Page 77 - Read Online

Benusa et al. Neuroimmunol Neuroinflammation 2020;7:248-63 I http://dx.doi.org/10.20517/2347-8659.2020.03 Page 255

myelin peptide (myelin oligodendrocyte glycoprotein peptide 35-55) accompanied by pertussis toxin and
an adjuvant to ignite an inflammatory response [117-119] , which transfers to the brain and results in chronic
neuroinflammation persisting for months. While neuroinflammation is present after induction and
[79]
throughout the EAE course, clinical symptoms do not begin to appear until ~15 days post-induction .
Mice exhibit a range of clinical symptoms from limp tail and loss of righting reflex (EAE 1 & 2, respectively)
[79]
to single- or double- hind limb paralysis (EAE 3 & 4, respectively) . AIS pathology begins to appear
at an early timepoint after clinical onset (~18 days post-induction), but only in mice that display more
severe clinical symptoms (EAE Early 3 & 4). Increased AIS pathology is observed with disease severity
[79]
and progression (EAE Late 1 & 2, 3 & 4, ~25 days post-induction) . In contrast, the LPS model is an
acute neuroinflammatory model induced by a single peripheral injection of LPS [120,121] . This results in
widespread peripheral inflammation that rapidly transfers to the brain (~3 h), but the neuroinflammation
is resolved by 2 weeks post-injection. In the LPS model, AIS pathology was present from as early as 24 h
post-injection and persisted until 1 week post-injection, coincident with the initiation and resolution of the
[80]
acute neuroinflammatory environment . In contrast to the immune-mediated neuroinflammatory models,
[79]
the cuprizone model is a demyelinating model where a copper-chelating toxin, cuprizone, is administered
through chow resulting in oligodendrocyte cell death and, consequently, loss of myelin [119] . Demyelination
is detectable 1-2 weeks after cuprizone treatment with peak demyelination occurring by 5-6 weeks of
exposure [122-124] . The cuprizone model yields substantial cell death and demyelination resulting in microglial
[79]
recruitment and neuroinflammation but no AIS pathology was observed .
We utilized these three models to further investigate microglial heterogeneity. AIS disruption only occurred
in the LPS and EAE models, while microglial-AIS contact was abundant in all three models. Thus, while
microglial reactivity and contact increased prior to and was coincident with disruption in EAE, contact
alone did not disrupt AIS integrity [79,80] . Therefore, we analyzed the inflammatory expression profiles
of cortical microglia across all three models to assess how microglial reactivity differentially influences
neuronal integrity. Our goal was to assess microglia expression profiles early in the disease process
to identify inflammatory changes that drive disease progression and are not consequential of disease
progression. Thus, cortical microglia were isolated from mice induced with EAE, Cuprizone, or LPS at time
[79]
points where neuronal pathology is detectable but had not peaked (EAE Early 3 & 4 , 3 week Cuprizone [123] ,
+
[80]
LPS 24 h) . Briefly, total RNA, collected from cluster of differentiation (CD) 11b cells isolated from
the cortex of c57black6 female mice, was submitted for NanoString mRNA expression analysis. (Further
details on model generation, cell isolation and NanoString analyses are provided in Supplementary
Materials [79,80,120-123,125-133] ). Cells were collected at time points in each model that corresponded to the early
presence of neuronal/myelin pathology, but prior to peak disease course in an effort to understand the
inflammatory profiles that drive pathogenesis [79,80] .

Microglia with reactive morphologies predominate in the cortex of all three models [79,80] [Figure 2A],
which is consistent with these cells presenting with a pro-inflammatory phenotype. However, based
on NanoString expression analysis of 248 inflammation-associated genes, microglia from all three
neuroinflammatory models displayed distinct regulation of inflammatory genes [Figure 2B], underscoring
the heterogeneity of morphologically similar cells. Of 248 analyzed genes, 95 were significantly upregulated
(1.3 fold-change or greater) [Figure 2C] and 175 were significantly downregulated (at least 1.3 fold-change)
among the three neuroinflammatory models when compared to microglia from naïve mice [Figure 2D]. 27
of 95 (28.4%) upregulated genes [Figure 2C] and 50 of 175 (28.6%) downregulated genes [Figure 2D] were
similarly changed across all three models but model-specific differences were observed for both categories.
Numerous genes [Figure 2C] associated with a pro-inflammatory (“M1”) phenotype (such as interferon
regulatory factor 1, lymphotoxin beta, C-C chemokine receptor type 7, C-C motif chemokine ligand 7,
C-C motif chemokine ligand 17, lymphotoxin Alpha, Il1a, signal transducer and activator of transcription
2, and tumor necrosis factor super family 14) were upregulated uniquely in EAE Early 3 & 4 and LPS 24 h