Innate sensing of mechanical properties of brain tissue by microglia

Abstract

External organic or inorganic objects (foreign bodies) that are inadvertently or purposefully placed in the human or animal tissues can trigger local tissue responses that aim at the elimination and/or segregation of foreign bodies from the tissue. The foreign body response (FBR) may have major implications for neurodegeneration associated with the formation of aberrant protein-based aggregates or plaques. The distinct physical features of the plaques, including high rigidity and varying surface properties, may trigger microglial mechanosensing of the plaque as a foreign body. The microglial FBR may have a dual function by promoting and/or suppressing the plaque driven neurodegeneration. Microglial contact with the plaque may trigger inflammatory activation of microglia and support microglia-driven neuronal damage. Conversely, persistent microglial activation may trigger the formation of a microglia-supported cell barrier that segregates and compacts the plaques thus preventing further plaque-induced damage to healthy neurons.

Adaptive nature of tissue-specific macrophage diversity

The macrophage owes its name to their initial discovery by Eli Metchnikoff, who observed the engulfment of a splinter—artificially placed into the star fish tentacle—by large cells that congregated around the foreign object [1]. During the following century, our knowledge of the macrophages’ oeuvre of functions expanded significantly and well beyond their role as “phagocytes” [2]. The myeloid lineage is represented by circulating monocytes and tissue-resident macrophages that fight infectious agents both by using potent innate response mechanisms and by facilitating adaptive immune responses via antigen presentation and cytokine production that recruit other immune cells [2]. Tissue-resident macrophages can also operate as accessory cells supporting highly specialized parenchymal cells, such as hepatocytes, cardiomyocytes, and neurons [2]. This supportive function relies on the ability of macrophages to sense chemically distinct cues derived from neighboring cells and to respond to them in a variety of ways that can range from cell killing to the production of ligands supporting cell survival and specialized functions. Macrophages display a tissue-specific heterogeneity that is exemplified by their polymorphic morphology, and their distinct patterns of cell surface receptor proteins and gene expression [2]. This tissue specific heterogeneity implies either a developmentally predetermined ability of macrophages to operate within a specific tissue setting and/or a remarkably stable macrophage adaptation to a particular environment. The latter mechanism is supported by findings that show the phenotypic plasticity of macrophages when transplanted into a “new” tissue environment [3-5].

The mechanism of macrophage diversity is particularly puzzling in the brain, where resident macrophages—termed microglia by Rio del Hortega [6]—may adjust not only to specific cell types, i.e. neurons or oligodendrocytes, but also to functionally distinct neurons in different brain areas (Fig 1). The progenitor cells of microglia are produced in the yolk sac and arrive in the brain alongside the developing blood vessels during early embryonic development [7-9]. The microglia population subsequently expands during brain development and maintains itself as a self-renewing population throughout life [10,11]. As the brain develops and many of the newly generated neurons that fail to establish connection to other neurons die, microglia scavenge dying cells en mass [12]. Microglia further contribute to the robustness of neuronal networks by pruning weak/afunctional synapses [13-15]. These clearing functions of microglia are thought to be essential for brain development as ablation of microglia leads to impaired brain formation and premature death [16-18].

Figure 1: Brain region specific microglia specification.

Microglia display region specific differences in morphology, density, and gene expression patterns. Schematic shows the mouse brain with digitized immunofluorescence images of IBA1+ microglia from the cortex (yellow), striatum (gray), hippocampus (green), cerebellum (red), and brain stem (black), percentages indicate the number of microglia as compared to total cell number per corresponding brain region [68,69]. A representative image of an individual cerebellar IBA1+ microglia (green) containing two large lysosomal structures (CD68: red) is shown (Imaris, DAPI: blue) [20]. Heat map shows the differential gene expression pattern of microglia by brain regions [22].

Microglia clearing activities generally subside in the adult brain, except in the cerebellum, where neurons die continuously—albeit at a low rate—throughout life (e.g. ~4% of Purkinje cells are lost per month in mice between 4-12 months [19]) (Fig 1). The physiological attrition of the cerebellar neurons is associated with increased phagocytic activity of microglia that is reflected by increased lysosomal load [20], directed motility [21], and the expression of genes encoding proteins associated with lysosomal/clearance functions [20,22,23] (Fig 1). Similar to the cerebellum, microglia in the subventricular zone, hippocampus, and the olfactory bulb—sites of de novo neurogenesis and associated neuron death—are maintained in a state of heightened clearance activity [12,24]. Collectively, these observations suggest that clearance phenotype of microglia occurs in response to dying neurons. Indeed, independent studies showed that the phagocytic phenotype of microglia can be induced by mere exposure to apoptotic neurons in and ex vivo [20,25].

The clearance potential of microglia becomes particularly apparent during ageing and neurodegenerative diseases. The analyses of gene expression changes in isolated microglia and brain tissue of mice and humans affected by Alzheimer’s disease (AD), Parkinson’s disease, Huntington’s disease, amyotrophic lateral sclerosis, frontotemporal dementia or Pick’s disease, show up-regulation of genes that are linked to microglia clearance activity and inflammation [25-31].

Cellular and extracellular triggers of microglial activation during neurodegeneration.

Dying neurons, however, are not likely to be the only signal that triggers microglia activation during neurodegeneration. Many neurodegenerative diseases are associated with the emergence of extracellular protein aggregates that share common features with foreign bodies [32]. Indeed, the aberrant protein/lipid-rich plaques in neurodegenerative conditions are densely surrounded by microglia (Fig 2), evoking Metchnikoff’s description of macrophages as a scavenger of foreign bodies [1]. Metchnikoff’s original experiments as well as numerous follow-up studies [33] show that the encounter of a macrophage with a foreign object will first lead to the migration and polarization of the macrophage toward the object followed by its engulfment [34]. A failed attempt to phagocytose (and/or an object size > 10 μm in diameter) will trigger macrophage fusion, inflammatory activation, and the secretion of bioactive molecules [34-36] to mitigate the extracellular breakdown the object. In cases when the foreign substance cannot be eradicated, the persistent inflammatory activation of macrophages will lead to the recruitment of other immune and stromal cells [36] and the subsequent isolation and neutralization of the foreign material. Microglia responses to plaques are highly reminiscent of the described foreign body response (FBR) (Fig 2) and involve microglial migration and cell polarization toward the plaques [37,38], microglial adhesion to the plaque surface [39], inflammatory activation [40], and the physical insulation and potential compaction of the plaques (Fig. 2).

Figure 2: Microglial response to amyloid plaques resembles macrophage response to foreign bodies.

(left) Schematic shows microglia (green) responses to amyloid plaques (orange/red), amyloid plaques are characterized by a soft lipid-rich halo (orange) surrounding the rigid core of the plaques (red), which is formed by densely folded amyloid β (Aβ) sheets [49]. (right) Representative images of IBA1+ microglia (IBA1: green) interactions with amyloid plaques (Thioflavin S (Thio S): red) in the 5xfAD mouse model [86] of Alzheimer disease are shown (DAPI+ nuclei: blue). Similar to macrophage FBR, microglia polarize and migrate toward the plaque [37], display ruffling of cell membranes and adhesion to the plaque surface, release enzymes for extracellular degradation [57], produce an inflammatory response [57], and have their processes loosely interdigitated [34-36]. The inability of microglia to clear the plaque may trigger the formation of a microglia-supported barrier leading to the physical isolation/segregation of the plaques [63,64]. Microglia barrier formation has been suggested to prevent plaque growth, facilitate its compaction, and reduce plaque-induced toxicity to healthy neurons [63-65].

There are many components of the plaque that may trigger microglia recruitment and activation (Fig 3), similar to those elicited by foreign bodies. The FBR is triggered by recognition of molecular signals, such as pathogen-associated patterns (PAMPs) and damage-associated patterns (DAMPs) [41-45], as well as mechanical cues [46] (Fig. 3). In addition to localized tissue damage caused by plaque aggregation, amyloid plaques contain proteins, carbohydrates, nucleic acids, lipids, and metal ions, many of which have opsonizing and clearance- or inflammation-inducing activities on macrophages [47,48]. These molecules together with a soft lipid-rich halo plaster the rigid core of the plaques, which is formed by densely folded amyloid β (Aβ) sheets [49]. Physical measurements reveal that the level of stiffness of the plaque's core is similar to the stiffness of the bone or extracellular collagen fibers [49-51], which is more than 10⁶-fold above the stiffness of the brain, one of the softest tissues in the entire body [52]. Amyloid plaques as well as other pathogenic protein aggregates further possess the capacity to produce different aggregation forms with variable surface structures and roughnesses [53]. Based on these characteristics of the plaques, it is plausible that microglia deem the plaques as bona fide foreign objects, not overly dissimilar from the splinter inserted into the tissue of a starfish [1] or any other foreign object placed in the human tissue including the brain [54]. The clustering of microglia around the plaques was originally thought to reflect the process of microglia-mediated phagocytosis of the plaques. However, multiple recent studies assessing microglia-mediated plaque clearance suggest that microglia, while engulfing oligomeric Aβ [55-57], are rather inefficient/unable to remove fibrillar Aβ deposits [58,59] and conversely, may even contribute to dense-core plaque formation [60-62]. Reminiscent of FBR in other well described scenarios [34-36], the inability of microglia to clear the plaques may trigger the formation of a microglia-supported barrier around the plaques [63,64] (Fig. 2). Microglia barrier formation can lead to the physical isolation/segregation of the plaques and by preventing plaque growth and facilitating its compaction may reduce plaque-induced toxicity to healthy neurons [63-65].

Figure 3: Cellular and extracellular triggers of microglial activation during neurodegeneration.

Schematic displays the variety of molecular signals that can be involved in mediating FBR including activation of pathogen-associated pattern recognition receptors, such as toll-like receptors and scavenger receptors, to initiate inflammatory or phagocytic response [44,45]. The same receptors can also recognize damage-associated molecular patterns, that are released or exposed upon host danger, such as HMGB1 (released upon necrotic cell death) and fibrinogen (exposed upon disruption of extracellular matrix [44,45]. Another potent signal representing cellular damage is ATP, which is recognized by purinergic receptors expressed on microglia [45]. Identification of foreign particles can be mediated by the presence of “eat-me” or the lack of “don’t-eat-me signals”. These are pairs or macrophage-expressed receptors that bind their cognate ligands [42]. “Don’t-eat-me” signals can be mediated e.g. by the expression of CD47 on functional neurons and synapses, which by binding to SIRPα on myeloid cells triggers inhibition of (unwanted) phagocytosis. Conversely, opsonization of pathogen and foreign bodies by the complement system [43] can mediate strong “eat-me” signals via the activation of Fc receptors, such as FcγR3. A more recently characterized family of receptors are called triggering receptor expressed on myeloid cells (TREMs) which can recognize anionic lipids and ApoE found on plaques [41]. Given the chemically complex composition of amyloid plaques that can contain a variety of lipids, nucleic acids, carbohydrates, and misfolded proteins [47,48], it is likely that many of these receptors are simultaneously activated to drive a microglia response. In addition to chemical cues, mechanical cues, such as stiffness and surface roughness, are also potent triggers of FBR [46] and can directly affect the level of inflammatory response elicited [36]. Mechanical cues can be detected among others by specific mechanosensors expressed on microglia [46,76].

Mechanosensing by microglia

The stark disparity between the physical features of the plaque and the brain tissue suggests that mechanosensing could be a potential integral part of plaque-driven microglia activation. Microglia, similar to other cells, have the ability to sense changes in mechanical forces and can adjust their morphology according to the stiffness of their environment [66]. Microglial mechanosensing may allow microglia to discriminate between cellular densities and the associated tissue stiffness [67] and may contribute to microglia region-specific morphological and functional specification in the different brain regions [68,69] in the healthy and diseased brain (Fig. 1). If given a choice, microglia in vitro and in vivo display a preference toward stiff materials [66]. This preference, which is called durotaxis, could explain the recruitment of microglial to the amyloid plaques. Moreover, the degree of environmental stiffness is directly correlated with the inflammatory activity of macrophages [70], microglia and astrocytes [71].

While the underlying molecular mechanism for mechanically induced macrophage/microglia activation remains to be determined, various studies have implicated molecules regulating adhesion signaling, cytoskeleton signaling [72], and podosomes [73]. Primary mediators of cellular mechanosensing, however, are mechanically-gated channels that are located in lipid bilayers of cell membranes and that can undergo conformational changes upon shifts in the direction of the mechanical forces applied on the bilayer [46]. Even though the presence of mechanically-activated sensors were first described in 1979 in vertebrate hair cells [74], their identity was not known until the discovery of the Piezo channels in 2010 [75]. Today, multiple families of mechanically-gated ion channels are identified, including two-pore-domain potassium channels TREKs and TRAAK [76], transient receptor potential (TRP) channels [76], and degenerin/epithelial sodium channels [76], in addition to Piezo channels [75]. One of the mechanically-gated channels that has been shown to directly regulate innate immune functions is TRPV4, which can tune inflammatory response of macrophages to the stiffness of their microenvironment [77]. Moreover, two recent studies implicated mechanical sensing through PIEZO1 in both peripheral innate immune functions and oligodendrocyte progenitor cell (OPC) differentiation [78,79]. In one study, the authors use an in vitro model to mimic the pressure changes in the lung and show that mechanical forces acting on monocytes drive a calcium influx via PIEZO1 that leads to an HIF1α-mediated inflammatory response [79]. The activation of PIEZO1 dictates the level of systemic immune response both in the context of pathogens and autoinflammation [79]. The second study shows that aging leads to the stiffening of the OPC microenvironment, which impairs the proliferation and differentiation rates of OPCs through mechanosensing by PIEZO1 [78].

Based on their distinct physical properties, it is plausible that amyloid plaques present a mechanical stimulus to microglia that is sensed through mechanically-gated ion channels or yet to be identified mechanosensing properties of other immunoreceptors [80]. Among the well-studied mechanically-gated ion channels, PIEZO1 in the brain is highest expressed in endothelial cells and microglia [81] and displays increased expression in astrocytes around the plaques in a rat model of AD [82]. Even though the function of PIEZO1 in plaque-associated microglia remains unknown, it is an exciting candidate for microglial mechanosensing of dense-core plaques. Notably, AD patients display a significant dysregulation of the unsaturated fatty acid metabolism [83], which may have a direct impact on lipid membrane composition and fluidity, and subsequently the sensitivity of mechanically-gated channels. Recent data show that lateral membrane tension activates PIEZO1 [84]. These findings suggests that the composition of lipid bilayers—which directly affects membrane fluidity [85]—may play a role in regulating PIEZO1-mediated mechanosensation. Accordingly, a recent study showed that saturated lipids, which decrease membrane fluidity, lead to inhibition of PIEZO1 function, which can be rescued by an increase in unsaturated fatty acids [85]. Therefore, one of the direct effects of lipid imbalance during AD could be alterations in microglial mechanosensing.

The mechanosensing by macrophages involves the production of the peptide ligand EDN1 that mediates the PIEZO1-driven impact on HIF1α stability and supports inflammatory gene expression [79]. Provided that similar mechanisms can operate in the brain, the pharmacological manipulation of the ligand/receptor pair induced by mechanical stress may prevent microglial inflammatory responses and attenuate neuronal degeneration. The next exciting steps will be to investigate the specific extrinsic signals, receptors, and downstream signaling pathways that govern microglial mechanosensing. Elucidating the regulatory mechanisms underlying microglial mechanosensing and its contribution to microglia activation states will expand our understanding of how microglia contribute to neurodegeneration and has the potential to reveal novel therapeutic avenues in the treatment of neurodegenerative diseases.

Highlights:

• Microglia display diverse and brain region-specific phenotypes.

• Microglial response to protein aggregates/plaques during Alzheimer disease resembles macrophage response to foreign bodies.

• Microglial mechanosensing of the physical properties of a plaque as a potential driver of microglial activation.