30 normal controls (NC), 43 aMCI subjects and 35AD subjects, who were matched in age, gender and education. All subjects diagnosed with AD fulfilled the following criteria:

General demographic data
General demographic data, such as age, gender, education, smoking history, drinking history, tea drinking history and disease history(diabetes and hypertension) were collected with standardized questionnaires, and in the form of self-report.

Plasma NfL
Plasma NFL was examined by a highly sensitive assay on the Single Molecule Array (Simoa™) platform 22 . SiMoA NfL assay advantage kits (product #103186) were commercially obtained from Quanterix corporation(MA, USA). Critical reagents including buffer quality controls and recombinant human NfL standards were supplied frozen for single use only 23 . And all the samples were measured in duplicate.

Neuropsychological Tests
Montreal Cognitive Assessment-Chinese Version (MoCA-CV) 24 was used to assess the overall cognitive function of all subjects. The MOCA is an evaluation tool used to evaluate MCI, developed by Nasreddine 25 et al. And it is able to differentiate between MCI and early dementia as well as between normal and MCI 26 . The MoCA cut-offs (-1 to -2 standard deviations) for cognitive impairment was ranged from <25 to <21 for the lowest educated and <26 to <24 for the highest educated, depending on different age groups 27 .

Magnetic Resonance Imaging
MR images were scanned by using Siemens Magnetom Verio 3.0T scanner (Siemens, Munich, Germany). T1-weighted images were obtained from 176 sagittal slices using 3D magnetization prepared rapid gradient echo acquisition sequence with the following parameters: Spatial resolution = 1 * 1 * 1.2 mm3, TE = 2.98 ms ,TR = 2300 ms, Flip angle = 9°2 8  Automated procedures were utilized to ascertain volumetric data. The automated assessment was described by using the Learning Embedding for Atlas Propagation (LEAP) algorithm 30 . The volume and asymmetry with hippocampus and amygdala as well as the brain size index of each subject were extracted.

Statistical Analyses
Continuous variables were expressed as mean ± SD and categorical variables were expressed as frequencies (%). Single sample Kolmogorov-Smirnov test was used to check whether the data conform to the normal distribution. Single factor ANOVA was used to compare the data of normal distribution among AD group, aMCI group and normal group, while Kruskal-Wallis H(K) test was used to compare the data of non -normal distribution. And chi square test was used to categorical variables among the three groups. Then partial correlation analysis(controlled for hypertension) was used to explore the association between plasma NfL and volume of cognitive related brain areas(hippocampus and amygdala). Two-tailed tests were utilized at a significance level of P < 0.05, and all statistical analyses were performed using SPSS 22.0 (IBM Corporation, Armonk, NY, USA).

Results
There were statistical differences (p<0.05) in hypertenion, MoCA, left hippocampus, right hippocampus, left amygdala and right amygdala among AD group, aMCI group and normal group, but no significant differences (p>0.05) in age, education, gender, smoker, drinker, tea drinker and diabetes. Table 1 shows the results. Then further comparison showed that the MOCA scores of normal group, aMCI group and AD group decreased in turn. The volume of hippocampus and amygdala and in aMCI group and normal group were larger (p<0.05)than that in AD group, but there was no significant difference(p>0.05) between aMCI group and normal group. However, the concentration of plasma NfL in aMCI group and normal group were lower (p<0.05)than that in AD group, but there was no significant difference(p>0.05) between aMCI group and normal group(figure 1). And table 2 presents the results. Then by using partial correlation analysis and controlled of hypertension, we found that plasma NfL was negatively correlated (p<0.05)with MoCA total score (r=-0.415, p=0.013), right hippocampal volume(r=-0.335,p=0.036) and right amygdala volume(r=-0.337, p=0.048).

Discussion
In the present study, we analyzed and compared the plasma concentration of NfL as well as the volume differences in hippocampus and amygdala among patients with AD, patients with aMCI and normal elderly. Finally, we reached some interesting conclusions: a) the concentration of plasma NfL in aMCI group and normal group were lower (p < 0.05)than that in AD group, but there was no significant difference(p > 0.05) between aMCI group and normal group; b) the volume(both left and right) of hippocampus and amygdala in aMCI group and normal group were larger (p < 0.05)than that in AD group, but there was no significant difference(p > 0.05) between aMCI group and normal group; c) plasma NfL was negatively correlated (p < 0.05)with MoCA total score (r=-0.415, p = 0.013), right hippocampal volume(r=-0.335,p = 0.036) and right amygdala volume(r=-0.337, p = 0.048).
NfL is an important cytoarchitectural protein present primarily in large-caliber myelinated axons 31 ,and increased NfL in CSF will indicate damage or degeneration of these axons. This protein appears to be relatively independent of tau and amyloid levels, and might correlate with symptomology, progression, and survival 32 . Now NfL has been considered as an especially promising biomarker for neurodegeneration because it can be measurable in plasma 33 . Single molecule array (SiMoA) digital immunoassay platform can offer heightened sensitivity by miniaturizing the reaction volume 34 , and several biomarkers indicative of peptide fragmentation and neuronal dysfunction can be accurately quantified in peripheral blood. In our study, we used this Simoa technology to compare the plasma NFL concentrations of AD, aMCI and normal elderly people(their gender, age and education are matched), and found the concentration of plasma NfL in AD group were significantly higher(p < 0.05) than that in aMCI group and normal group. Previous studies had confirmed that AD was associated with elevated NFL in plasma, so our findings were consistent 8,9 . However, we did not find any difference in plasma NfL between patients with aMCI and normal elderly, which was inconsistent with the conclusion of Zhow W 7 et al. So the relationship between NfL and aMCI needs to be further verified.
Atrophy of the medial temporal lobe, a critical region involved in memory formation, is considered as a recognized marker for AD 35 . The hippocampus and amygdala, both residing in the medial temporal lobe, are proved to be associated with declarative memory and emotional processing, respectively 36 .
As the atrophy of hippocampal and amygdalar are already evident in the prodromal stage of AD 37 , the volumetric measurements of hippocampus and amygdala have been used to assist the clinical diagnosis of AD 38 and to predict the cognitive status of the elderly 39 . As expected, the volume of hippocampus and amygdala in AD patients was significantly smaller than that in aMCI and normal old people, in concordance with previous publications [39][40][41] . Then by using partial correlation analysis (adjusting hypertension), we found plasma NfL was negatively correlated with MoCA total score And Weston PSJ 43 et al also found that serum NfL correlated with baseline brain volume and wholebrain atrophy rate in patients with familial Alzheimer disease. So we had jointly confirmed that there was a certain correlation between plasma NfL and cognitive related brain regions. However, we only found that plasma NfL was related to the right hippocampus and right amygdala, but not to the left.
Wang L 44 et al found that the functional connectivity between the right hippocampus and a set of regions(such as medial prefrontal cortex, right inferotemporal cortex, right cuneus extending into precuneus, ventral anterior cingulate cortex, left cuneus, right superior and middle temporal gyrus and posterior cingulate cortex)was disrupted in AD. Sohn WS 45 et al also believed that the right hippocampal connectivity was relatively by AD progression. And López-Jaramillo C 46 et also found that lithium-treated bipolar I disorder patients had larger right thalamus than unmedicated patients and controls. Therefore, we speculated that plasma NFL had some intrinsic relationship with the right hippocampus and the right amygdala, although the mechanism was not fully understood.
Our study has certain limitations: first, it's just a cross-sectional study that could not establish a causal link; second, the relatively small sample size reduced the reliability of the study; third, we lacked the gold standard to diagnose AD and aMCI. 6. Conclusions 9 NFL in plasma of AD patients was significantly increased, and the protein was related to atrophy of right hippocampus and right amygdala. University School of Medicine. All participants had signed the informed consent written informed consent before the start of the study.

Consent for publication
Not applicable.

Availability of data and materials
The data base generated and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Authors' contributions
Wei Li and Lin Sun contributed to the study concept and design. Ling Yue acquired the data. Ye Wu collected the data. Shifu analyzed the data and drafted the manuscript. All authors have read and approved the final manuscript.

Acknowledgement
Not applicable.