The Central Nervous System (CNS) contains a diverse array of cell types, each playing an important role throughout development. This includes astrocytes, oligodendrocytes, microglia, and ependymal cells as well as many subtypes within each group. Myeloid cells in the CNS are of particular interest due to their role in the pathology of many diseases as well and their utility as a biomarker.
The myeloid cells within the CNS are heterogeneously-distributed and include perivascular, choroid plexus, and meningeal macrophages as well as microglia which are the most abundant. Historically, non-parenchymal CNS macrophages were thought to be derived from bone marrow and undergo steady replacement by short-living blood monocytes. However, current research has shown that microglia has the same prenatal origin as leptomeningeal, perivascular-space, and choroid-plexus macrophages. Additionally, rather than receiving steady contributions of cells from bone marrow, they are long-lived cells that undergo self-renewal.1
While immunity in the CNS is generally associated with neuron-associated microglia in the parenchyma, several other specialized macrophage populations are involved at CNS interfaces with largely uncharted roles in health and disease.3 Other cell types localized to CNS interfaces include dendritic cells responsive to FLT3 and leptomeningeal mesothelial cells. In conjunction with pericytes and astrocytic foot processes, these cells contribute to the immunological blood-brain-barrier.2
Microglia are specialized macrophage-like cells found in the CNS capable of orchestrating a potent inflammatory response. They are involved in a variety of synaptic and neuronal activities including synaptic organization, apoptotic cell phagocytosis in the developing brain, and brain protection and repair. Activation of microglial cells is dependent on disease context and a variety of other factors including aging and environment.2
An area of particular interest that has not received much attention in the past is the development of microglia in an embryo. This period is marked by neurogenesis, neural migration, and synaptogenesis. Microglia can be seen as early as the 4th week of gestation. Their early development is of note as they are thought to contribute to the development of various portions of the brain as well as overall homeostasis. In the reverse direction, microglial phenotypes are also heavily influenced by the brain microenvironment.1
Microglia have been found to be involved in almost all brain diseases, including neurodevelopmental, neuroinflammatory, neurodegenerative, neurooncological, and neuropsychiatric disorders.2 Among them, neurodevelopmental disorders (NDDs) are a group of disorders that cause abnormal development of the brain, including autism spectrum disorder, attention deficit hyperactivity disorder (ADHD), intellectual disabilities (IDs), etc. Children with NDD may experience delay in brain function development that controls motor, social and sensory processes.4 Genes associated with specific NDDs including autism, IDs, and schizophrenia are found to express more in microglia compared to genes in the whole cortex.1 Within these brain disease progressions, various characteristic macrophage functions are seen including phagocytosis, antigen presentation, and the production of immunomodulatory factors. Because of this, microglia are an attractive target for modulation in CNS diseases.2
Microglia continuously communicate with various cell types including neurons, astrocytes, oligodendrocytes, and other microglia in order to accurately assess conditions in the CNS and respond accordingly. Ligands, their corresponding receptors, and respective downstream pathways that regulate microglial function have been studied to various degrees. These functions are diverse and include immune response, maintenance of biochemical homeostasis, scavenging cellular debris and dead neurons, and neuronal circuit maturation throughout development.
CD200 is a transmembrane glycoprotein which plays a neuroprotective role and is primarily found on neurons, astrocytes, and oligodendrocytes but also epithelial cells, endothelial cells, fibroblasts, and lymphoid cells. The target receptor (CD200R), exclusively found on microglia and other macrophages, downregulates the expression of pro-inflammatory molecules such as TNF-alpha, various interferons, and iNOS.
Fractalkine (CX3CL1), a chemokine constitutively expressed in neurons that targets receptors on microglia, is also involved in neuronal protection and keeps microglia in a surveying phenotype. Activation of the CX3CR1 receptor in microglia trigger several intracellular secondary messengers such as P13K, AKT, and NF-κB which in turn regulate apoptotic, proliferative, transcriptional, and migratory functions in microglia. Fractalkine is implicated in embryonic and postnatal maturation of the CNS, promoting recruitment of microglia to neuronal circuits during synaptic remodeling. Microglia in turn regulate neuronal function via the release of trophic factors.
One of the key functions of neuron to microglia fractalkine signaling occurs is synaptic development during the fetal and postnatal period. Fractalkine has been implicated in microglial pruning of dendritic spines, remodeling of neuronal circuits, and recruitment of microglia to develop thalamocortical synapses. Microglia to neuron signaling then occurs. For instance, BDNF released by microglia modulates spine density and expression of AMPA and NMDA in cortical neurons.
Numerous other chemokines are also involved in neuron to microglia and astrocyte to microglia paracrine signaling as well as autocrine microglial signaling. CCL2/MCP-1 and CXCL12/SDF-1 are released by astrocytes and neurons in neuroinflammatory conditions, while CCL21 (targeting CXCR3 on microglia) is expressed exclusively by endangered neurons following injury. Overall, chemokines and other molecules play a key role in the bidirectional signaling between neurons and microglia. They modulate inflammatory responses, regulate cell survival, exert neuroprotection, and affect axonal sprouting and synaptic plasticity.
The utility of computational tools has grown tremendously over the last 2 decades. Older methods utilizing data from large-scale assays were mostly limited to static models. Models investigating signal transduction, co-expression networks, and various correlations were utilized but their utility was limited. A common drawback was the type of data used to generate the model. Additionally, intermediate signaling molecules were almost exclusively omitted. These networks were mostly based on protein-protein interaction data and microarray expression data.10
Even so, the results of these studies were significant. In one study, a library of protein-protein interactions was generated via a large-scale yeast two-hybrid screen which found 957 putative interactions.10 These interactions included proteins within the same biological function as well as proteins within different biological functions. Interactions linking the biological functions together give larger cellular processes.11
One older computational model still in use today is WGCNA, an R package for weighted correlation network analysis.12 Correlation networks are commonly used for bioinformatics applications. One such manifestation is a weighted gene co-expression network analysis, a systems biology method for describing correlation patterns among genes. Weighted correlation network analysis (WGCNA) can be used to find clusters of highly correlated genes, summarizing such clusters, and relating them to each other. Correlation networks find utility in identifying candidate biomarkers or therapeutic targets. Portions of the correlation network methodology described in the WGCNA paper had been described previously but WGCNA provides a consistent and comprehensive software implementation of such a correlation network.12
More recently, a computational tool used to study intercellular communication and predict ligand-target interaction has emerged. Known as NicheNet, this tool is particularly useful for modeling gene expression of a cell influenced by interacting cells. NicheNet works with single-cell gene or bulk expression data and combines it with prior knowledge of cell signaling and gene regulatory networks collected from multiple existing databases.13 NicheNet is then able to predict ligand-receptor interactions with potential to drive gene expression changes in cells of interest.