Microglia consist of at least two subpopulations that coexist in the adult CNS and derive from different sources: one that originates from bone marrow-derived cells and migrates to the CNS during embryonic development for colonization of the nervous system parenchyma, the second that develops from myeloid progenitor cells and enters the brain after birth
, becoming fundamental for microglial maintenance of the CNS homeostasis
[6, 33]. Since adult microglia can play a twofold role, either amplifying the effects of inflammation and mediating cell degeneration, or protecting the nervous system from pathological insults
, efficient chemotaxis can acquire either neuroprotective or inflammatory and detrimental roles. Whereas we are currently capable of distinguishing amoeboid-phagocytic from branched-surveilling microglia
[3, 5], we are unaware of how these morphological shapes can denote a neuroprotective or neurotoxic role. In other words, it is still an open matter how the vast morphological heterogeneity existing within the mixed microglia population might relate to functional diversification.
Cell behavior strictly depends on the stiffness and shape of the microenvironment, and the response to surface topography is a very critical determinant of cell morphogenesis. Microglia respond to micro- nano-structured pits, protrusions and grooves with altered morphology, adhesion, and directional growth
. In our work, we establish that primary microglia retain in vitro the intrinsic competence of modifying their structure in response to contact guidance cues. When grown on pillar-shaped substrates, microglia acquire a center-stage multi-polar morphology and develop abundant button-like pseudopodia emerging from the cell body with a nearly radial orientation. This situation might very well depict the in vivo condition of microglia not committed to migrate, but exploring the environment with short forward, backward and sideways steps, with the final purpose to orient their migration. When placed on a line-grating substrate in the presence of parallel and symmetrical grooves, microglia elicit a longitudinally flattened appearance with elongated bipolar shape. This in vitro condition might favor a polarized extension of filopodia at the leading edge of the cell, in preparation of a forward translocation of the cell body, with retraction of the rear of the cell
. The adjustment of microglia to a line-grating geometry is thus highly suggestive of commitment to follow a straight path in vivo. Thus, surface topography can induce a synchronous and homogeneous metamorphosis of microglia in vitro, distinguishing the center-stage multipolar (on micropatterned pillars) from the elongated bipolar (on line-grating micropatterns) cells. With the use of specific markers and cytoskeleton-perturbing drugs, future work will aim to establish if these subpopulations can be correlated to: a) microglia generated from bone marrow rather than from myeloid progenitor cells; b) microglia colonizing the CNS during development, rather than maintaining CNS homeostasis in adulthood; c) more importantly, “beneficial” rather than “detrimental” microglia. Moreover, future analysis of parameters such as cell adhesion, distribution of actin cytoskeleton and microtubules, phagocytosis, antigen presenting power, beneficial or detrimental chemokine/cytokine expression and release, will allow to determine the functional identity of multipolar versus bipolar microglia.
Moreover, by the use of bio-mimetic 3D microfluidic substrates, our work demonstrates that primary microglia appear intrinsically different from immortalized N9 cells. During cell polarization inside the microchannels, primary microglia tend to align longitudinally and individually, whereas immortalized N9 microglia assume a more compact morphology with the propensity to establish cell contacts. This might be reminiscent of the immortalization program that likely confers to N9 cells a compact morphology optimizing cell division and contact formation, rather than the aptitude of primary cells to explore the environment with a direction-oriented polarization. In both cases, specific receptors and structural molecules are found enriched respectively at the cell-to-cell surface (N9 cells) or leading edges (primary microglia), as evinced by increased phalloidin and P2Y12 receptor signals at these sites.
Cell adhesion and morphology are dynamically mutable during cell migration
. The last issue that we addressed in our work is thus the free motility of microglia in culture. While numerous publications cite motility and short migration of microglia in vivo, only few works describe microglia moving for long distances. This is the case for example of the work by Carbonell and coauthors
, demonstrating by intracerebroventricular injection of rhodamine and time-lapse confocal microscopy that subventricular microglia at the interface of the cerebrospinal fluid and brain parenchyma, exhibit the “in situ” ability to migrate for several hundred microns into the parenchyma, towards a deafferentation injury of the hippocampus. Conversely, Nimmerjahn and colleagues indicate only static movement of microglia, branch motility, but not cell bodies migration
. Using transcranial two photon microscopy, Davalos and coauthors
 confirm that only microglial processes are highly dynamic in intact brain, reaching up to several micrometers in length, or retracting until they completely disappear. This baseline dynamism of microglial processes is thus in sharp contrast to the stability of the microglial cell bodies and surrounding neuronal processes
In vitro studies on this subject have instead shown only transwell device infiltration directed by a chemical gradient. Lee and Chung
 describe that ADP stimulates chemotaxis of immortalized BV2 microglia from the upper to the lower side of the transwell membrane. Karlstetter and co-authors
 present evidence that curcumin inhibits basal and LPS-induced relocation of BV2 microglia from the upper to the lower transwell membrane surface
. However this passage across a physical barrier limits the possibility of investigating long distance migrations and, moreover, provides only indirect evaluation of kinetic parameters. Honda and collaborators
 using the Dunn chemotaxis chambers report that primary rat microglia have weak motility in the absence of ligands, but perform a mean displacement of about 50 μm in the presence of ATP or ADP. Nasu-Tada et al.
 confirm that primary microglia are almost static in the absence of stimulants, but show chemotactic responsiveness to ADP with a maximal displacement of about 140 μm. Finally, Haynes and coauthors
 prove that the leading edge of mouse primary microglia moves for a maximal displacement of 50 and 120 μm after 30 minutes, in a gradient of ATP or ADP, respectively. However, the average distance spontaneously migrated in the absence of any provided stimulus is only 0.8 μm.
All considered, spontaneous long distance migration of microglia is described in vivo with conflicting results, but never in vitro until now. Here, we demonstrate that biocompatible polimeric channels constitute a more suited environment than transwell devices or Dunn chemotaxis chambers for permitting and measuring spontaneous long distance migration in vitro. Indeed, the 500 μm traveled by primary microglia inside the microfluidic channels in the absence of stimuli are much above the known average distance so far reported and this achievement might be perhaps improved.
In summary, by the use of interdisciplinary techniques, our data indicate that microfluidic technologies and microstructured patterns with control over the presentation of adhesion sites, are able to sort out different morphological structures within a mixed microglia population and allow free motility in vitro. By reproducing a specialized niche for the cells, these technologies can help to understand the role of structural determinants in priming morphogenesis and free motility of microglia and may be exploited for translational research on functional tissue engineering and implantable device design.