Introduction

Microglia activation is a characteristic feature of Alzheimer’s disease (AD). Although microglia cells have been found to contact neurofibrillary tangles in the brain of Alzheimer’s patients ([1]), the association of microglia to amyloid plaques is a more frequent and distinguished event ([2]; [3]). Plaques are extracellular proteinaceous aggregates mainly composed of the β-amyloid peptide (Aβ). Proteolytically derived Aβ monomers successively aggregate into oligomeric species and thence amyloid fibrils, ultimately forming the pathognomonic amyloid plaques which gradually grow in size ([4]). In the vicinity of amyloid plaques neurons are damaged, as apparent by dystrophic neurites, the progressive loss of dendritic spines as well as the loss of axons and neurons ([5]; [6]; [7]; [8]).

Microglia are the resident immune cells of the brain actively surveying the brain parenchyma with their fine processes. After brain injury or immunological challenges, microglia are activated leading to the upregulated expression of immune-related molecules, including interleukin-1 (IL-1) and transforming growth factor β ([9]; [10]). Once activated, microglia prominently change their morphology: the ramified processes swell and withdraw, whilst their cell bodies enlarge ([11]). In the AD brain, activation of microglia and clustering around amyloid deposits tend to occur early in the development of plaque pathology. Nevertheless, little is known about the initiating event of this interaction. Here, we specifically focus on very small newly-formed amyloid plaques and attempt to characterize the process by which microglia are attracted to these lesions.

Material and Methods

Results

On brain slices of 12-month-old APP-PS1(dE9) mice, microglia cells were immunohistologically stained and fibrillar amyloid deposits were labelled with Methoxy-XO4. Rarely, we observed small amyloid plaques that were devoid of any detectable microglial contact (Fig. 1A); in most cases, similar sized plaques were clearly contacted by microglia, partly exhibiting an activated appearance (Fig. 1B). To characterize the relationship between plaque size and microglial activation, we specifically imaged small amyloid plaques (which are supposed to correspond to novel plaques) in cortex or hippocampus. Therefore, plaques with a diameter of up to ~10 μm (equates a volume of 524 μm³) were chosen randomly, while the microglia channel was not visible to the investigator. Image-stacks were recorded by confocal microscopy covering the total plaque volume with additional 10 μm beneath and above each plaque. Subsequently, the microglial status around the plaque was classified into five levels: (0) Amyloid plaque devoid of any detectable microglial contact; (1) Amyloid plaque contacted by a single surveying microglial process; (2) Amyloid plaque surrounded by various microglial processes; (3) Amyloid plaque enclosed by activated (explicitly swollen) microglial processes; (4) Amyloid plaque insulated from neuropil by activated microglial cell bodies (Fig. 1A-B). Whereas level 3 and level 4 exhibit distinct morphological signs of microglia activation, at level 2 the microglia start to recognize and orientate toward the plaque. At level 0 and level 1, shape and number of microglia processes at and nearby the amyloid plaque are indistinguishable from plaque-free brain areas. It should be noted that all plaques shown in Fig. 1 A-B are of comparable size. Hereunder, we will refer to activated microglia cells based on their morphological appearance (level 3 and level 4).

Download:

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

Fig 1. Activation of microglia by small amyloid plaques.

On brain slices of 12-month- old APP-PS1(dE9) mice microglial cells were immunohistochemically labelled for Iba1 (green) and amyloid plaques were stained with Methoxy-X04 (red). Small amyloid plaques were imaged and classified according to the microglial status. A Image projections of an amyloid plaque lacking any microglial contact (level 0) at three magnifications. Below, the single x-y-planes clarifying that no microglial process touches the plaque. B Exemplary images of the stages of plaque-associated microglial (MG) activation (level 1 to level 4). The upper lower magnification illustrates the microglial environment around the plaque, the higher magnification below highlights the individual plaque. Notice that, although each plaque has a similar size, microglial reaction is diverse. C Small plaques were classified according to their microglial activation level and plotted against the plaque volume. D The number of microglial cells associated to the individual plaques was plotted against the plaque volume. (The brains of five animals were analysed with a total of 188 small plaques. Red bars show mean with standard deviation.)

https://doi.org/10.1371/journal.pone.0119768.g001

In total, 188 small amyloid plaques were investigated from the brains of five animals. Six very small plaques (volume between 0.9 and 33.5 μm³) were lacking any microglial detectable contact (level 0) (Fig. 1C). An additional 32 plaques were classified as level 1, with volumes reaching up to 168.3 μm³ while 40 plaques were classified as level 2 (volume between 3.6 and 143.8 μm³). About 50% of small amyloid plaques were associated with clear morphological signs of microglia activation (level 3 and 4; Fig. 1C). The recruitment and activation of microglial cells correlate weakly with plaque size (R² = 0.1040). For example 73% of plaques in level 3 are smaller than 100 μm³, 12% are even smaller than 20 μm³. Analyzing the number of microglia, we found that most small amyloid plaques were associated with only one microglial cell (Fig. 1D). Only a very tiny quantity of small plaques were surrounded by three or four microglia, with the volume of these plaques ranging from 37.4 μm³ to 434.9 μm³ (Fig. 1D). These results indicate that small plaques present a variable recruitment and activation of microglial cells around them.

To confirm the ex vivo data and explore the dynamic of microglia activation by amyloid plaque formation we utilized in vivo two-photon microscopy. Therefore, we crossbred APP-PS1(dE9) mice with CX3CR1 heterozygous knockout mice, which carry the EGFP gene at the CX3CR1 locus providing a specific fluorescent labelling of microglia in the brain. Although previous studies have shown an effect of heterozygous and homozygous knockout of CX3CR1 on amyloid plaque burden in Alzheimer-transgenic mice ([16]; [17]), in our hands plaque density, plaque area, plaque size and microglia area were indistinguishable in cortex and hippocampus of APP-PS1(dE9) and APP-PS1(dE9)xCX3CR1^+/- (S1 Fig.).

Starting at an age of six months we repetitively imaged superficial volumes in the somatosensory and the primary motor cortex of APP-PS1(dE9)xCX3CR1^+/- mice to scan for newborn, Methoxy-X04 labelled amyloid plaques. To distinguish very small plaques from autofluorescent spots we recorded additionally to the Methoxy-X04 emission (short pass 485 nm) nonspecific emission at 590–650 nm (S2 Fig.). New amyloid plaques were identified as Methoxy-X04 specifically-labelled structures that remained in place and typically grew in size over time (S3 Fig.). In four mice, we were able to detect 21 new appearing plaques in total and followed the microglial reaction on these plaques for up to 106 days (Fig. 2A). Interestingly, at the imaging time-point of first appearance out of the 21 new appearing plaques we could not observe a single plaques without any detectable microglial contact (level 0) (Fig. 2B). Except for one plaque, at the first detection time point all plaques were classified as level 1 or level 2, lacking clear signs of microglial activation. Over time, most plaques induced a more serious morphological activation of microglia, typically terminating at level 4. Nevertheless, the time point when microglial activation was reached differed for individual plaques, ranging for level 4 between 7 and 42 days after first detection (Fig. 2B). Similarly, the plaque volume for microglial activation was different, with the smallest plaque inducing level 4 at a size of 10.3 μm³ and the largest at a size of 104.8 μm³ (Fig. 2C). There was no correlation between plaque volume and plaque lifetime at the time point, when microglial activation level 3 as well as level 4 was reached (Fig. 2D). Occasionally we also observed a deactivation of microglia from a high level to a lower one (Fig. 2A-C). Together these results suggest that the formation of individual amyloid plaques precedes the activation of microgial cells. Furthermore, the recruitment of microglia does not seem to be directly dependent on the lifetime or the initial size of the corresponding plaque. Although, eventually, all amyloid plaques are surrounded by activated microglial cells (level 3–4), there is substantial variability between individual plaques before reaching this level.

Download:

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

Fig 2. In vivo monitoring of amyloid plaque induced microglial activation.

In the cortex of PS1(dE9)xCX3CR1^+/- mice, individual new appearing plaques (red, Methoxy-X04 stained) and the subsequent microglia activation (green) were followed using two-photon microscopy. A Time series of image projections of a newly formed plaque (starting with day 1) and the corresponding microglial reaction. The upper row indicates the microglial environment, the lower row shows the plaque in higher magnification. Notice, at day 1 a microglial process is already contacting the plaque (level 1). At day 9 this plaque was classified as level 3 and level 4 was reached at day 29. Interestingly, at day 88 the plaque is mainly liberated from enclosing microglia, which at day 92 is reverted to level 4 again. B Time dependent development of microglial activation level for each plaque (colour coded) beginning with the time point of appearance. The box plots below indicate the distribution of initial time points when microglia activation was reached (level 3 and 4). C Size dependent development of microglial activation for individual plaques (colour coded). The box plots below indicate the distribution of initial volume when microglia activation was reached (level 3 and 4). D No correlation between plaque lifetime and plaque volume when microglial activation level 3 or 4 are reached.

https://doi.org/10.1371/journal.pone.0119768.g002

Discussion

In the brain of AD patients, the amyloid burden is accompanied by clustering of activated microglia around these amyloid plaques. The role of microglia in this pathology is controversial. Hypotheses range from clearance of Aβ-aggregates to potential negative inflammatory effects and neuron elimination mediated by microglia ([18]; [19]; [20]). In this descriptive study, but the first of its kind, we aimed to analyse the initiation of microglial interaction with amyloid plaques in APP-PS1(dE9) transgenic mouse line. For very small plaques we distinguished five levels of increasing interaction in brain slices, four of which could be confirmed by in vivo two-photon microscopy. We observed that the degree of microglia reaction showed a great variation with regards to size as well as lifetime of individual small plaques.

In a more detailed description, we therefore herewith confirm the original finding by Meyer-Luehmann and colleagues that the formation of amyloid plaques precedes the recruitment of microglia ([21]). Typically, the plaque is touched by fine microglial processes before swollen processes and finally activated microglial cell bodies envelope the whole plaque. However by the data it cannot be ruled out, that microglia might be involved in the formation of plaques by releasing certain factors into the microenvironment.

An important question is whether the amyloid deposit itself is sensed by the microglia and induces the subsequent microglial reactions. It has been shown that fibrillar Aβ can directly activate microglia IL-1 expression ([22]). Nonetheless, it is not clear why the here observed reaction of microglia to the individual small plaques is so diverse. After emergence of a plaque, we observed that the morphological activation of microglia takes between 7–42 days to start. The variation in initiation sizes and initiation time points after plaque formation of unequivocal microglial activation supports the hypothesis that secondary consequences of amyloid deposition lead to the recruitment of microglia. This view is further supported by the fact that indeed sporadically small plaques without any detectable association to microglia have been found. One possible explanation could be that a developing amyloid plaque directly injures at some time point its surrounding neuropil, and the subsequent release of neuron-derived signalling molecules induces the recruitment of nearby microglial cells. Accordingly, neuritic dystrophies which accompany the development of amyloid plaques might represent the site of acute neuronal tissue damage (S4 Fig.). The different degree of neuropil alteration around a plaque over time could explain the different levels of microglial activation, but at a certain point all plaques are surrounded by activated microglia. However, we cannot rule out that neuronal injury might also be a consequence of microglia activation. Although the microglia in humans might respond differently to fibrillar Aβ, in AD brains microglial activation has been shown to progress with the neuropathological stage of the disease as well as with the progression of individual amyloid plaques ([23], [24]). About one fifth of early, non-fibrillar amyloid plaques in AD brains were found not to be associated with microglia, again questioning whether Aβ aggregates per se can recruit these cells ([23]). Interestingly, the late-stage dense core plaques showed less microglial association, paralleled by a reduced number of dystrophic neurites ([23]). Presumably, at this final stage the immediate vicinity of the plaque lacks neuronal tissue that can be damaged.

In summary, we have provided a detailed description of the initial interaction of microglia with newly developing amyloid plaques in an AD mouse model. Our results indicate that the appearance of plaques alone might not be the direct cause of microglia recruitment. Alternatively, the progressive tissue alteration around the plaques might be related to the microglia activation. Clearly, the trigger which actually attracts the microglia to the amyloid lesion remains to be identified.

Supporting Information