Abstract
Background/Aim: We investigated whether mastication affects microglia, whose activity is thought to be associated with cognition and brain tumor progression. Materials and Methods: We kept mice by feeding either a hard or soft diet for 2, 4 or 8 months. After each period, we removed the whole brains and isolated microglia. The total RNA extracted from each brain's microglia was subjected to DNA microarray analysis. Results: Many genes were found to be significantly differentially expressed between hard- and soft-diet-fed mice in each group of the same feeding period. The expression of several genes involved in the regulation of actin cytoskeleton was down-regulated in the soft-diet-fed mice. Conclusion: Mastication may affect microglia's roles in cognition as well as their neuroimmune activity through their activity of patrolling the brain.
Weijenberg et al. (1) have concluded that mastication is related to cognition as well as nutrition and activities of daily living in human and animals. The review also suggested the existence of causality for these relations and the possible underlying physiological mechanisms. For example, an elderly Swedish population study suggested an association between mastication ability and cognitive impairment regardless of natural teeth or prostheses if there is no mastication difficulty (2). An animal study showed that soft-diet-fed mice have reduced synaptic formation in the cerebral cortex and impaired spatial learning ability compared with hard-diet-fed mice (3). Another animal study demonstrated that soft-diet feeding impairs neurogenesis in the subventricular zone and olfactory functions in mice and that this impairment can be recovered by feeding mice with hard diet (4).
Cognition is generated by the brain that consists of neurons and glial cells, which can be subdivided into several types such as astrocytes, oligodendrocytes, and microglia. It is considered that cognition would be better understood by investigating a comprehensive map of neural connections in the brain (termed connectome). Fields et al. (5) proposed that the connectome should include glia, because the neural connectome is regulated by glia through local and long-range communications such as stimulation of synaptogenesis by astrocytes, determination of the patterns of neuronal connectivity through myelination by oligodendrocytes, and synapse pruning by microglia.
Microglial cells, which constitute ~10% of the glial cells in the brain, are derived from yolk sac differently from neurons and the other types of glia, which are from the ectoderm. Microglia were once considered to work only as inflammatory mediators like macrophages of the central nervous system. This view of microglia has been drastically changed by recent findings including the unique transcriptome signatures of microglia (6-9). Microglia are found to play physiological roles in synaptogenesis, trophic support, neurogenesis, and surveilling their surroundings. Also, their functional alterations are also implicated in neurodegenerative diseases such as Alzheimer's disease and chronic pain. Furthermore, it has been shown that microglia polarization affects glioma progression (10, 11).
We investigated whether mastication affects microglia by comparing microglia transcriptomes from hard-diet-fed mice with those from soft-diet-fed mice. We found that many genes were differentially expressed between these two groups. A functional annotation analysis showed that the expression of several genes involved in the regulation of actin cytoskeleton was down-regulated in the soft-diet-fed mice, suggesting that mastication is important for the motility of microglia that patrol the brain.
Materials and Methods
Mice. Twelve 10-weeks old C3H/HeNJcl male mice were purchased from Charles River Laboratories (Kanagawa, Japan). After a one-week acclimation-breeding period, half of them were fed with a pelleted (hard) diet (CE-2, FEED ONE, Kanagawa, Japan) and another half with a powdered (soft) diet (CE-2) for 2, 4 or 8 months. These diets contained the same ingredients. The experiments were performed in accordance with the Guidelines on the Care and Use of Laboratory Animals issued by Niigata University of Pharmacy and Applied Life Sciences. The protocol was approved by the Committee on the Ethics of Animal Experiments of Niigata University of Pharmacy and Applied Life Sciences (Permit Number: H28-3).
Microglia preparation. After 2, 4 or 8 months, four mice, each two of which were fed with either a hard or soft diet, were euthanized by cervical dislocation, and the whole brains were removed. The brain/body weights (g/g) of 2-month-fed mice were 0.34/27 and 0.32/27 for the hard-diet-fed mice, and 0.33/27 and 0.33/27 for the soft-diet-fed mice. Those of 4-month-fed mice were 0.33/27 and 0.36/27 for the hard-diet-fed mice, and 0.36/32 and 0.34/32 for the soft-diet-fed mice. Those of 8-month-fed mice were 0.36/33 and 0.33/28 for the hard-diet-fed mice, and 0.33/30 and 0.33/30 for the soft-diet-fed mice.
Microglia were isolated from each brain according to the established method (12, 13). The brain was removed, kept in 2–3 ml of DMEM (Wako, Tokyo, Japan) containing 1.2 units/ml dispase II (Roche, Basel, Switzerland), 1 mg/ml papain (Wako) and 20 units/ml DNase I (Takara Bio, Shiga, Japan), and cut finely with a sharp razor. After incubation of the brain pieces at 37°C for 10 min, they were passed through an 18G needle (Terumo, Tokyo, Japan) 5 times and subsequently through a Pasteur pipette (Iwaki, Shizuoka, Japan) 5 times to make them finer. The resulting brain homogenate was further incubated at 37°C for 10 min. The last two steps were repeated. After terminating the enzymatic reactions by adding DMEM containing 10% fetal bovine serum (FBS; Sigma-Aldrich, MO, USA), the separated cells were centrifuged for 5 min at 300 × g at room temperature, and the pellets were re-suspended in phosphate buffered saline (PBS) containing 0.5% bovine serum albumin (BSA; Wako). Tissue debris was removed by passing the cell suspension through a 40-μm cell strainer (Falcon, MA, USA). The cells collected by centrifugation were re-suspended in PBS containing 0.5% BSA and brain magnetic myelin removal beads (Miltenyi Biotec, Bergisch Gladbach, Germany). According to the manufacturer's protocol, myelin was removed with a magnetic cell separation system, MACS (Miltenyi Biotec). After removing myelin, CD11b positive cells were isolated using CD11b microbeads (Miltenyi Biotec) with the MACS. The isolated CD11b positive cells were stained with Trypan blue, and the number of live and dead cells was counted under the microscope. We regarded the CD11b positive cells (75–100% viable) as microglia.
DNA microarray analysis. Total RNA from microglia of the 2-month-fed mice was extracted with RNAiso Plus (Takara Bio, Otsu, Japan). The RNA samples were amplified with an Ovation Pico WTA System V2 (NuGEN), and subjected to DNA microarray analysis with a SurePrint G3 Mouse GE v2 8x60K Microarray (Agilent Technologies, Tokyo, Japan). This microarray analysis was carried out by Takara Bio Inc. (Shiga, Japan).
Total RNA from microglia of the 4-month- and 8-month-fed mice was extracted with an miRNeasy Micro Kit (Qiagen), and the RNA samples were subjected to DNA microarray analysis with the SurePrint G3 Mouse GE v2 8x60K Microarray. This microarray analysis was performed by Subio (Kagoshima, Japan).
Bioinformatic analysis. Raw signal data obtained from the analysis with the SurePrint G3 Mouse GE v2 8x60K Microarray were imported to a data analytics platform, Subio Platform version 1.22. The microarray contains multiple probes for one gene in a subset of the genes. The raw data of each experimental group were normalized globally to adjust means of the raw signal values, and probes that contained at least one abnormal peak signal (gIsFeatNonUnifOL=1) were removed to obtain filtered data sets (QC1). In addition, probes whose signals were too low (gIsWellAboveBGs=0) to be reliably measured in all four samples were also excluded to obtain further filtered data sets (QC2). Then, we transformed the raw signal values to log2 ratio values by setting the mean of two raw signal values of each probe in two transcriptomes of hard-diet-fed mice microglia to zero. We excluded probes whose mean log2 ratio values in two transcriptomes of soft-diet-fed mice microglia were in the range of −0.5 and 0.5, and termed these filtered data sets QC3. Similarly, the combined transcriptome data from the 4-month- and 8-month-fed groups were analyzed.
Volcano plots for the data sets QC3 were drawn with an analysis tool, Compare to Control on Subio Platform. Significance (p-value of one sample t-test) and fold-change were put on the y and x axes, respectively. Each set of probes (termed QC4) corresponding to points on a volcano plot that had p-values of <0.05 and fold-change values of >1.5 were subjected to hierarchical clustering with uncentered correlation using an analysis tool, Tree Clustering, to obtain clustering trees.
Data sets of differentially expressed genes were further analyzed with a functional annotation tool on the Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.8 to obtain KEGG pathways implicated with the differentially expressed genes (14-16).
Mouse microglia genes implicated in health and disease were analyzed by extracting them from the QC2 data sets and normalizing their raw signal values against those of GAPDH. The difference in the expression level of each gene was presented on a heat map drawn with the Microsoft Excel 2016 for each mice group of the same feeding period. The mean of ratios of a mean value for two soft-diet-fed mice to that of two hard-diet-fed mice was calculated and graphed with a standard deviation for each category of health and disease.
Accession number. DNA microarray data were deposited in the DNA Data Bank of Japan (17) with the accession number E-GEAD-303.
Results and Discussion
Gene expression characteristic for mouse microglia. Ten weeks old C3H/HeNJcl male mice were fed either a hard or soft diet for 2, 4 or 8 months. After each period, we removed the whole brains from four mice, two of which were fed with either a hard or soft diet, and isolated microglia from each brain. The total RNA extracted from each brain's microglia was subjected to DNA microarray analysis.
Firstly, we examined if each transcriptome shows gene expression characteristics for mouse microglia. As expected, because microglia were isolated using an anti-CD11b antibody, the CD11b gene was expressed comparably to or more highly than the two housekeeping genes GAPDH and RPL5 (data not shown). The AIF-1/Iba1 gene, which is known to be expressed in microglia, was moderately expressed. We also examined the expression levels of 13 cell surface genes Krtcap2, Slco4a1, Mcoln3, Tmem55b, Slc30a5, Stab1, Pap2c, Tmem48, Mfsd10, Lrp8, Cmtm4, CD40, and Lpcat3, which have been reported to be specific to mouse microglia (18).
Mastication affects transcriptomes of mouse microglia. A. Volcano plots for the QC3 data sets are shown with respect to the 2-month-, 4-month-, and 8-month-fed mice groups. The shaded areas correspond to the QC4 data sets. The numbers of the gene probes of the QC3 and QC4 data sets are at the lower right of each plot. B. Clustering trees for the QC4 data sets were obtained by hierarchical clustering with uncentered correlation.
The expression profiles of the genes. 56, 6, and 12 KEGG pathways were implicated in the significant differences in the gene expressions for the 2-month-, 4-month-, and 8-month-fed mice groups, respectively. And 7 pathways were shared by 2 groups. The TNF signaling pathway (A) and the NOD-like receptor signaling pathway (B) were shared by the 2-month- and 4-month-fed groups, and the phosphatidylinositol signaling pathway (C) and the thyroid hormone signaling pathway (D) were by the 2-month- and 8-month-fed groups. The expression profiles (log2 ratio values) of the genes shared by the 2 groups in each pathway are shown with respect to all 6 individual mice that were fed with a soft diet. The numbers, which were assigned to the individual mice in the same group, correspond to those in Figure 1.
Most of the genes were expressed comparably to or less than the Iba1 gene. Further, we examined the expression levels of 14 transcriptional regulator genes, 24 genes encoding for cell surface proteins, and 5 other microglia signature genes, the expression levels of which are known to be specific to mouse microglia (18-20), and found that many of the genes were expressed comparably to or more than the Iba1 gene (data not shown).
In several cases, the values of raw signals (e.g. in CD11b and Krtcap2) differed largely in microglia transcriptomes between 2-month-fed mice and 4-month-fed and 8-month-fed mice. This is probably mainly because the amplification procedure for low amount RNA samples was applied to the former analysis. Since in the subsequent analyses we compared gene expression levels between hard-diet-fed mice and soft-diet-fed mice that were kept for the same period of time, these differences would not matter.
Mastication affects transcriptomes of mouse microglia. Next, we investigated how mastication affects the transcriptomes of mouse microglia by comparing microglia transcriptomes from hard-diet-fed mice with those from soft-diet-fed mice. We extracted significantly differentially expressed genes, which are shown as dots in the shaded areas of the volcano plots, and termed the extracted data set QC4 (Figure 1A). The areas correspond to p-values of <0.05 and fold-change values of >1.5. In the 2-month-fed mice group, 3,959 of 25,000 genes in the QC3 data set were significantly differentially expressed. In addition, 682 of 5,139 genes and 1,828 of 11,001 genes were expressed significantly differentially in the 4-month-fed group and in the 8-month-fed group, respectively (data not shown).
The significantly differentially expressed genes (the QC4 data set) in each group were analyzed by hierarchical clustering with uncentered correlation to obtain a clustering tree (Figure 1B). In all groups, the gene expression profiles of two soft-diet-fed mice were very similar, while those of two hard-diet-fed mice were not. In the 2-month-fed group, 65% of the genes were significantly highly expressed, whereas the expression levels of only 10% and 8% of the genes were significantly high in the 4-month- and 8-month-fed groups, respectively.
Differentially expressed genes in commona.
Furthermore, we examined which pathways are affected by mastication by analyzing the significantly differentially expressed genes in each group with DAVID, a functional annotation tool. The above 682 and 1,828 genes of the 4-month- and 8-month-fed groups were analyzed, respectively. However, 2,862 genes of the 3,959 genes of the 2-month-fed group, which had p-values of <0.05 and fold-change values of >2 were analyzed. The reduction in the gene number from 3,959 to 2,862 in the 2-month-fed group was due to the limitation of the upload list size of 3,000 genes (21). The functional annotation analysis for the 2-month-, 4-month-, and 8-month-fed mice groups showed 56, 6, and 12 KEGG pathways, respectively, of statistical significance (data not shown). None of the pathways was shared by all three groups, and 7 pathways were shared by two groups. The TNF signaling pathway and the NOD-like receptor signaling pathway were shared by the 2-month- and 4-month-fed groups, whereas the phosphatidylinositol signaling pathway, the thyroid hormone signaling pathway, the pathways in cancer, the proteoglycans in cancer, and the amoebiasis were shared by the 2-month- and 8-month-fed groups. The expression profiles of the genes shared by the two groups in each signaling pathway are shown in Figure 2.
The expression of the genes selected in the four signaling pathways tended to decrease with feeding periods with the exception of Plce1 (Figure 2). The observation that the expressions of Il6 and Tnfrsf1a in the TNF signaling pathway changed from relatively high levels in the 2-month-fed group to relatively low levels in the 4-month- and 8-month-fed groups suggests that mastication affects the neuroimmune activity of microglia.
Mastication affects regulation of actin cytoskeleton in mouse microglia. We extracted genes that all three QC3 data sets from 2-month-fed, 4-month-fed, and 8-month-fed mice groups have in common and analyzed them with the functional annotation tool of DAVID (Table I). The analysis showed 2 KEGG pathways of statistical significance, the regulation of actin cytoskeleton and the RAS signaling pathway. In the regulation of actin cytoskeleton, the 5 genes SOS1, GSN, PFN1, ITGA1, and FGFR1 were down-regulated with respect to at least one probe, and in the RAS signaling pathway, the 4 genes SOS1, FGFR1, VEGFA, and RASGRP1 were down-regulated and the gene FOXO4 was up-regulated with respect to at least one probe (Figure 3). SOS1 and FGFR1 were involved in both pathways, and the 2 pathways are known to be moderately related with each other (21). These observations suggest that feeding mice with a soft diet affects the regulation of actin cytoskeleton and may result in dampening microglia's activity of surveilling their surroundings.
Mastication affects expression of mouse microglial genes implicated in health and disease. Lastly, to examine how mastication affects expression of mouse microglial genes implicated in health and disease (7, 9), we compared their expression levels between the hard-diet-fed mice and the soft-diet-fed mice. A heat map of the difference in expression of a gene among 4 mice was presented for 2 health categories, synaptic plasticity and neuronal programmed cell death, and 5 disease categories, Alzheimer's disease, frontal temporal dementia, neuropathic pain, genes induced in the microglial neurodegenerative phenotype (MNP), and genes suppressed in the MNP (Figure 4A). And the means of ratios of a mean value for 2 soft-diet-fed mice to that for 2 hard-diet-fed mice were graphed for each category of health and disease (Figure 4B). Although the relative expression profiles differed with genes and feeding periods, interestingly the mean values in microglia of the soft-diet-fed mice were ~1.5–2-fold higher and ~1.3-fold lower for the 2-month-fed mice and 4-month-fed mice, respectively, in each category. The mean values for the 8-month-fed mice were almost the same in most of the categories.
Mastication affects regulation of actin cytoskeleton in mouse microglia. The functional annotation analysis of the genes that each QC3 data set has in common showed the 2 KEGG pathways of statistical significance the regulation of actin cytoskeleton and the RAS signaling pathway. The expression profiles (log2 ratio values) of the genes selected in both pathways are shown with respect to all 6 individual mice that were fed with a soft diet. FGFR1 and VEGFA are members of RTK and GF, respectively, and FOXO4 is a synonym of AFX. The numbers, which were assigned to the individual mice in the same group, correspond to those in Figure 1.
In summary, the microglial expression of genes implicated in health and disease increased by feeding mice with a soft diet for 2 months and the elevated gene expression levels reached those for the hard-diet-fed mice after feeding with a soft diet for 6 more months. And this happened through the point where the expression levels further went down. These observations suggest that although a shift from a hard diet to a soft diet perturbs the expression of these genes in 2 months, microglia can restore this perturbation by acclimating to the soft diet in six more months.
Potential effects of mastication on cognition and brain tumor through microglial activities. In general, the microglial expression of the genes involved in brain's important physiological properties, synaptic plasticity and neuronal programmed cell death, were augmented by feeding mice with a soft diet for 2 months, but were restored to the normal levels in 6 more months (Figure 4). TNFα is a common key player in both properties (22-24), and the expression of the genes Il6 and Tnfrsf1a, important in the TNF signaling pathway, were also changed in a similar fashion (Figure 2). These observations imply that mastication could affect cognition through microglial physiological activities. In addition, mastication could affect brain tumor progression by changing microglia polarization through the TNF signaling pathway.
In contrast to the above genes, the microglial expression of the five genes involved in the regulation of actin cytoskeleton was down-regulated for all soft-diet-fed mice during the whole experimental period with one exception (Figure 3C). This suggests that the perturbation of the regulation of actin cytoskeleton by soft-diet feeding lasts for a longer period of time and that mastication would affect microglia's roles in synaptogenesis and neurogenesis as well as their neuroimmune activity against brain tumor through their activity of surveilling their surroundings.
Mastication affects expressions of mouse microglial genes implicated in health and disease. Mouse microglia genes implicated in health and disease were analyzed with respect to synaptic plasticity, neuronal programmed cell death, Alzheimer's disease, frontal temporal dementia, neuropathic pain, genes induced in the MNP, and genes suppressed in the MNP. A. The difference in the expression level of each gene was presented on a heat map for each mice group of the same feeding period. B. The mean of ratios of a mean value for two soft-diet-fed mice to that for two hard-diet-fed mice was graphed with a standard deviation for each category of health and disease. The numbers, which were assigned to the individual mice in the same group, correspond to those in Figure 1.
Acknowledgements
The Authors would like to thank Dr. Akihiko Komuro for helpful discussion in the early stage of this study, and Dr. Akihiko Komuro and Dr. Koichi Kawahara for critical reading of the manuscript. This work was supported by the Nikkyoko Research Award, and the Council for Science, Technology and Innovation (CSTI), Cross-ministerial Strategic Innovation Promotion Program (SIP) and ‘Technologies for creating next-generation agriculture, forestry and fisheries’ (funding agency: Bio-oriented Technology Research Advancement Institution, NARO).
Footnotes
Authors' Contributions
MS, HI, GS, HT and MN designed the study. AH and TI performed the experiments. HI and GS provided the experimental technique for microglia preparation. MS and MN conducted the data analysis. HT and MN supervised the study. MN wrote the manuscript, and all Authors approved its final version.
Conflicts of Interest
The Authors declare no conflicts of interest regarding this study.
- Received May 23, 2020.
- Revision received June 16, 2020.
- Accepted June 17, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.