Retinal dendritic cell recruitment, but not function, was inhibited in MyD88 and TRIF deficient mice
Journal of Neuroinflammation
Retinal dendritic cell recruitment, but not function, was inhibited in MyD88 and TRIF deficient mice
Neal D Heuss 0
Mark J Pierson 0
Kim Ramil C Montaniel 0
Scott W McPherson 0
Ute Lehmann 0
Stacy A Hussong 0
Deborah A Ferrington 0
Walter C Low 1
Dale S Gregerson 0
0 Department of Ophthalmology & Visual Neurosciences, University of Minnesota , Lions Research Bldg. Rm 314, 2001 6th St SE, Minneapolis, MN 55455 , USA
1 Department of Neurosurgery, University of Minnesota , Minneapolis, MN 55455 , USA
Background: Immune system cells are known to affect loss of neurons due to injury or disease. Recruitment of immune cells following retinal/CNS injury has been shown to affect the health and survival of neurons in several models. We detected close, physical contact between dendritic cells and retinal ganglion cells following an optic nerve crush, and sought to understand the underlying mechanisms. Methods: CD11c-DTR/GFP mice producing a chimeric protein of diphtheria toxin receptor (DTR) and GFP from a transgenic CD11c promoter were used in conjunction with mice deficient in MyD88 and/or TRIF. Retinal ganglion cell injury was induced by an optic nerve crush, and the resulting interactions of the GFPhi cells and retinal ganglion cells were examined. Results: Recruitment of GFPhi dendritic cells to the retina was significantly compromised in MyD88 and TRIF knockout mice. GFPhi dendritic cells played a significant role in clearing fluorescent-labeled retinal ganglion cells post-injury in the CD11c-DTR/GFP mice. In the TRIF and MyD88 deficient mice, the resting level of GFPhi dendritic cells was lower, and their influx was reduced following the optic nerve crush injury. The reduction in GFPhi dendritic cell numbers led to their replacement in the uptake of fluorescent-labeled debris by GFPlo microglia/macrophages. Depletion of GFPhi dendritic cells by treatment with diphtheria toxin also led to their displacement by GFPlo microglia/ macrophages, which then assumed close contact with the injured neurons. Conclusions: The contribution of recruited cells to the injury response was substantial, and regulated by MyD88 and TRIF. However, the presence of these adaptor proteins was not required for interaction with neurons, or the phagocytosis of debris. The data suggested a two-niche model in which resident microglia were maintained at a constant level post-optic nerve crush, while the injury-stimulated recruitment of dendritic cells and macrophages led to their transient appearance in numbers equivalent to or greater than the resident microglia.
Dendritic cells; Injury; Microglia; Retinal ganglion cells; NFB
Neurodegenerative processes adversely affect vision in a
significant portion of the human population, and are
associated with glaucoma, age-related macular
degeneration (AMD), diabetic retinopathy, ischemia, retinopathy
of prematurity and traumatic injuries [1-4]. Irrespective
of the cause of the degenerative process, it is evident
that the immune system can be protective or pathogenic
in neurodegeneration . Studies show that microglia
(MG) are not simply the scavengers of the nervous
system; instead, they appear to play complicated, even
contradictory, roles [6,7]. MG have been associated with
clean-up of dead/dying neurons with minimal
inflammation [8,9], restoration of health to damaged peripheral
nervous system (PNS) neurons , and promotion of
inflammation through secretion of proinflammatory
molecules [11,12]. They also appear to kill neurons
through effector mechanisms that include production of
reactive oxygen species which disrupt axonal transport
, and cytotoxic cytokines including TNF. FasL
expression by MG may contribute to the death of Fas+
Regardless of the injury, MG are rapid responders.
Within minutes of a focal injury to brain or retina,
neighboring MG extended processes to the site [16-18].
At later time points post-injury, other cells participate.
The optic nerve crush (ONC) has been used as a model
for neural injury, giving a discreet injury to a limited
number of neurons, retinal ganglion cells, that
progresses on a well-studied, consistent course [19,20]. By
two to three days post-ONC, we found evidence that
dendritic cells (DC) were recruited to the retina,
increasingly associated with the ganglion cell layer (RGC) and
the nerve fiber layer (NFL) of the retina, and increased
in number for approximately ten days, equaling the MG
in total number per retina . The number of DC
remained elevated for more than two months, gradually
declining in number. Many DC were closely associated
with the axons of RGC post-ONC . DC dominated
this close interaction for at least three weeks post-injury,
raising the question of whether their response was
protective, harmful, or unrelated to survival of RGC.
The well-known function of DC as antigen presenting
cells may be important, consistent with reports that
suggest that adaptive immunity mediated by T cells may be
neuroprotective [22-25]. However, very few lymphocytes
were found after this sterile injury, suggesting that the
activity of these DC may not be limited to antigen
presentation. We present evidence that DC are active
participants in the injury response. Analysis of DC from
MyD88 and TRIF single and double knockout mice
showed that DC in the knockout mice were much less
efficiently recruited to the retina post-injury, but the
smaller number that was present remained fully able to
take up DiI+ debris via phagocytosis of DiI-labeled RGC
and axon debris after an ONC. MG compensated for the
reduced number of DC in MyD88 and/or TRIF deficient
mice by increasing their uptake of DiI+ debris.
Materials and methods
CD11c-DTR/GFP mice (CDG) mice on the B6
background (B6.FVB-Tg(Itgax-DTR/EGFP)57Lan/J), express
a chimeric protein comprised of GFP and the diphtheria
toxin receptor (DTR) under control of the CD11c
promoter . Wild type (wt) C57BL/6J (B6) mice were
obtained from Jackson Laboratory (Bar Harbor, ME, USA).
MyD88/TRIF double knockout mice (MTdko) on the
B6 background were bred from pairs obtained from
Dr. Stephen Jameson, University of Minnesota. The
double knockout mice were backcrossed to B6 mice to
generate the single knockout mice, MyD88-deficient
(Mko) mice and TRIF-deficient (Tko) mice. The single
and double knockout mice were backcrossed to CDG
mice to allow visualization of the GFP reporter for CD11c.
All mice were CD45.2. Transgenic mice were bred in
house. All mice were rd8 negative. Mice were handled in
accordance with the Association for Research in Vision
and Ophthalmology Statement for the Use of Animals in
Ophthalmic and Vision Research and University of
Minnesota Institutional Animal Care and Use Committee
Optic nerve crush
The optic nerve crush (ONC) was performed as
described [21,27], except for use of #2197E DSAEK forceps
(Ambler Surgical Corp., Exton, PA, USA) for the crush
procedure. Briefly, the optic nerve was clamped for three
seconds at a point 1 to 2 mm from the posterior pole of
Injection of fluorescent dye, either Fluorogold (FG) or
di-alkyl-indocarbocyanine (DiI), into the superior
colliculus was done to retrogradely label the RGC.
Manipulations were done in a stereotactic device. A midline
incision was made in the scalp to expose the skull. A
unilateral 1 mm hole was drilled at 3.5 mm from
bregma and +1.2 mm from midline. A 10-l syringe and
non-coring needle attached to a micromanipulator was
inserted to a depth of 2.5 mm from the surface of the
brain. Four percent dye in 1.5 l of saline was injected
over the course of 2 minutes. After slow removal of the
syringe, the scalp was sutured with 4-0 silk. FG was used
to count surviving RGC and was administered after the
ONC four days before retina harvest. The FG diffuses
rapidly to the opposite hemisphere of the brain, so that
the RGC of both retinas become equivalently stained
even if the dye was injected unilaterally in the brain. DiI
was used to label the RGC with red fluorescence for
experiments to detect labeled RGC debris in the
mononuclear cells of the retina by flow cytometry and
fluorescence microscopy. DiI was administered seven to
ten days before the ONC.
Flow cytometry of retinal cells
Mice were euthanized, perfused, and the retinas
removed as described . Retinas were dissociated in
0.5 g/ml Liberase/Blendzyme3 (Roche, Indianapolis, IN,
USA) and 0.01% DNase in Dulbeccos
phosphatebuffered saline (DPBS), stained with indicated
antibodies, and analyzed as described [21,28,29]. Analyses
were based on the examination of all immune cells
collected from one or more retinas, as specified.
Immunostaining of retinal flatmounts
Retinal flatmounts were prepared, stained, and analyzed
as described . Primary antibodies included: rat
antiCD11b to stain myeloid cells (clone M1/70, BD
Biosciences, San Jose, CA, USA) and anti-3-tubulin to stain
RGC and their axons (clone TU-20). Secondary
antibodies (Invitrogen, Life Technologies, Grand Island, NY,
USA) included: Alexa Fluor 594 donkey anti-rat IgG for
anti-CD11b; and Alexa Fluor 405 rabbit anti-mouse IgG
for anti-3 tubulin. Cell nuclei were stained with
4,6diamidino-2-phenylindole (DAPI, Vector Laboratories,
Burlingame, CA, USA).
TUNEL-stained retinal sections
Eyes were enucleated and immediately snap-frozen in
Tragacanth (Sigma, St. Louis, MO, USA). Retinas were
sectioned (12 m) through the optic nerve. Detection of
apoptotic nuclei was accomplished by terminal
deoxynucleotidyl transferase-mediated dUTP nick-end labeling
(TUNEL) using the In Situ Cell Death Detection Kit,
Fluorescein (Roche, Indianapolis, IN, USA). Slides were
cover slipped with VECTASHIELD Mounting Medium
containing DAPI (Vector Laboratories, Burlingame, CA,
USA) to visualize the nuclei.
Retinal morphology measurements
The density of nuclei in the ganglion cell layer (GCL)
was measured on DAPI-stained retinal sections. For each
retinal section, three images were taken on either side of
the optic nerve at 500 m intervals. The length of the
GCL was measured using BIOQUANT NOVA PRIME
6.90.10 (BIOQUANT Image Analysis, Nashville, TN,
USA). The number of DAPI-stained nuclei in the GCL
was counted from these images and normalized to the
length of the retinal section to calculate the nuclei
density (nuclei/m). The same retinal sections were also
stained with TUNEL. The TUNEL+ nuclei in these same
sections were counted and cross-referenced with the
DAPI stained nuclei to ensure presence of a nucleus.
Data, whether expressed numerically or graphically as a
mean, included the standard deviation; standard error
was not used. Data analyses for significance were done
with the InStat3 package from GraphPad Software (San
Diego, CA, USA). Comparisons of three or more data
sets were done with one way analysis of variance
(ANOVA) using the Dunnett multiple comparisons test
with a designated control set. Comparisons of two data
sets were done with a two-tailed, unpaired t-test.
Optic nerve crush injury
The ONC procedure yielded a significant, reproducible
sterile injury to the RGC (Figure 1), providing the
opportunity to study factors that may affect recruitment of
mononuclear cells and their interactions with the injured
RGC. Survival of RGC from wt B6 mice and CDG mice
was similar (Figure 1). The autocount procedure and
results are described in the Additional file 1: Figure S1.
Extrapolating our counts/field to the entire normal C57BL/
6J retina yielded a total of 46,324 1,968 RGC, which
compares well with values from other reports obtained
by counting axons (44,860  and 47,113 ) or
retrograde labeled RGC (49,823 ).
Identification of retinal DC and MG/macrophages
We previously showed that the CD11b+GFPhi cells in
quiescent CDG retina were morphologically similar to
MG . Flow cytometry showed that the CD11b+GFPlo
and CD11b+GFPhi DC in the CD45med region commonly
associated with MG were indistinguishable by several
parameters, including expression of F4/80, whether from
quiescent or post-ONC retinas (Figure 2A). The CD45hi
region also contains a small number of CD11b+ cells,
Figure 1 CDG and B6 retinal ganglion cells (RGC) survive similarly after an optic nerve crush (ONC). RGC survive in greater numbers using
DSAEK forceps for the ONC. *All post-ONC counts of RGC differed from normal controls, as well as the contralateral RGC, P < 0.05. Normal control
RGC counts represent the RGC counts of both retinas after unilateral injection into the superior colliculus. Ipsilateral - manipulated side; contralateral
opposite, unmanipulated side. Counts are average RGC numbers/retinal field where each retina count is the average of eight fields/retina, SD.
N = number of mice. For example, an N of 4 represents 32 fields. Field size = 0.190 mm2.
Figure 2 (See legend on next page.)
(See figure on previous page.)
Figure 2 Retinal GFPhi cells are CD11b+ and express F4/80. The GFPhi cells in the retina of CDG mice are found in the CD45med region
associated with microglia (MG), and in the CD45hi populations. (A), Analysis of F4/80 and GFP expression in retinal CD45med cells from naive B6,
naive CDG, and CDG mice post-ONC. (B), Expression of GFP and F4/80 in CD45hi cells from the same retinas shown in (A). The quadrants are
labeled with the percent of cells contained in each quadrant.
most are F4/80+ (Figure 2B). The influx of GFPhi cells
post-ONC was easily detected by sequentially gating on
CD45+ cells, and then confirming that the GFPhi cells
expressed CD11b. At 7 days post-crush, the GFPhi cells
accounted for approximately 40% of the total CD45+11b+
cells. The CD11b+GFPhi cells do not express detectable
levels of Ly6G (clone 1A8) or Ly6C/G (Gr-1, clone
RB68C5 (data not shown)), showing that neutrophils did not
contribute to the counts. These properties allowed
identifying these cells in four groups (Table 1) for the studies
Retinal DC associate with RGC axons and soma after
In quiescent retina, a small number of ramified GFPhi
DC were distributed in the inner plexiform layer (IPL)
immediately below the soma of the RGC (Figure 3A).
This morphology was unchanged at day 1 post-ONC
(Figure 3A). DC in close association with injured RGC
were first detected three days post-ONC, when a careful
search of three retinas found a few GFPhi DC closely
associated with RGC nerve fibers (Figure 3A). Dendrites
(small blue arrows) show GFPhi DC in the IPL at day 1
and day 3a. Large blue arrows show axons bundled into
nerve fibers (day 3b). Somata (white/blue arrows) are
shown in days 3b and 7a. By five days post-ONC,
numerous close associations between the GFPhi DC and
nerve fibers were found (Figure 3A). Counts of
representative 20X fields from three retinas harvested at day 5
post-ONC showed that 16% of the total GFPhi DC in
retina were closely associated with nerve fibers or soma.
Close association of the GFPhi DC with the NFL peaked
at 7 to 11 days (Figure 3A). We previously showed that
GFPhi DC dominated the interaction of CD11b+ cell
populations with the NFL at seven days post-ONC .
By 11 days post-ONC, the density of 3-tubulin+ fibers
monocyte/macrophage (CD45hi) lo
in the NFL was reduced, but GFPhi DC were still found
in close contact with remaining fibers (Figure 3A). At
five or more days post-ONC, DC were found in contact
with the RGC soma, and these appeared to be engulfing
the RGC (Figure 3B). The close association of the DC
with the RGC and their axons raised questions about
the molecular basis for these interactions that may affect
RGC survival following an ONC.
Roles of DC versus MG/macrophages in clearance of
Candidate phagocytic cells in the retina that may clear
damaged RGC are CD45+CD11b+Ly6G mononuclear
cells. To detect the relative contributions of MG,
recruited macrophages, and recruited GFPhi DC to the
clearance of injured RGC post-ONC, the RGC were
prelabeled with DiI to detect CD45+CD11b+ cells that had
phagocytosed DiI-labeled RGC. The CD11b+ phagocytic
cells in the retina are found in two populations based on
CD45 staining intensity, CD45hi and CD45med (Figure 4A,
Table 1). The CD45hi population of CDG retina contains
CD11b+ cells that include GFPhi DC, GFPlo monocytes,
macrophages, and polymorphonuclear granulocytes (PMN).
The PMN were excluded by staining with clone 1A8,
which identifies Ly6G+ PMN. Although PMN are
recruited onto the retinal surface at early time points
postONC, a PMN in contact with the NFL or RGC was not
observed (data not shown), and they were routinely
excluded from further analysis. The CD45med population
post-ONC includes MG, recruited macrophages and
GFPhi DC (Figure 4A, Table 1).
As shown by others, and in Figure 4B, RGC were
welllabeled seven days after injection of DiI into the superior
colliculus . After seven days DiI labeling, mice were
given an ONC. Seven days post-ONC, DiI+ RGC were
being engulfed by GFPhi DC, as shown in a series of 1-m
optical sections from confocal microscopy (Figure 4B).
Slices 2, 4, 6 and 8 show the association of the DC with
individual RGC somata at 2-m increments. Further
verification that CD11b+ cells would take up DiI-labeled
cellular debris after an ONC was found by flow cytometric
analysis of retinas after seven days DiI labeling followed
by an ONC. The retinas were harvested at day 6
postONC. DiI labeling was found in both the GFPhi DC and
GFPlo MG/macrophages from the CD45med population
(Figure 4C). This gating strategy was used for further
analysis of DiI+ DC and macrophages.
Figure 3 Recruitment, redistribution and interaction of retinal GFPhi DC with retinal ganglion cells (RGC) and their axons following an
optic nerve crush (ONC) injury. (A), GFPhi DC in naive retina, and at 1, 3, 5, 7, and 11 days post-ONC. GFPhi DC, green; isolectin B4-stained blood
vessels, red; anti-3 tubulin stained RGC somata (blue/white arrows), nerve fibers (large blue arrows), and dendrites (small blue arrows). (B), GFPhi
DC surround and engulf RGC somata ten days post-ONC. DAPI, blue; GFPhi DC, green.
Figure 4 Detection of retinal ganglion cell (RGC) phagocytosis by GFPhi and GFPlo CD11b+ cells in CDG retina post-ONC. (A) Depiction
of CD45hi and CD45med gating, elimination of 1A8+ PMN, and selection for CD11b+ cells that are GFPlo or GFPhi. (B) Sequential confocal sections
of GFPhi DC engulfing DiI-labeled RGC seven days post-ONC. DiI-labeled RGC, red; GFPhi DC, green; DAPI-stained nuclei, blue. (C) Detection and
quantitation of DiI-labeled CD45medCD11b+GFPhi DC and GFPlo cells by flow cytometry. Ipsilateral retinas were labeled by injection of DiI into the
superior colliculus. After seven days, mice were given an ONC; retinas were harvested six days later.
The time course of DiI uptake into CD11b+ cells was
determined next. Based on the progression of RGC
apoptosis, estimated by counts of TUNEL+ cells in the
RGC layer (Figure 5A), DiI-labeled retinas were
harvested from 7 to 17 days post-ONC and analyzed by flow
cytometry to detect DiI uptake by phagocytic CD11b+
Figure 5 Phagocytosis of retinal ganglion cells (RGC) following an optic nerve injury (ONC). (A) Time course of RGC apoptosis following
an ONC. (B) Analysis of DiI uptake in GFPhi DC and GFPlo MG/macrophages following an ONC. Retinas were labeled by injection of DiI into the
superior colliculus. After 7 days the mice were given an ONC; retinas were harvested 7, 10, 13, and 17 days later. CD45+ cells were examined by
flow cytometry as shown in Figure 4. All cells in (B) are CD11b+.
cells in CD45med and CD45hi populations (Figure 5B). The
number of DiI+CD11b+ cells peaked at ten days
postONC in both CD45hi and CD45med populations. Although
the total number of CD11b+ cells in the CD45hi
population was much lower than the CD45med population, the
frequency of CD45hiDiI+ cells at day 10 post-ONC was
much higher (14%, 27 of 191 cells), than that found in the
CD45med population (2.5%, 115 of 4,638 cells) (Figure 5B).
Of the total CD45hiDiI+CD11b+ cells, 89% were GFPhi,
whereas only 35% of the CD45medDiI+ cells were GFPhi.
These results showed that mononuclear cells were actively
phagocytosing DiI-labeled RGC, and proportionately more
of the DiI+ cells were CD45hi DC. Since day 10 post-ONC
was the peak of DiI uptake, day 10 was used for
MyD88/TRIF and recruitment of CD11b+ myeloid cells to
Given the active participation of mononuclear cells in
the clearing of RGC debris, we sought evidence for the
contribution of toll-like receptors (TLR) to this response.
A number of recent reports suggested roles for TLR in
the responses of CNS tissue to injury and/or
neurodegeneration [34,35]. The adaptor proteins MyD88 and
TRIF link TLR ligation to NFB activation. Their
deletion has been reported to reduce neural inflammatory
responses to infectious agents [36-39], and has been
associated with an increase in neural deficits or
diminished neurotoxicity . Since the GFP+ DC appeared in
substantial numbers in the retina following injury, were
closely associated with injured neurons, and
phagocytosed RGC and axon debris, we asked if these activities
correlated with TLR-mediated sensing of the local
environment. Mko, Tko, and MTdko mice were backcrossed
to the CDG background to examine the effects of
eliminating these proteins on the influx of GFPhi DC and
GFPlo MG and macrophages in response to the ONC.
Prior to injury, Tko and MTdko mice had reduced
numbers of CD45medGFPhi DC in the retina compared to
control mice (Figure 6, Top). Deficiencies in either
MyD88 or TRIF or both also led to a substantial
reduction in the number of GFPhi DC that were recruited in
response to an ONC (Figure 6, Bottom). The double
knockout mice gave the greatest decline in retinal GFPhi
DC for both basal and post-ONC conditions, suggesting
that the NFB pathway regulated the response of
mononuclear cells to an ONC. The flow cytometry conditions
shown in Figures 4 and 6 were then analyzed for DiI to
generate the data for Figures 7 and 8, which follow.
MyD88/TRIF deficiency diminished uptake of DiI-labeled
RGC debris by GFPhi DC post-ONC
The CD45hi and CD45med populations were examined
for the effects of MyD88 and/or TRIF deficiency on
Figure 6 Recruitment of CD45medGFPhi cells to retina in Tko, Mko, or MTdko mice on the CDG background. Retinas were harvested ten
days after an ONC. Numbers in the quadrants are percents, and are the averages of at least four samples. P-values for differences in the
populations are shown.
Figure 7 DiI uptake by CD45hi cells. (A) Flow cytometric analysis of DiI uptake by GFPhi and GFPlo cells in the CD45hi population of retina from the
indicated strains harvested after retrograde labeling of RGC with DiI. The retina was DiI labeled by injection of DiI into the superior colliculus seven days
prior to the ONC. Retina was harvested ten days post-ONC where indicated. (B) Summary of cell counts and P-values in mice receiving DiI and an ONC.
clearance of DiI-labeled RGC debris. The mononuclear
cells in the CD45hi population became labeled with DiI
after an ONC, regardless of the presence or absence of
MyD88 or TRIF (Figure 7); the differences due to
MyD88 and TRIF activity were found in the number of
cells that were recruited by the injury. In control CDG
mice, 39% of the GFPhi DC were DiI labeled (yellow versus
green bars), compared to 6% of CD11b+GFPlo cells that
were found to be DiI+ (red versus gray bars) (Figure 7B).
Although the recruitment of GFPhi DC into the CD45hi
Figure 8 DiI uptake by CD45med cells. (A) Flow cytometric analysis of the uptake of DiI by GFPhi and GFPlo cells in the CD45med population of
retina from the indicated strains harvested after retrograde labeling of RGC with DiI via injection of DiI into the superior colliculus seven days prior to
the ONC. Retina was harvested ten days post-ONC. Application of the DiI label, and performance of the ONC were as indicated on the axis. (B) Summary
of cell counts and P-values.
population declined in the MTdko mice, the fraction of
DiI+GFPhi cells relative to GFP+ cells did not change
significantly (Figure 7B). The fraction of DiI+GFPlo cells
ranged from 6 to 10%, and was not significantly different.
Although the total of CD45hiGFPhi cells declined
substantially (from 82 to 23), the total number of DiI+ cells was
not different (Figure 7B); the GFPlo cells appeared to
compensate for the smaller number of GFPhi DC in the uptake
of DiI-labeled debris. Clearly, the DiI+GFPhi DC and the
DiI+GFPlo cells responded differently to the deficiencies in
MyD88 and/or TRIF.
As shown above, CD11b+DiI+ cells were also found in
the CD45med population, but their frequency was much
lower than seen in the CD45hi populations (Figure 8A).
The effect of MyD88/TRIF deficiency differed, relative to
the results seen above for CD45hi cells. For example, the
number of DiI+ cells, relative to the total number of
CD11b+ cells, was constant in control versus MTdko
mice (1:46 in CDG retina versus 1:47 in CDG x MTdko
retina) (Figure 8B). Although the total number of cells in
the CD45med population of CDG retina was
approximately 27-fold higher than in the CD45hi population, the
total number of CD45medDiI+ cells was only 4-fold
higher. A similar comparison of these populations in
MTdko retina showed an 11-fold difference in total cell
number and a 2.5-fold difference in DiI+ cells.
The basis for maintaining the DiI uptake in the face of
a substantial decrease in GFPhi DC is the significant shift
in the distribution of DiI+ cells. Their frequency in GFPhi
cells from CDG x MTdko retina was higher than in
GFPhi cells from CDG retina (0.015 in CDG retina versus
0.068 in CDG x MTdko retina; P < 0.05). Conversely, the
proportion of DiI+GFPlo cells was not different in the
CDG versus CDG x MTdko retina (0.020 versus 0.033,
NS) (Figure 8B). Since the number of GFPhi DC/retina
substantially declined in the MTdko mice (P < 0.001), but
the DiI+ cells did not, showed that uptake of debris was
not significantly compromised by MyD88 and/or TRIF
deficiency. The effect of MyD88/TRIF deficiency was more
closely associated with recruitment of GFPhi cells. The
GFPhi DC clear debris with similar efficiency, even though
their numbers declined.
Depletion of GFPhi cells from retina by systemic DTx
reveals their replacement with GFPlo cells
The results above revealed an altered response to injured
RGC following an ONC in MyD88 and/or TRIF
deficient mice. The GFPhi DC population was most actively
associated with injured RGC, and most affected by
manipulations of MyD88 and TRIF. Since they are DTx
sensitive, DTx treatment of the mice allowed studies of
the participation of DC in RGC degeneration post-ONC,
within the context of MyD88 and TRIF deficiencies.
Several DTx treatment protocols were examined to confirm
the DTx sensitivity of retinal DC, and devise DTx
treatment protocols that would allow comparison of the
ONC injury response in the presence or absence of
To explore differences between GFPhi DC versus GFPlo
macrophages with respect to the close association with
RGC axons in the NFL post-ONC, systemic injections of
DTx were examined. The ip (intraperitoneal) route was
chosen to facilitate in vivo imaging, since the optical
properties of the cornea would be preserved, and to establish
that DTx readily crossed the blood/retinal barrier. Serial
fundus photographs showed that DTx treatment via ip
injection effectively depleted GFPhi DC recruited to the
retina by an ONC (Figure 9A-C). Panel A showed an
elevated number of GFPhi DC at 12 days post-ONC. Panel
B showed noticeable depletion in the same retina one day
after ip injection of 200 ng DTx. Panel C revealed that two
days post-DTx injection gave near-total depletion of GFPhi
cells. Six serial ip injections of 10 ng DTx, or two ip
injections of 100 ng, substantially depleted the GFPhi DC from
uninjured versus crush-injured retinas, respectively, based
on flow cytometry (Figure 9D). Importantly, both showed
retention of GFPloCD11b+ cells representing the
MG/macrophages. Direct counts of GFPhi cells in retinal flatmounts
confirmed their depletion (Figure 9E). However, repeated
systemic use of DTx for more than one week, even at low
doses, was toxic. Accordingly, the short-term, high-dose
protocol was used to explore the observation above that
the MG replaced the DC interaction with RGC and axons,
physically and functionally, in the MTdko mice.
Analysis of DTx depletion of GFPhi DC by confocal
microscopy of retinal wholemounts of ONC-injured
CDG retinas given saline or 200 ng DTx ip showed
depletion of the GFPhi DC from the nerve fibers of the
RGC (Figure 9F-K). In the saline-treated control mouse,
multiple GFPhi DC in close contact with the nerve fibers
were found, and only a single GFPlo cell can be seen
(Figure 9I). GFPhi DC were prominent in the underlying
IPL (Figure 9J and K). We previously showed that at
18 hours post-DTx, the retinal GFPhi DC were depleted
and the axons showed no association with GFPhi cells or
with GFPloCD11b+ MG/macrophages. Instead, the MG/
macrophages were seen intact, in the same field, below
the RGC/NFL . We show here that extending the
time after DTx treatment to 48 hours before harvest
revealed replacement of the GFPhi DC by the GFPlo
macrophages/MG on the nerve fibers (Figure 9F), and
continued presence of GFPlo MG/macrophages in the
underlying IPL (Figure 9G and H).
A two-niche model
Taken together, the data supports a two-niche model for
resident and recruited cells in retina (Figure 10). Niche 1
(N1) is occupied by MG/macrophages. In the resting
Figure 9 Systemic administration of DTx to the CD11c-DTR/GFP mice leads to depletion of the GFPhi DC. (A-C) Serial fundus photographs
of the same retina following an ONC and DTx treatment via ip injection. (A) GFPhi DC at 12 days post-ONC. (B) Fundus photograph of retina one
day after ip injection of 200 ng DTx. (C) Photograph of retina two days post-ip DTx injection. (D) Flow cytometry of retina following ip injection
of DTx under two different protocols; six daily injections of 10 ng into a naive mouse, and 2 injections of 100 ng into a mouse at days 11 and 12
post-ONC. (E) Daily serial injections of 10 ng DTx in naive mice. (F-K) Confocal microscopy of retinal flatmounts in ONC-injured retinas. (F-H) DTx
(200 ng) was injected ip 48 hours prior to harvest, at 15 days post-ONC. (F) Cells and nerve fibers in the NFL post-DTx. (G-H) Cells at two levels in
the IPL. All images in each vertical column are from the same confocal field at different depths into the retina. (I-K) Saline, 1 l, was injected ip
48 hours prior to harvest, at 15 days post-ONC. (I) Cells and nerve fibers in the NFL. (J-K) Cells at two levels in the IPL. GFPloCD11b+ cells stained
red; GFPhiCD11b+ cells stained yellow-green. NFL, nerve fiber layer; IPL, inner plexiform layer at two depths is shown, immediately below the
retinal ganglion cell layer (RGC) soma, and adjacent to the INL.
retina, this population far outweighs the GFP+ DC in
niche 2 (N2). Upon stimulation, the majority of cells
occupy an expanding N2 population. Shrinkage of N2
occurs with resolution, although levels remain somewhat
elevated over time. Evidence from the MTdko mice
suggests a role for NFB signaling in the migration of cells
into the retina under both resting and stimulated
We previously reported that DC identified as CD11b+
GFPhi cells in CDG mice strongly responded to retinal
injury, and that their numbers often equaled those of the
MG, or other recruited GFPlo macrophages . DC are
central components in innate and adaptive immunity,
expressing a wide range of receptors that sample the
environment for molecules indicative of cell health,
infection or disease. Their extensive and close association
with RGC and their axons after an ONC raised the
possibility that they were active participants in the RGC
injury response. TLR, and several other receptors, are
important sensors linked to NFB via the adaptor
proteins MyD88 and TRIF . Some studies have shown
that neural inflammation was reduced, and neuron
survival enhanced following injuries in MyD88- or
TRIFdeficient mice [42-45], while others reported that
MyD88 deficiency led to exacerbation of the injury
response and less neuroprotection . We explored the
retinal DC response in mice deficient in one or both
Unlike MG, DC were transient participants in the injury
response, first observed interacting with the NFL at three
days post-injury, rising rapidly to peak at seven to thirteen
days post-injury, and then declining in number. The
subpopulation of retinal GFPhi DC that expressed CD45 at a
high level, resembling that of circulating myeloid DC
progenitors, was most active in acquiring the DiI label
contained in the RGC. GFPhi cells engulfing DiI-labeled RGC
soma post-ONC were readily identified by fluorescence
microscopy. GFPhi DC with CD45 expression at a lower
level similar to MG also took up DiI post-injury, as did
MG/macrophages that were CD45med and GFPlo.
Figure 10 A two-niche model for resident and recruited cells in retina post-injury. The sizes of the niches, N1 (MG/macrophage niche) and
N2 (GFPhi DC niche), as well as the arrows denoting pathways, were drawn to suggest relative sizes, and changes due to injury. VEC, vascular
A key difference between GFPhi DC and MG is that
relatively few GFPhi DC remain in the retina long-term,
whereas MG are defined as long-term resident
macrophages. Emigration of GFPhi DC from retina during
injury resolution appears unlikely to occur by lymphatic
drainage, but their numbers decline by an unknown
mechanism. Their decline in numbers suggests that the
numbers of DiI+GFPhi DC may be underestimated due
to turnover. A few thousand GFPhi DC appear, have the
opportunity to acquire DiI+ debris, and disappear with
their label over the course of two to three weeks, while
the number of MG remains relatively constant.
T cell adaptive immune responses may contribute to
neuroprotection. A number of neurodegenerative
diseases/injuries are sterile, so that the antigens are limited
to self-antigens. Nervous system self-antigen targets for
immunopathogenic CNS autoimmune responses were
reported to promote RGC survival post-ONC ,
perhaps by activated autoreactive T cells secreting
neurotrophic factors promoting neuron survival .
Selfreactive Tregs were found to reduce RGC survival [23,24].
DC are the prototypical antigen presenting cells, and we
showed elsewhere that retinal GFP+ DC present antigen
to antigen-specific T cells , and that retinal DC
upregulate MHC class II expression following ONC . These
properties may further link the role of retinal DC to T
cells and influence neuroprotection.
Although infiltrating monocytic cells were the major
source of TLR+ cells in traumatic brain injury ,
signaling pathways to NFB activation in retina are not
straightforward, a result that has been observed by
others [49,50]. In preliminary studies, we found that the
canonical signaling pathway via RelA is not well-used
in vivo by mononuclear cells in murine retina (data not
shown). Further, since the entire mouse shares in adaptor
protein deficiency in these knockout mice, some outcomes
may negate each other. This may account for results
showing neuroprotection  or a lack of effect [51,52] in
MyD88 and/or TRIF-deficient mice.
A two-niche model for resident microglia and transient
Several observations suggest the presence of two
overlapping niches in retina, the resident MG niche (N1)
maintained at a relatively stable number, and a transient
niche (N2) created by injury or inflammation that
becomes occupied by recruited mononuclear cells,
including GFPhi DC (Figure 10). The finding that sublethal
irradiation protects from glaucoma in DBA/2 J mice by
reducing monocyte entry in the irradiated retina/ON
 is consistent with a two-niche model. Retinal DC
appear to occupy N2, distinct from MG/macrophages in
N1. Relative to the GFPhi DC in CDG retina, their
recruitment in MTdko retina is substantially reduced (note
Figure 6), but the GFPhi cells that were recruited
exhibited a similar ability to take up DiI-labeled RGC
following an ONC. While the number of GFPhi cells peaked at
seven to thirteen days post-ONC followed by a rapid
decline, moderately increased numbers persisted for at
least two months post-ONC . No definitive marker
has been found to distinguish MG from DC, nor have
factors that sustain N1 or N2 been identified. If these
cells occupied the same niche, one would expect that
both MG and the GFP+ cells would decline in
proportion to their numbers, as the total number of cells
returned to pre-injury levels. However, in the absence of
catastrophic injury, N1 appears to continue to be
occupied by MG, as allelic or other markers of recruited cells
The chemokines and receptors that support
recruitment of circulating precursors into multiple tissues, C-C
chemokine receptor-2/C-C chemokine ligand-2 (CCR2/
CCL2) and C-X3-C chemokine ligand-1; CX3CR1,
CX3-C chemokine receptor-1 (CX3CL1/CX3CR1) , do
not appear to be strong candidates for mediating the
recruitment, function or maintenance of the cells in retinal
N1 , but may have more effect on N2. The effect of
CX3CR1-deficiency on experimental autoimmune
uveoretinitis (EAU) pathogenesis in mice was found to be
insignificant by one lab , but was associated with more
severe EAU by another . Further, no evidence for
effects on retinal development or injury repair was
associated with CX3CR1 deficiency . Evidence for
differences in migration or tracking was found in mice
deficient in both CCL2 and CX3CR1, but degenerative
changes in the retina were not found . Experimental
autoimmune encephalomyelitis (EAE) was found to be
more severe in CX3CR1-deficient mice , but in
models for ischemia-reperfusion injury in kidney ,
spinal cord injury , and atherosclerosis , disease
was less severe in CCR2/ or CX3CR1/ mice, and
attributed to reduced myeloid cell recruitment. Knockout of
either or both of these receptor/ligand pairs yields mice
whose retinal myeloid cells bear some resemblance to the
MTdko mice; that is the MG niche is largely intact, but
recruitment is diminished. This is likely due to the role of
NFkB in production of CX3CL1 on endothelial cells 
stimulated by IL-1 and TNF . Similarly, upregulation
of CCR2 mediated by TLR2 and TLR4 ligation  would
be diminished in MTdko mice, leading to reduced
expression of accessory molecules and integrins associated with
migration of cells into the retina [68-70]. Recent
experiments with depletion protocols that may allow specific
depletion of CNS MG may yield valuable information on the
N1 niche , and strategies to manipulate N2.
Use of MyD88 and/or TRIF deficient mice on the CDG
background provided insights into the retinal injury
response. Mice with single or double deficiencies in these
adaptor proteins had reduced responses, based on
finding fewer GFPhi DC in injured retinas post-ONC.
Although fewer in number, GFPhi DC retained their ability
to acquire DiI+ debris, suggesting that expression of
these adaptor proteins, and their role in NFB signaling,
was not a factor in their ability to phagocytose dying
RGC. Recruitment of GFPhi DC, and their phagocytic
activity, were distinct processes in which recruitment
was diminished, but phagocytic activity was not. Instead,
GFPlo MG were more likely to become DiI+, suggesting
that they filled the void left by the reduction in GFPhi
DC in MyD88/TRIF-deficient mice.
Additional file 1: Figure S1. Comparison of RGC counts/field obtained
by manual and autocounting protocols. FG labeling combined with
automated RGC counting was done to reduce observer bias. Field
size = 0.190 mm2.
AMD: age-related macular degeneration; ANOVA: analysis of variance;
B6: C57BL/6J; CCL2: C-C chemokine ligand-2; CCR2: C-C chemokine receptor-2;
CDG: CD11c-DTR/GFP; CX3CL1: C-X3-C chemokine ligand-1; CX3CR1: C-X3-C
chemokine receptor-1; DAPI: 4,6-diamidino-2-phenylindole; DC: dendritic cell;
DiI: di-alkyl-indocarbocyanine; DPBS: Dulbeccos phosphate-buffered saline;
DTR: diphtheria toxin receptor; DTx: diphtheria toxin; EAE: experimental
autoimmune encephalomyelitis; EAU: experimental autoimmune uveoretinitis;
FG: Fluorogold; GCL: ganglion cell layer; GFP: green fluorescent protein; IL-1:
interleukin-1; IPL: inner plexiform layer; MG: microglia; MyD88: myeloid
differentiation primary response gene (88); Mko: MyD88 knockout;
MTdko: MyD88/TRIF double knockout; NFL: nerve fiber layer; NFB: nuclear
factor kappa-light-chain-enhancer of activated B cells; ONC: optic nerve crush;
PMN: polymorphonuclear granulocytes; RGC: retinal ganglion cell; TLR2: toll-like
receptor 2; TLR4: toll-like receptor 4; TNF: tumor necrosis factor; TRIF:
TIR-domaincontaining adapter-inducing interferon-; Tko: TRIF knockout; TUNEL: terminal
deoxynucleotidyl transferase-mediated dUTP nick-end labeling; wt: wild type.
NH developed and validated procedures for flow cytometry and performed
all flow cytometry studies. MP performed all surgical procedures,
fluorescence microscopy and prepared photos. KM developed and validated
the RGC autocounting protocol. SM supervised all molecular biology
procedures for use and analysis of the transgenic mice. UL developed and
performed studies using local and systemic DTx to deplete dendritic cells. SH
performed and analyzed the TUNEL analysis of RGC. DF performed data
analysis and statistics, manuscript editing and major revisions. WL adapted
and demonstrated surgical procedures using the stereotaxic manipulator for
retrograde labeling of RGC. DG conceived of the study, oversaw design and
drafted the manuscript. All authors contributed to editing. All authors read
and approved the final manuscript.
Supported by R01-EY021003 (DSG); the Wallin Neuroscience Discovery Fund
(DSG); T32-EY07133 (UL); T32-AG029796 (SH); P30-EY011374; the Minnesota
Lions Clubs; and Research to Prevent Blindness, Inc. The authors thank Heidi
Roehrich and Thien Sam for technical assistance. We also thank Dr. J Douglas
Cameron for a helpful critique of the results.
All labwork on this manuscript was performed by the authors while they
were present at the University of Minnesota.
1. Voleti VB , Hubschman JP : Age-related eye disease . Maturitas 2013 , 75 : 29 - 33 .
2. Payne AJ , Kaja S , Sabates NR , Koulen P : A case for neuroprotection in ophthalmology: developments in translational research . Mo Med 2013 , 110 : 429 - 436 .
3. Chou CF , Cotch MF , Vitale S , Zhang X , Klein R , Friedman DS , Klein BE , Saaddine JB : Age-related eye diseases and visual impairment among U.S. adults. Am J Prev Med 2013 , 45 : 29 - 35 .
4. Thumann G : Prospectives for gene therapy of retinal degenerations . Curr Genomics 2012 , 13 : 350 - 362 .
5. Czeh M , Gressens P , Kaindl AM : The yin and yang of microglia . Dev Neurosci 2011 , 33 : 199 - 209 .
6. Amor S , Puentes F , Baker D , van der Valk P : Inflammation in neurodegenerative diseases . Immunology 2010 , 129 : 154 - 169 .
7. Peferoen L , Kipp M , van der Valk P , van Noort JM , Amor S : Oligodendrocyte-microglia cross-talk in the central nervous system . Immunology 2014 , 141 : 302 - 313 .
8. Streit WJ , Mrak RE , Griffin WS : Microglia and neuroinflammation: a pathological perspective . J Neuroinflammation 2004 , 1 : 14 .
9. Takahashi K , Rochford CD , Neumann H : Clearance of apoptotic neurons without inflammation by microglial triggering receptor expressed on myeloid cells-2 . J Exp Med 2005 , 201 : 647 - 657 .
10. Sanders VM , Jones KJ : Role of immunity in recovery from a peripheral nerve injury . J Neuroimmune Pharmacol 2006 , 1 : 11 - 19 .
11. Van Eldik LJ , Thompson WL , Ralay Ranaivo H , Behanna HA , Martin Watterson D : Glia proinflammatory cytokine upregulation as a therapeutic target for neurodegenerative diseases: function-based and target-based discovery approaches . Int Rev Neurobiol 2007 , 82 : 277 - 296 .
12. Akiyama H , Barger S , Barnum S , Bradt B , Bauer J , Cole GM , Cooper NR , Eikelenboom P , Emmerling M , Fiebich BL , Finch CE , Frautschy S , Griffin WS , Hampel H , Hull M , Landreth G , Lue L , Mrak R , Mackenzie IR , McGeer PL , O'Banion MK , Pachter J , Pasinetti G , Plata-Salaman C , Rogers J , Rydel R , Shen Y , Streit W , Strohmeyer R , Tooyoma I , et al: Inflammation and Alzheimer's disease . Neurobiol Aging 2000 , 21 : 383 - 421 .
13. Stagi M , Dittrich PS , Frank N , Iliev AI , Schwille P , Neumann H : Breakdown of axonal synaptic vesicle precursor transport by microglial nitric oxide . J Neurosci 2005 , 25 : 352 - 362 .
14. Qiu J , Whalen MJ , Lowenstein P , Fiskum G , Fahy B , Darwish R , Aarabi B , Yuan J , Moskowitz MA : Upregulation of the Fas receptor death-inducing signaling complex after traumatic brain injury in mice and humans . J Neurosci 2002 , 22 : 3504 - 3511 .
15. Ju KR , Kim HS , Kim JH , Lee NY , Park CK : Retinal glial cell responses and Fas/FasL activation in rats with chronic ocular hypertension . Brain Res 2006 , 1122 : 209 - 221 .
16. Davalos D , Grutzendler J , Yang G , Kim JV , Zuo Y , Jung S , Littman DR , Dustin ML , Gan WB : ATP mediates rapid microglial response to local brain injury in vivo . Nat Neurosci 2005 , 8 : 752 - 758 .
17. Lee JE , Liang KJ , Fariss RN , Wong WT : Ex vivo dynamic imaging of retinal microglia using time-lapse confocal microscopy . Invest Ophthalmol Vis Sci 2008 , 49 : 4169 - 4176 .
18. Damani MR , Zhao L , Fontainhas AM , Amaral J , Fariss RN , Wong WT : Agerelated alterations in the dynamic behavior of microglia . Aging Cell 2011 , 10 : 263 - 276 .
19. Blanch RJ , Ahmed Z , Berry M , Scott RA , Logan A : Animal models of retinal injury . Invest Ophthalmol Vis Sci 2012 , 53 : 2913 - 2920 .
20. Berry M , Ahmed Z , Lorber B , Douglas M , Logan A : Regeneration of axons in the visual system . Restor Neurol Neurosci 2008 , 26 : 147 - 174 .
21. Lehmann U , Heuss ND , McPherson SW , Roehrich H , Gregerson DS : Dendritic cells are early responders to retinal injury . Neurobiol Dis 2010 , 40 : 177 - 184 .
22. Kipnis J , Yoles E , Schori H , Hauben E , Shaked I , Schwartz M : Neuronal survival after CNS insult is determined by a genetically encoded autoimmune response . J Neurosci 2001 , 21 : 4564 - 4571 .
23. Kipnis J , Mizrahi T , Hauben E , Shaked I , Shevach E , Schwartz M : Neuroprotective autoimmunity: naturally occurring CD4 + CD25+ regulatory T cells suppress the ability to withstand injury to the central nervous system . Proc Natl Acad Sci U S A 2002 , 99 : 15620 - 15625 .
24. Kipnis J , Avidan H , Caspi RR , Schwartz M : Dual effect of CD4 + CD25+ regulatory T cells in neurodegeneration: a dialogue with microglia . Proc Natl Acad Sci U S A 2004 , 101 (Suppl 2): 14663 - 14669 .
25. Wax MB , Tezel G , Yang J , Peng G , Patil RV , Agarwal N , Sappington RM , Calkins DJ : Induced autoimmunity to heat shock proteins elicits glaucomatous loss of retinal ganglion cell neurons via activated T-cellderived fas-ligand . J Neurosci 2008 , 28 : 12085 - 12096 .
26. Jung S , Unutmaz D , Wong P , Sano G , De los Santos K , Sparwasser T , Wu S , Vuthoori S , Ko K , Zavala F , Pamer EG , Littman DR , Lang RA : In vivo depletion of CD11c(+) dendritic cells abrogates priming of CD8(+) T cells by exogenous cell-associated antigens . Immunity 2002 , 17 : 211 - 220 .
27. Li Y , Schlamp CL , Nickells RW : Experimental induction of retinal ganglion cell death in adult mice . Invest Ophthalmol Vis Sci 1999 , 40 : 1004 - 1008 .
28. McPherson SW , Heuss ND , Gregerson DS : Local 'on-demand' generation and function of antigen-specific foxp3+ regulatory T cells . J Immunol 2013 , 190 : 4971 - 4981 .
29. Heuss ND , Lehmann U , Norbury CC , McPherson SW , Gregerson DS : Local activation of dendritic cells alters the pathogenesis of autoimmune disease in the retina . J Immunol 2012 , 188 : 1191 - 1200 .
30. Jeon CJ , Strettoi E , Masland RH : The major cell populations of the mouse retina . J Neurosci 1998 , 18 : 8936 - 8946 .
31. Cone FE , Gelman SE , Son JL , Pease ME , Quigley HA : Differential susceptibility to experimental glaucoma among 3 mouse strains using bead and viscoelastic injection . Exp Eye Res 2010 , 91 : 415 - 424 .
32. Pang JJ , Wu SM : Morphology and immunoreactivity of retrogradely double-labeled ganglion cells in the mouse retina . Invest Ophthalmol Vis Sci 2011 , 52 : 4886 - 4896 .
33. Schlamp CL , Montgomery AD , Mac Nair CE , Schuart C , Willmer DJ , Nickells RW : Evaluation of the percentage of ganglion cells in the ganglion cell layer of the rodent retina . Mol Vis 2013 , 19 : 1387 - 1396 .
34. Zhang Z , Zhang ZY , Wu Y , Schluesener HJ : Immunolocalization of Toll-like receptors 2 and 4 as well as their endogenous ligand, heat shock protein 70, in rat traumatic brain injury . Neuroimmunomodulation 2012 , 19 : 10 - 19 .
35. Babcock AA , Toft-Hansen H , Owens T : Signaling through MyD88 regulates leukocyte recruitment after brain injury . J Immunol 2008 , 181 : 6481 - 6490 .
36. Esen N , Kielian T : Central role for MyD88 in the responses of microglia to pathogen-associated molecular patterns . J Immunol 2006 , 176 : 6802 - 6811 .
37. Miranda-Hernandez S , Gerlach N , Fletcher JM , Biros E , Mack M , Korner H , Baxter AG : Role for MyD88, TLR2 and TLR9 but not TLR1, TLR4 or TLR6 in experimental autoimmune encephalomyelitis . J Immunol 2011 , 187 : 791 - 804 .
38. Torres M , Guiton R , Lacroix-Lamande S , Ryffel B , Leman S , Dimier-Poisson I : MyD88 is crucial for the development of a protective CNS immune response to Toxoplasma gondii infection . J Neuroinflammation 2013 , 10 : 19 .
39. Kumar H , Kawai T , Akira S : Pathogen recognition by the innate immune system . Int Rev Immunol 2011 , 30 : 16 - 34 .
40. Pimentel-Coelho PM , Michaud JP , Rivest S : Evidence for a gender-specific protective role of innate immune receptors in a model of perinatal brain injury . J Neurosci 2013 , 33 : 11556 - 11572 .
41. Barton GM , Medzhitov R : Linking Toll-like receptors to IFN-alpha/beta expression . Nat Immunol 2003 , 4 : 432 - 433 .
42. Lin S , Liang Y , Zhang J , Bian C , Zhou H , Guo Q , Xiong Y , Li S , Su B : Microglial TIR-domain-containing adapter-inducing interferon-beta (TRIF) deficiency promotes retinal ganglion cell survival and axon regeneration via nuclear factor-kappaB . J Neuroinflammation 2012 , 9 : 39 .
43. Hanafy KA : The role of microglia and the TLR4 pathway in neuronal apoptosis and vasospasm after subarachnoid hemorrhage . J Neuroinflammation 2013 , 10 : 83 .
44. Lim JE , Kou J , Song M , Pattanayak A , Jin J , Lalonde R , Fukuchi K : MyD88 deficiency ameliorates beta-amyloidosis in an animal model of Alzheimer's disease . Am J Pathol 2011 , 179 : 1095 - 1103 .
45. Michaud JP , Richard KL , Rivest S : MyD88-adaptor protein acts as a preventive mechanism for memory deficits in a mouse model of Alzheimer's disease . Mol Neurodegeneration 2011 , 6 : 5 .
46. Downes CE , Wong CH , Henley KJ , Guio-Aguilar PL , Zhang M , Ates R , Mansell A , Kile BT , Crack PJ : MyD88 is a critical regulator of hematopoietic cellmediated neuroprotection seen after stroke . PLoS One 2013 , 8 : e57948 .
47. Mizrahi T , Hauben E , Schwartz M : The tissue-specific self-pathogen is the protective self-antigen: the case of uveitis . J Immunol 2002 , 169 : 5971 - 5977 .
48. Moalem G , Gdalyahu A , Shani Y , Otten U , Lazarovici P , Cohen IR , Schwartz M : Production of neurotrophins by activated T cells: implications for neuroprotective autoimmunity . J Autoimmun 2000 , 15 : 331 - 345 .
49. Mao X , Moerman AM , Barger SW : Neuronal kappa B-binding factors consist of Sp1-related proteins. Functional implications for autoregulation of N-methylD-aspartate receptor-1 expression . J Biol Chem 2002 , 277 : 44911 - 44919 .
50. Mao XR , Moerman-Herzog AM , Chen Y , Barger SW : Unique aspects of transcriptional regulation in neurons - nuances in NFkappaB and Sp1- related factors . J Neuroinflammation 2009 , 6 : 16 .
51. Famakin BM , Mou Y , Ruetzler CA , Bembry J , Maric D , Hallenbeck JM : Disruption of downstream MyD88 or TRIF Toll-like receptor signaling does not protect against cerebral ischemia . Brain Res 2011 , 1388 : 148 - 156 .
52. Drouin-Ouellet J , Gibrat C , Bousquet M , Calon F , Kriz J , Cicchetti F : The role of the MYD88-dependent pathway in MPTP-induced brain dopaminergic degeneration . J Neuroinflammation 2011 , 8 : 137 .
53. Howell GR , Soto I , Zhu X , Ryan M , Macalinao DG , Sousa GL , Caddle LB , MacNicoll KH , Barbay JM , Porciatti V , Anderson MG , Smith RS , Clark AF , Libby RT , John SW : Radiation treatment inhibits monocyte entry into the optic nerve head and prevents neuronal damage in a mouse model of glaucoma . J Clin Invest 2012 , 122 : 1246 - 1261 .
54. Kaneko H , Nishiguchi KM , Nakamura M , Kachi S , Terasaki H : Characteristics of bone marrow-derived microglia in the normal and injured retina . Invest Ophthalmol Vis Sci 2008 , 49 : 4162 - 4168 .
55. Niess JH , Brand S , Gu X , Landsman L , Jung S , McCormick BA , Vyas JM , Boes M , Ploegh HL , Fox JG , Littman DR , Reinecker HC : CX3CR1-mediated dendritic cell access to the intestinal lumen and bacterial clearance . Science 2005 , 307 : 254 - 258 .
56. Kezic J , Xu H , Chinnery HR , Murphy CC , McMenamin PG : Retinal microglia and uveal tract dendritic cells and macrophages are not CX3CR1 dependent in their recruitment and distribution in the young mouse eye . Invest Ophthalmol Vis Sci 2008 , 49 : 1599 - 1608 .
57. Kezic J , McMenamin PG : The monocyte chemokine receptor CX3CR1 does not play a significant role in the pathogenesis of experimental autoimmune uveoretinitis . Invest Ophthalmol Vis Sci 2010 , 51 : 5121 - 5127 .
58. Dagkalis A , Wallace C , Hing B , Liversidge J , Crane IJ : CX3CR1-deficiency is associated with increased severity of disease in experimental autoimmune uveitis . Immunology 2009 , 128 : 25 - 33 .
59. Zhao L , Ma W , Fariss RN , Wong WT : Retinal vascular repair and neovascularization are not dependent on CX3CR1 signaling in a model of ischemic retinopathy . Exp Eye Res 2009 , 88 : 1004 - 1013 .
60. Luhmann UF , Carvalho LS , Robbie SJ , Cowing JA , Duran Y , Munro PM , Bainbridge JW , Ali RR : Ccl2, Cx3cr1 and Ccl2/Cx3cr1 chemokine deficiencies are not sufficient to cause age-related retinal degeneration . Exp Eye Res 2013 , 107 : 80 - 87 .
61. Garcia JA , Pino PA , Mizutani M , Cardona SM , Charo IF , Ransohoff RM , Forsthuber TG , Cardona AE : Regulation of adaptive immunity by the fractalkine receptor during autoimmune inflammation . J Immunol 2013 , 191 : 1063 - 1072 .
62. Li L , Huang L , Sung SS , Vergis AL , Rosin DL , Rose CE Jr, Lobo PI , Okusa MD : The chemokine receptors CCR2 and CX3CR1 mediate monocyte/ macrophage trafficking in kidney ischemia-reperfusion injury . Kidney Int 2008 , 74 : 1526 - 1537 .
63. Donnelly DJ , Longbrake EE , Shawler TM , Kigerl KA , Lai W , Tovar CA , Ransohoff RM , Popovich PG : Deficient CX3CR1 signaling promotes recovery after mouse spinal cord injury by limiting the recruitment and activation of Ly6Clo/iNOS + macrophages . J Neurosci 2011 , 31 : 9910 - 9922 .
64. Liu P , Yu YR , Spencer JA , Johnson AE , Vallanat CT , Fong AM , Patterson C , Patel DD : CX3CR1 deficiency impairs dendritic cell accumulation in arterial intima and reduces atherosclerotic burden . Arterioscler Thromb Vasc Biol 2008 , 28 : 243 - 250 .
65. Garcia GE , Xia Y , Chen S , Wang Y , Ye RD , Harrison JK , Bacon KB , Zerwes HG , Feng L : NF-kappaB-dependent fractalkine induction in rat aortic endothelial cells stimulated by IL-1beta, TNF-alpha, and LPS . J Leukoc Biol 2000 , 67 : 577 - 584 .
66. Popovic M , Laumonnier Y , Burysek L , Syrovets T , Simmet T : Thrombininduced expression of endothelial CX3CL1 potentiates monocyte CCL2 production and transendothelial migration . J Leukoc Biol 2008 , 84 : 215 - 223 .
67. Souto FO , Alves-Filho JC , Turato WM , Auxiliadora-Martins M , Basile-Filho A , Cunha FQ : Essential role of CCR2 in neutrophil tissue infiltration and multiple organ dysfunction in sepsis . Am J Respir Crit Care Med 2011 , 183 : 234 - 242 .
68. Jimenez F , Quinones MP , Martinez HG , Estrada CA , Clark K , Garavito E , Ibarra J , Melby PC , Ahuja SS : CCR2 plays a critical role in dendritic cell maturation: possible role of CCL2 and NF-kappa B . J Immunol 2010 , 184 : 5571 - 5581 .
69. Ley K , Laudanna C , Cybulsky MI , Nourshargh S : Getting to the site of inflammation: the leukocyte adhesion cascade updated . Nat Rev Immunol 2007 , 7 : 678 - 689 .
70. Pober JS , Sessa WC : Evolving functions of endothelial cells in inflammation . Nat Rev Immunol 2007 , 7 : 803 - 815 .
71. Parkhurst CN , Yang G , Ninan I , Savas JN , Yates JR 3rd, Lafaille JJ , Hempstead BL , Littman DR , Gan WB : Microglia promote learning-dependent synapse formation through brain-derived neurotrophic factor . Cell 2013 , 155 : 1596 - 1609 .